Sunday, July 11, 2010

More Regenerative Braking Results. It gets worse...

In our first initial drives with the Mini, we were getting apparent ranges that matched my calculations for range initially without regen. We had assumed we would get some extension in range from having our first car with regenerative braking.

After just a few drives, it became apparent to me that with 100AH pack, and just under 1 AH per mile, that our range would be limited to about 100 miles - really quite similar to the BMW MIni-E. We had a larger pack, but also a larger car as we used the Clubman. But I was assuming we would get something from regenerative braking.

The question was of course what? Most comparisons indicate AH consumed and AH recovered and the Toyota RAV-4 even has a little indicator showing you your efficiency gains from regen based on this. I rather suspected that a real world comparison didn't yield the 21-25% gains that this measurement routinely provided. But I assumed 6-8% would be "real."

In the case of the Mini, it didn't look like we were getting much.

But we were busy completing the Speedster build and we have already heard from a number of people who want to purchase it from Special Editions Inc as a finished car when they can produce it. They can't produce it until we send them the prototype and bill of materials, so priorities being what they are, we've been playing Speedster.


Our first Speedster featured a series DC motor from Netgain, the venerable Warp 9, and a Kelly DC series controller with two banks of 90AH Thundersky cells. This makes a powerful package, but there just isn't an easy way to have regenerative braking with a series DC motor. The longest distance we drove this Speedster was 107 miles on a single charge, but to 80% DOD it's probably an 85 mile car safely.

In doing the new Speedster, we used just a bit smaller motor, the AC-50 from High Performance Electric Vehicles, along with a newish Curtis 1238-7501 3 phase controller. This does give us a regenerative braking capability that is quite flexible, and also allowed us a badly needed 1/4 inch to wedge in CALB 180AH cells in a single string. The AC induction motor is purportedly more efficient, though in very small numbers.

I did a max range test a couple of weeks ago and what number do you think came up on the odometer at the end? It was just over 107 miles.

What happened to the range gain from regen? Surely I must get something. In fact, the car is a bit lighter with the CALB cells but still got about the same range as the original. Now that's a rough approximation. Bringing cells to the brink of destruction WITHOUT destroying them is something I'm pretty good at, (though bettter at actually destroying them) but it's not precise. But we were pretty much run out. Even at 6 or 8% gain from the regenerative braking, I shoudl have hit 115 miles on the drive. What happened to the regen?

Regenerative braking, of course, is the concept of using the motor as a generator to slow down and stop a vehicle, instead of using the "friction" brakes that basically convert forward motion into heat in order to stop the vehicle.

So like Jerry McGuire - SHOW ME the REGEN. Where did it go.

Such questions are actually very good for me. They distract me from my full time vocation drinking whiskey and chasing younger women - though I'm easily distracted from that important work by technical questions.

To find out, we devised a "test drive" featuring lots of "regen opportunities" and of some length. Actually it was 48 miles or 76.6 km and takes an hour and a half to drive it.

We did the drive three times in three different modes, fully charging the car between each drive.

In the first drive, we used regenerative braking on the brakes only. We had fitted a 0 500 psi 5v pressure transducer to our Speedster and used it as an input to the brake pot input of the Curtis 1238. Curtis allows you to map this quite flexibly across the 5 volts and we used 100% regen braking current mapped across the first 2.5 volts of the pressure transducer output. This gave a great feel and we could do regenerative braking pretty much to max before the friction brakes came smoothly into play. We observed braking currents of up to 100 amps using this technique and typically still made current down almost to a full stop.

The drive departs EV Motor Verks up Spanish Street to William Street/Route K and proceeds west through a very busy area to U.S. Highway 61, known locally as the KIng's Highway. It was originally El Camino Real a Spanish Trail throught he territory.

It is now an unlimited access four lane with hundreds of shops and restuarants and commercial enterprises stretching about 8 miles out to Jackson Missouri.

We continued the drive through Jackson and out Country Road 335 to the site of Southeast Taxidermy, where on the first drive we had fried fish and Stag beer. We then backtracked to Jackson, went through the downtown area, and back onto Highway 61 out to Route 177. All of this is quite urban and quite busy.

177 is a two lane new "highway to the Proctor and Gambol plant. Speed limit 55 with really nothing along its length and so no stop and go here, though gentle elevation changes and curves. We turned off on county road V which is a narrow two lane blacktop road that is quite curvy and quite hilly. This reconnects to 177 back into Cape Girardeau, which we take through downtown main street back to EV Motor Verks.

The results of drive 1 indicated we had travelled 76.6 kilometers amd consumed 78.2AH of energy from the pack for a net 1.02088 AH per kilometer. Brian went along as camera man and so he was rather required on the two subsequent drives so that the car would be loaded the same.

On the second drive, we added a 50% of max regen current level to the "neutral braking" of the accelerator. What this means is that a certain map of the first volt of throttle signal does not accelerate the car, in fact it progressively decelerates it - though in reverse. At the top of the volt, you have minimum regen and as you remove pressure from the throttle it builds to 50% of the max current available from regen.

Driving this way, you can actually come to a full stop as if you had applied brakes simply by taking your foot off the accelerator. Indeed we hardly touched the brake pedal and accomplished nearly the entire drive just by modulating the throttle in this way.

We DID miss the turn to 335. So we turned around and went back into Jackson anyway, cutting about a kilometer from the drive.

This, what I consider more "aggressive" regen is actually pretty interesting when you get used to it. You are pretty much always either accelerating or decelerating. The "sweet spot" of zero regen and zero acceleration is not very broad. The results of drive two were 75.2 km traveled, 77.7 AH and 1.03324 AH per kilometer. With MORE regen, it appeared it was actually LESS efficient????

In drive three, we disabled all the regen. I DID make the turn and so we recorded the exact 76.6 kilometers for 78 AH or 1.01827 AH per kilometer.

We're not answering questions here. We're just bringing up new ones. Not only was there NO gain from regenerative braking, but our most efficient energy use was WITH NO REGEN AT ALL. And the three drives were in a direct line from MAX REGEN being least efficient, to SOME REGEN in the middle to NO REGEN as the best. How can this be?

This week we repeated the test and we used the same drive. But we used a different car. The 2009 Mini Cooper Clubman.

This unit features regenerative braking as well. But it has a different motor - the MES-DEA 200-250 and a different controller - the TIMS 600. All the parameters are of course different. But it is interesting in that it is FRONT WHEEL DRIVE instead of rear wheel drive. And instead of 2000 lbs, it weighs 3500 lbs. It is a different pack size at 375 volts and 100 AH, as opposed to the Speedster's 120v and 180AH. In fact, almost everything about it is different except for the measurement device, a TBS Expert Pro with a 50 mv 500A shunt.

The regen is used on both brakes and throttle. On the throttle, it is very mild, simulating the back pressure from the compression of an ICE engine very nicely. On the brakes, we did not employ a pressure transducer but trigger it on the brake light, which comes on at the lightest touch of the brake pedal. And it is progressive in that it builds over a 2 second rate. Often, on long stops, this leads to a pulsing feeling. As the braking builds, you let off the pressure but it resets and builds again very quickly so that the effect is of a gently pulsating regenerative braking.

On both cars, we can easily observe regenerated currents of up to 100 amperes. And indeed on a long downhill you can observe the AH counter tick backwards.

On drive 1 in the Mini Cooper, our odometer recorded 48 miles and our elapsed time, which we tracked for this version of the test, was one hour 28 minutes for an average speed of 32.72 mph. The drive end to end required a total net energy from the pack of 47.1 AH.

For drive 2, we simply switched the regen off totally using a switch we had mounted on the dash for this purpose early on. The car feels best taking off in 3rd gear, and in fact, we drove both these drives in 3rd gear the entire time. In first gear or reverse, the additional gear leverage makes the regen very uncomfortable at low speeds. So we had wired this switch to turn it off while backing up, parking or in inclement weather.

One of the downsides of regenerative braking is that it is reported to be a bit hard to control the car on ice and snow. So we had installed this disable switch.

We recharged and drove the same drive the next morning with regen totally disabled, again recording 48 miles on the odometer which does not provide tenths, and required one hour 25 minutes to complete for an average driving speed of 33.88 mph. It required 43.8 AH - some 3.3 AH less or 7% less than our original drive with regen.

Again, I find this result astounding, but very much in keeping with what we found on the Speedster. The reason we were not enjoying any gain in range in either the Speedster or Mini Cooper is that regen wasn't providing any. Indeed, it seems to carry a penalty.

The comments since have been both frustrating and entertaining. The heart of the angst seems to be that they cannot envision what happened to the recovered energy during braking. They seem strangely willing to forfeit the throttle gain, but unwilling to do so in the case of the brake peddle.

The results are the results as far as I'm concerned at this point. We started with the observation that we were not seeing any gains in efficiency from regen, and after five 1.5 hour drives, we pretty much know why. Regenerative braking, either throttle actuated or brake actuated, does not provide any.

But it is not unusual for the answer to a question to trailer along a series of further questions. The regenerative braking traces back to very early electric cars. The 1906 Krieger Electric Landaulet featured regenerative braking and sported a range of 50 miles and a speed of 18-20 mph.

So the theory is not precisely recent.

So the question remains: Why no efficiency gains, either in experienced range or measured energy usage, from two different cars with entirely different regenerative braking schemes?

We don't know. But it has always seemed to me that the basic concept was a little flawed. The highest and best use of the kinetic energy stored in the forward motion of a vehicle has to be to remain in motion. Traffic laws being what they are, you eventually have to brake to slow down. And regenerative braking is pictured as a way to recover some of that energy. Whether it is a large amount that is recovered or a small amount is not really the problem. You don't have to be faster than the cheetah. You only have to be faster than the smallest gnu in your herd.

So how do we have a NEGATIVE amount???

There really aren't a lot of options here. We must have a phantom drive at play. And I think this is where the answer lies. The drive not driven. The path not taken.

Cars and drivers interact. Adding regenerative braking absolutely alters the driving characteristics of the car. It is my observation that with regenerative braking on the throttle, you really don't coast along very much. You are always either accelerating or decelerating. There is a bit of a sweet spot at around 0 amps but it's a little tricky to find and stay in. And it is just easier to accelerate and decelerate as traffic demands. You quickly adjust to the car and you do indeed drive it differently because of that.

What if the drive with NO regenerative braking is significantly different than the drive with regenerative braking?

By way of a thought experiment in simplest form, let's picture a drive of 5 miles on perfectly straight, perfectly flat road. In the first case, we will accelerate to 45 mph and on achieving that speed we will maintain it precisely at 45 mph until 200 feet before the end of the drive, where we will manfully haul in the binders and bring the car to a stop.

There is a certain level of energy required to accelerate the car to 45 mph. This energy must be sufficient to overcome the rolling resistance of the car, the wind resistance, but it also must be sufficient to accelerate the full mass of the entire vehicle to 45 mph ON TOP OF THAT. And that is beginning at an inertia of zero.

Once there, the energy required to maintain 45 mph is relatively trivial. We need only replace the energy from the parasitics of rolling resistance and wind resistance.

Let us say that the total energy on this drive requires 5 AH.

On a second drive, we accelerate to 50 mph. At the moment we hit 50 mph we immediately decelerate to 40 mph. Once at 40 mph we immediately accelerate again to 50 mph. And we continue this oscillation until the end of the 5 miles, averaging 45 mph for the entire distance.

It took more energy to accelerate to 50 than it did to 45. We recapture some of the kinetic energy stored in the forward motion of the car when we decelerate using regenerative braking. But our efficiency in doing this is less than unity and in fact let's use the number 80% for the sake of argument. On reaching 40 mph we again have to apply energy to accelerate back to 50 - in fact 100% of the energy required to do so. But once there, we again decel at an 80% efficiency. We do 200 such oscillations in the course of the 5 miles.

It would appear obvious that it would require notably and measurably more energy to drive the car with the oscillations, even with the resulting recoveries from regenerative braking, than to maintain a steady state of 45 mph.

But our real world drive was much closer than that. Ok. Let's revise the thought experiment. Let's say we accelerated to 47 mph. And we decelerated to 43 mph. Are we getting closer?

Over the 48 miles of stop and go driving, a lot is going on. We know with some certainty that our measured results showed not only little gain, but negative gain. And that is the clue.

The regen induces oscillations in the driving characteristics of the car, and we as drivers adapt to them. If we compare the measured gains from the regenerative braking to the THEORETICAL drive of the SAME energy use that we WOULD have had without regen, the gains are real and apparent.

But if we compare to a REAL drive that we ACTUALLY drove without regen, we get a very different result. And this is because we lack the changes induced by the regen itself.

We do know that driving style can dramatically alter the energy consumption in both gasoline and electric cars. Electric cars SEEM more sensitive to this. We can readily observe that just such oscillations are induced when we have regen on the throttle.

With regen on brakes only, most people's ability to picture this breaks down. However, the drives indicate the following:

Throttle plus brake - least efficient
Brake regen only - in the middle
No regen - most efficient.

It would fit that while the effect may be LESS pronounced with brake only regen, it remains a factor. And this would explain a penalty for regenerative braking of any kind, compared to a REAL drive with no regenerative braking at all.

So that's our working theory at the moment.  I'm actually quite confident of the test results.  They match the initial observation that we weren't actually gaining any advantage from regenerative braking.  And they explain that rathe nicely - there aren't any such gains.

On the why, I'm a little less confident.  I like the scenario in that it accounts for all the observations.  But it is curiously difficult to communicate succinctly.  Usually, if I struggle to present, it is because I don't REALLY understand it yet viscerally.

Jack Rickard


  1. " But I assumed 6-8% would be "real." "

    HaHa. Your assumption was correct, and it was real. It was just not in the direction that you assumed it would be in.

  2. Jack, I am interested in the battery voltage reading during accel, during steady state, and during aggressive braking with regen. I would want them to all be read close to the same time so the battery state of charge is essentially the same. Do you have these numbers or at least approximately?



  3. Jack,

    I want to be sure I understand your current theory. Is it that the act of using regen changes the driving technique such that it nullifies any benefit that might have been realized? And I think I hear you saying that regen works, but when it comes to practical use in the real world of driving it not only has no benefit but even becomes a liability. Do you suppose this result extends to the "typical" american driver should plug-in electric cars become the norm?

    Thanks for the dialog. I have really enjoyed it!



  4. All very interesting. I don't have an electric car but I'm following your scientific methods intently. My wife detests your rambling but did remark on your haircut this past week. Keep up the good work.


  5. I think it pretty much wipes out regen in all forms I am aware of. And yes, what I'm saying is that it modifies your driving in a disadvantageous way.

    The best I've done is the 5 mile drive. No one seems to object that an oscillation uses more energy than a steady state 45 mph. And everyone seems onboard that this would be the case with the throttle based regen.

    The intuitive leap we don't seem to want to make seems to focus on the brake regen. I keep saying it is the same thing, just not as much of it. And indeed, brake only regen gets us a better test result than throttle regen, so everything is marching in sync here.

    But you have to put on the brake eventually anyway. So why can recovering energy be a bad thing???

    Let's take a look at that. On the Mini, we turn on regen just on the brake light signal. On the Speedster we use a pressure transducer that is variable. And note that the results were WORSE on teh Mini than on the Speedster. So stay with me, every observation is still lined up correctly.

    Even on the Speedster with brake transducer, we WANT regen to kick in before friction brakes. That's the whole point. And as I've said, they feel like power brakes. But Speedsters don't HAVE power brakes.

    Now if I put my foot on the brake pedal, and don't really apply pressure, the brake light on the Mini will come on, and regen begins anyway.

    Similarly, but perhaps to a lesser degree, just touching the brake will start regen, and by design we start it before the regular friction brakes, and indeed we want some regen before we even use friction brakes. But of course, we made no modification to the friction brakes to make them occur LATER. We incorporated regen where IT would come on SOONER.

    So now, with a very light touch on the brake pedal, we start to generate regenerative braking, where normally we would do nothing. In turn, we start to decelerate, where without the regen we normally would not have that effect.

    To maintain speed, then we have to accelerate PAST the steady state speed. And so we need to DECELERATE - and of course as soon as our foot TOUCHES the pedal, we are indeed decelerating and generating current. Do you see how the oscillation occurs? Its basis was a more sensitive brake. And we did that on purpose to preempt the regular brakes.

    And my theory is that if you set up this oscillation, you can never surpass the energy efficiency of the steady state speed maintained by the tortois. Instead, you are doomed to living the life of the hare. There IS no recovery.

  6. So as I have said REPEATEDLY, the brake regen is EXACTLY like the throttle regen, just less so and a little harder to conceptualize.

    Rewind. No regen. We're cruisin. We don't get braking until we HIT the brakes, not when we're playing around on the pedal ready to brake. And the amount of energy required to maintain speed plummets because we are not oscillating about the steady state average speed. Remember, we arrived 3 minutes SOONER on the Mini as the tortois.

    End to end, real world, an hour and a half, it comes out a wash. Yes, there was energy recovered in regen, but it really WAS yet another version of perpetual motion all along. The costs exceed the gains.

    But to measure it you actually have to DO the alternate drive, and you have to DO it without regen at all.

    It really is a very dynamic environment, and truly is hard to model or hard to visualize. For a hundred years there really hasn't been a question that regen is cool.

    Note that almost everyone in Formula I is also dropping out of KERS. I suspect the same thing. It doesn't even pay for the weight of the components. They don't really know why. They just aren't getting what they "should be getting." Which is where we started on this topic.

    But right now EVERYONE is talking about an electric car design, and they ALL include regen, and they want US to pay for it. And it appears to be entirely illusion. Buy me bigger batteries.

    Jack Rickard

  7. Might be interesting to put a cell log 8 on some of the cells and see what they look like on a short steady state section with and without regen. If you are correct you should see more oscillations on the regen trip, you can zoom in pretty far on the graph for fine resolution.
    You also might find a way to give yourself a larger zero regen sweet spot on the throttle. A longer pedal arm might help, maybe change some parameters in the controller so the change between drive and regen is less abrupt. Can't remember the exact parameters but something like accel ramp and deccel ramp, throttle release rate. I can look them up if you're interested. Seems as if the brake pedal regen loss as you describe it could be made up by simply braking a little later and harder than normal.


  8. Jack, thanks for the explanation of the shunt device and how you measure AMPs in and out of the battery. Your efforts are greatly appreciated.

    As you said you have to trust the meter. I've told my guys that same thing "you have to take the results that the meter gives you and live with it you can't second guess the equipment it just doesn't fail that often"

    I would just like to say I think what is happening is a clear example of the second law of thermodynamics.

    On your theoretical 5 mile drive you loose energy as you stated when you oscillate your speed.

    That law states in laymans terms that any time you move energy around you end up missing some energy (these losses are through friction and heating of components). These losses are known as entropy.

    The engineers at the OEMs should be on this, I can't believe they would intentionally sell a product that is more complicated and expensive than necessary.(just kidding)

    I have to stop now, it's starting to hurt this "double think". Thinking you must be able to get a benefit from regen and not only not getting any benefit but going in the energy hole!

    Regards, JMS


  9. Hi Jack, excellant videos! I believe that the regen loss is due to friction and heat generation. Maybe the batteries just can not recharge this extra energy as efficiently as it releases, hence using ultra capacitors to store and release the regen energy seems to be the best answer to this problem. I wish I had the funds to build and experiment electric vehicles. You are living my dream, keep up the good work.
    Jerry C.

  10. Hey Jack, I have another idea here, an experiment to perform would be to actually "drive" the AC-50 drivetrain with another motor and measure actual reharge capability with the entire drivetrain system. Its probably an experiment that would be cool to do but would involve too much effort and expense, what do you think? It sure would answer some questions.
    Jerry C.

  11. Jack , Just a quick thought. The bmv has a parameter called charge efficiency factor (cef). This can be set anywhere from 1.0 to 1.9. At 1.0 it assumes that all charge current is stored in the battery all the way down to 1.9 which assumes only 10% of the injected current is stored as charge. Not sure if the xantrex has this setup but if so it might be worth a look. The recommended cef for my batteries (agm) is 1.25. Not sure what it would be for lithium.

  12. "The recommended cef for my batteries (agm) is 1.25. Not sure what it would be for lithium."

    I would use 1.00.


  13. Hi Jack,

    This may be a silly the motor speed controllers reconvert 3 phase AC into DC during the regen process? does it have some sort of retification circuit before it back feeds into the batteries?



  14. JMS: Hmmm...

    I was taught that entropy is the tendency toward randomness, and that enthalpy is the heat of a closed thermodynamic system under constant pressure.

    Of course, these vehicles are neither closed nor under constant pressure. It is instructive to think about the heat, though.

    The service brakes dissipate heat extremely efficiently straight out of the car- vented rotors are big axial fans and they work well. The regen system is no doubt putting some heat into the car and its components that the brakes would be entirely removing from the chassis, and that would only be worth doing if there was enough of a benefit from regen to justify it.

    Its a marginal thing, but once again, the regen system isn't worth its drawbacks.

    Jack, I wouldn't make too much of the KERS outcome as an indictment of regen; the teams "agreeing" to abandon it really has a lot more to do with rules than anything intrinsic to the systems, except that the complexity and engineering burden was actually more than all but the top two teams were willing to tackle in the first place, so in that limited sense there is a familiar regen issue there...


  15. One last thing about KERS:

    It has nothing to do with efficiency.

    F1 is about the precisely directed discharge of extravagant amounts of energy, not its conservation. KERS works beautifully to make a car almost impossible to pass, or to give it that little extra boost to make it more capable of passing other cars. That's all it does.

    I don't think any team ever bothered to see if it extended range, and it certainly doesn't add any braking to these cars. Whether it cut top speed is something only Ferrari and McLaren engineers would know, but it wasn't apparent on the track last year.

    KERS was so limited by the rules, to just a few seconds of use per lap, that it really wasn't worth the trouble to develop, provided nobody else was doing it, which is what the teams agreed would happen this year.


  16. There IS a charge efficiency factor in both the Xantrex and TBS Expert Pro devices. If you set it to 100, it is effectively shut off. It occurred to me that it could have an effect, and we have it set to 100.

    Heat and friction are very real, but they are always with us and I do not believe they change anything in this scenario. We have them on expenditure, and we have them in equal measure on recovery.

    The more I've thought about it, the more I believe it has to do with unnecessarily accelerating the full mass of the car, to overcome the effects of decelerating the mass of the car - the oscillation. On a round trip oscillation, you are GOING to lose in this transaction. Perhaps not much, but you're going to lose some every time.

    To help with the lost math:

    On a drive with no regen - SPENT 78AH RECOVERED 0 AH - NET 78 AH

    On a drive with regen - SPENT 86AH RECOVERED 7.8 AH - NET 78.2 AH.

    In order to prove that, I would need TWO counters that counted only in one direction. An OUTIE and an INNIE. I'm giving some thought as to how to accomplish that now. Sounds easy. The more I look at it, the harder it becomes. Any ideas?

    Dan - I just refuse to be pulled off into the weeds on kWh. But I will give you a couple of hints.

    1. Plug in whatever voltages suit you and do the math. We're talking about 100wH and less than an AH however you do it.

    2. Batteries are charged by current. You do have to have a voltage above the terminal voltage to cause current. But you only get the benefit of the current itself - as measured by AH. You can apply 700volts or 200 v - they will RESULT in a different current, but it will BE whatever current goes through the shunt. That's what we're measuring already. Don't wander off into the weeds...

    And that is kind of the theme of the test. Let's cut all the crap and go to the END points and measure what the actual RESULT is end to end.

    The main result is there is no benefit to regenerative braking, in any form.

    The anomaly in the result is that somehow we actually generate, and I have to point out CONSISTENTLY generate, a NEGATIVE number. While initially surprising, when you stop and consider that all additional complexity USUALLY results in a negative number.

    Consider the perpetual motion guys. They ALWAYS think if they can get just ONE MORE step in there somehow the laws of the universe will invert and at SOME SCALE it will work. As we know, when you scale up a loss, you get a bigger loss, and another intermediary step will simply compound the losses.


  17. "The engineers at the OEMs should be on this, I can't believe they would intentionally sell a product that is more complicated and expensive than necessary.(just kidding)"

    I do not believe anyone has intentionally run any sort of scam. I think the regen thing has been pretty much thought a proven concept. As I pointed out, it has been around for over a hundred years. And you can easily measure the very apparent benefit - AH out, and AH in. But they never looked at the "drive not driven" - that is without regen at all, in the comparison.

    Now I AM surprised at that. It would seem SOMEONE would have done this at some point. It's easy enough to turn off. And there is no magic in driving a car around for an hour. We didn't need to call NASA in to supervise this or anything. We're using the amp-hour meter, we haven't gone to a different measurement metric.

    And yes, believe the meter. When you find yourself "tapping" the face of an LCD meter to try to get the needle to center, you know you've lived too long. But that's what I wanted to do here. When you get an unexpected result, it MUST be the meter. It's just a natural reaction.

    The corollary is repeat the test. We did. Did we do it enough? I pretty much don't care. We're not the National Institute of Science and Technology. We just want to know why Speedster Duh has no more range than the first Speedster. And we want to know just how MUCH we should spend on the holy grail of EV land - the 200 kw AC induction 3-phase system. If regen is no benefit, how much will you spend to avoid changing brushes every 4 or 5 years?

    On the other hand, I find the HPEVS AC-50 to be just an excellent power train anyway and at $4500 it really is not any more expensive than you would generally spend on a drive anyway. A Soliton1 ($2895) and a Netgain Warp 9 ($1700) will give you more power I would think, but you're at the same price point.

    But as you go up to larger cars and larger power trains, the cost of AC power trains go up. We spent $13,000 on the components in the Mini power plant, when that same Soliton and Warp 9 would have done the job admirably.

    The AC Power system is $25,000, the UQM 150kw is $34,000. For regen?

    How about a couple of honkin Netgain 11's strapped to the same shaft with dual 1000 amp controllers? Tim the Tool Man - "Mo Power'. AND we can use lower voltages - 170 vdc. 340 kW on the ground.

    Jack Rickard

  18. Jack, How about trying my idea of using a cell log 8 for about a mile of steady state driving, with and without regen, and see what the graph looks like? If the oscillations are there as you say it would be a pretty good graphical representation of what you are claiming.


  19. "In order to prove that, I would need TWO counters that counted only in one direction. An OUTIE and an INNIE. I'm giving some thought as to how to accomplish that now. Sounds easy. The more I look at it, the harder it becomes. Any ideas?"

    What about switching the polarity on the current shunt? This would make any offset stand out and the average would remove the offset - assuming it it the same (which it usaully is).
    Standard procedure when measuring input/output efficiency on electrical converters and using non calibrated equipment.


  20. JRP:

    I wasn't ignoring you. I'm thinking. It hurts. I actually have cell logs on two of the cell access plugs on the mini and can record 16 cells.

    I would have to do two drives OVER, and that's painful. But I kind of like your idea. The problem is, it's almost TOO MUCH data. If I had it, what would I see and how would I make any sense of it? I would have to have some means mathematically of boiling down the extent of deviation from average across a one and a half hour drive into a value.

    If I did that twice, with regen and without, and had two numbers, they would either be the same or different. If substantially different, bingo. What is substantially?

    I might just do it and publish the graphs and see what the world makes of it. The coupling between cell voltage and current in and out is real enough although another step removed.

    Jack Rickard

  21. Jack,
    I notice that in all your tests so far, you've driven the regen drive first, followed by the no-regen drive. For future test, you might want to reverse the order just to remove another variable. You haven't said anything about cell temperatures. But if your cells are sitting at or close to ambient at the beginning of the regen drive, then they warm up during the drive and subsequent charging, they'll be at a higher temperature during the non-regen drive. My experience is only with the older Cobalt cells, but when they're up above 50 degrees C, they exhibit quite a bit less sag, which would lead to a lower Ah number due to the higher voltage.


  22. Martin:

    Yes, that would eliminate any offset. But what I wanted to see was total AH out, and total AH in separated - not net. Two meters, yes, reversed in polarity and we can use the same shunt. But how do I get them to quit totalizing the reverse?

    Sounds like I need to find two different meters than what I use.

    I would probably have to do this with an Arduino and LEM-HASS current sensor.

    Jack Rickard

  23. Jack,

    In my world of complex electro-mechanical machine design there are 2 necessary components for success; a good idea and a good implementation. If either is missing the outcome is not going to be good. You are saying regen doesn't work because because of oscillations (I'm not ready to buy that, but for the moment lets address it) in speed. If that is so, it sounds like an implementation problem so let's fix it! Any ideas?

    If you really, really, really want to know what is happening I would suggest the following data gathering. Get a data collection device (could be a laptop with some interface electronics and software) with 2 A/D channels. Monitor the shunt voltage bi-directionally for current on 1 channel and monitor the battery voltage on the other. Set it up to take a reading every second or 2. As you drive your course, write down the time at the start of every mile (if you have an iphone the digital stopwatch lap function would make this very convenient)so you can correlate the data positionally from multiple runs. I am most interested in the comparison of brake based regen vs no regen. The data will show exactly what is going on including oscillations. Anything short of this precise data gathering will leave us forever speculating.

    What do you think?



  24. Jack, your evidence is convincing. Whatever the explanation eventually turns out to be, the fact that no real world regen benefit exists is the bottom line for AC motors, and the discussion must move on to other reasons why AC technology should be deployed, if at all. On the Speedster, the SIZE of the motor is critical in order to fit in 36 cells, and the cost differential is minimal. Thus, the AC50 makes sense for the Speedster, perhaps with a little brake regen just to give a power brakes feel to the car.

    Size alone, however, is a thin advantage to base a business on (and will vary by type of application and cell size formats anyway), so AC motor/controller manufacturers must dramatically lower costs to remain viable. Without regen, AC has no significant TECHNICAL advantage over DC. It will come down to comparable costs.

    This evolution is very similar to trends in other technology areas, like computers, software, etc., where claims of game-changing "features" eventually just melted away to dollars and sense.

  25. Dan:

    You move from one thing to another all in a quest to GET ME TO WORK HARDER.

    Now you want me to design and build a data acquisition system to prove what I already know and have measured? And to produce what?

    I'll tell you what. YOU design and build it and send it to my house. I'll hook it up and run it if you make it simple enough to do so.

    Still puzzling over the cell log 8 idea...

    Jack Rickard

  26. Kriss Motors:

    Well yeah, that's the whole idea of doing the test and sharing the results. We need to make informed component selects when doing a build, and if you take regen off the table, the whole ecosystem geography just changed.

    Truly, the precise reasons for the failure to gain range when using regenerative braking are academically interesting. But only to a point for our purposes. There are many better equipped than I to charge off down that road.

    At this point, I'm pretty sure the guy didn't actually saw the lady in half, and she does NOT need medical attention.

    To take this a step further, we suspected we were not getting any additional efficiency from regenerative braking. At this point we've pretty much proven it. I don't think there is a magic switch or valve we can turn to get some. We can clearly see we are generating current in impressive amounts, though for short durations from regenerative braking and that in a real world drive, it does NOT result in any gain in efficiency for our car. From there, how many angels can dance on a pin becomes a bit academic. We may indeed be able to move things around and resolve the negative number back to zero somehow. Does that get me 6-8% gain? Maybe 5%? I'll take anything I guess but below about 3% I have to start counting ducats and comparing them to amp hours.

    JRP, yes I can change the attack elements on regen so that this gets better. If I change them enough, it won't come on at all and that will be as good as it gets. We WANT regen to come on before the friction brakes. If they come on after, we aren't doing much. And I can turn it off with a switch now.

    Jack Rickard

  27. "I notice that in all your tests so far, you've driven the regen drive first, followed by the no-regen drive. For future test, you might want to reverse the order just to remove another variable."

    A valid thought. I'll add it to my list of all the things I should have done differently.

    Truly, we have to start with a fully charged car. It's pretty much overnight. The tests were run on subsequent days. The ambient temperature on the Speedster tests were essentially identical. It's true, the second day on the MIni was substantially cooler.


  28. Jack,

    "You move from one thing to another all in a quest to GET ME TO WORK HARDER."

    First you accuse me of not letting go of an idea, now you say I am jumping around? I have been beating the same drum the whole time! C'mon now, you started the debate with bold proclamations. Not everyone is convinced we have a full understanding yet, including Jack!

    "I'll tell you what. YOU design and build it and send it to my house. I'll hook it up and run it if you make it simple enough to do so."

    I really would like to take you up on that, but additionally I would deliver it personally and assist in the hookup and test. Let me see what I can do.

    "Usually, if I struggle to present, it is because I don't REALLY understand it yet viscerally."

    This is a telling statement that I hope you hold on to!



  29. Tom Alvary, please don't confuse enthalpy with entropy.

    enthalpy is a measure of -potential energy- In a closed heat engine. Enthalpy is the capacity to do non-mechanical work plus the capacity to release heat.

    Entropy is the tendency of a system over time to move toward equilibrium without- external- inputs. Entropy is why you can not have a perpetual motion machine.
    Entropy is why my garage is a mess.
    It takes directed external energy inputs (my labor) to pick things up and arrange them on a shelf.
    Given enough time the potential energy of the items on the shelves will go to equilibrium (when the shelf finally collapses).
    Entropy is why a bag of nuts and bolts riding in the car vibrating and shaking do not assemble themselves. And in fact if they are preassembled they will disassemble when riding around with random vibrations.

    Please correct me if I am wrong.

    I checked around and the electric bike riders do not use regen. But the electric subway trains do use regen and claim 17% benefit. Then Jack gets even less efficiecy with a heavier car. Are the subway guys WRONG?

    regards, JMS

  30. "Yes, that would eliminate any offset. But what I wanted to see was total AH out, and total AH in separated - not net. Two meters, yes, reversed in polarity and we can use the same shunt. But how do I get them to quit totalizing the reverse?"

    A couple of diodes and two shunts?


  31. "Entropy is the tendency of a system over time to move toward equilibrium without- external- inputs. Entropy is why you can not have a perpetual motion machine."

    I'm much more comfortable with your definition. It was actually coined in terms of the second law of thermodynamics and was originally observed as the fact that heat will not move from a colder environment to a warmer one. This has come to mean a natural motion toward equilibrium. - two items of different temperatures will seek equilibirum.

    I am very uncomfortable transferring one set of experimental results to a different millieu except as postulate. I think most of the problem is with the wetware- not the hardware. The DRIVER adapts to the changed dynamicds of the car and negates the benefits of it. That might work wiht a bicycle, and it might not. It probably would NOT work with a subway. They basically select speed with a lever and leave it there. And there is huge inertia on their decel to a stop. Still, 17% is hard to fathom.

    Jack Rickard

  32. Haven't mainstream hybrid cars of the last decade been proving that regen is more efficient than no regen? Electric cars like the GM Volt and Nissan Leaf have tested and tuned their regen extensively. Unfortunately they often tune to help the heavy-footed drivers rather than tuning for optimal efficiency.

    I hate "coasting" regen; it should only come on when you hit the brake pedal.

    I recommend allot more test runs and staggering the orders of the regen setups. The data is not statistically significant yet to draw firm conclusions.

  33. Martin:

    They would have to be 500 amp diodes and I introduce loss through the forward voltage drop of the diodes.

    I've ordered an RS-232 communications kit for the Xantrex. I have never used it or the software, but it woudl make sense to me that it will allow me to break out AH out from AH in. We'll see if it does in fact allow that. Might even have a logging feature.... I just don't know from the product description and data sheet.

    Jack Rickard

  34. I purchased the TBS variant and it looks very straight forward. But one thing about the xantrex (and similar equipment) - you don't know the sampling rate and it will act as a lowpass filter.

    The xantrax will default dump measured data each second on the communication line (at least the TBS does).

    I might have a couple of 300A IGBT's lying around. They come equiped with an internal reverse diode of the same rating - and yes it would produce quite a loss, but that would not matter for the experiment. They would have to have a heat sink and a fan ;-)


  35. Jack,
    A motor does not make a good generator, in order to generate current you must have a lot of lines of force provided by the field coils. In order to generate current you must have a combonation motor with seperate fields that are switchable. Regen golf carts have smaller field coil windings with regen feature.

  36. "A motor does not make a good generator"

    Great point. Although a motor can be a generator, I'm sure it ends up being a suboptimal compromise for both. I was in Japan recently and was amazed to see a cutaway of Toyota's latest Hybrid with Regen. What surprised me was it had a separate electric drive motor and regen generator. It was way more complicated ($$) than using the motor for both purposes. I think those guys are pretty smart so there must be a good reason for it. I didn't get the chance to find out why.


  37. Jack, I wasn't suggesting you do another set of long runs with the Cell logs, just do a single, flat, steady state mile, say at 40 or 45 mph, with regen and without. You should see a difference if your theory holds true, no need to drive around pointlessly for hours, no need to even recharge between runs. Just look for more oscillations on the regen graph line, should stand out if it's there.


  38. Unless the generator is spinning above a certains speed during regen braking - it will not be effective at recapturing the energy during braking. Some kind of computer controlled variable gearing/flywheel mechanism might help recapture the energy at the proper speed that the generator needs. If braking is done suddenly such that the current level/voltage level is higher than can be safely used by the inverters, the inverter will *dump* the excess current to protect the battery system - maybe an ultra capacitor might help capture energy spikes during sudden recharging periods.

  39. "the inverter will *dump* the excess current"

    An interesting thought. If you are dumping 100 Amps into a dummy load rather than the batteries something is going to get hot quickly!

  40. I'll add another piece of data to consider. Say you are about to travel 10 miles on the highway, you accelerate to 60 miles per hour once and move a 2000 pound car through aerodynamic drag, the amount of energy used to get to 60mph is big on its own but in comparison to how much you used to travel at 60mph for 10 miles, you aren't getting much back. It's not very advantageous for the highway commute electric car like mine.

    I use a BLDC setup with a prototype controller built and paired together with a motor by a company in Iowa that didn't fare the beginning of the recession because they didn't market their product and were looking for that 'big B2B deal'. I don't use a throttle based regen because I don't want to slow down if I release the pedal because I prefer to coast when I can. I also don't want it on the brake because I don't want ANYTHING wasted through friction in the pads. I've setup a pot attached to a small lever to the left of my clutch so I have control of the regen through a pedal and mine will regen down to about 500rpm.

    With this setup there are no ups and downs in speed wasting energy when I don't want regen because I'm not making up speed as you are when you are inadvertently going through a battle of accelerate/decelerate olympics. It is always best to coast, as you've said, and that is exactly what I do but if I come about city driving, stop signs, stop lights, traffic and I don't have the patience or that driver behind you is freaking out that you are coasting on a one lane road, I'll use regen to replace ONLY what I would normally use the brakes for. If you use it in place of coasting, of course you have something to lose.

    Jack, try this. Take your regen OFF of other pedals, either use a slider pot on the shifter or, a squeeze throttle type pot, or an extra pedal. It might not seem, feel, or look like a gasoline car but this isn't a gasoline car and sometimes an extra piece must be there to get everything we can out of our setup.

    I don't get any of the same oscillating business you are going through and it doesn't take me extra power to get where I want to go, there are two options with my setup.

    1: Use brakes
    2: Do the exact same thing but use the regen pedal
    Everything else is the same and until I choose one of those options I cannot lose efficiency, there is NO WAY that brakes can be more efficient when doing it this way and otherwise driving the same.

    There is no way I'm using more power before I get to the point of wanting to use the brakes, of course coast when you can and only use the waste pedal or regen pedal when you need to.

    I really hope this helps you understand what is going on. I have used regen on and off and have made my comparisons over the same route when I originally set things up, similar to how you did it. The gain isn't that high because you really lose more to aerodynamic drag, driveline losses, and rolling resistance than you do to get up to speed so there isn't much to gather. I don't often drive in the city and have a highway commute of 40 miles. I went the brushless DC route to avoid brush wear, stuck brushes, comm corrosion, and to experiment with something others haven't and I had a company willing to hand me some components to test with. There is a gain and it is extra efficiency, you need to fix your oscillation problem to get the same gains. Either add more deadband in your throttle setup or get away from using a pedal that has no way of actuating without putting at least some pressure into the brake lines and causing you to slow down in more ways than one.

    Hope This Helps!
    Gerald Williams

  41. I'm a bit confused by all clearly I'm no expert and it's *possible* I didn't read the whole post. However, "regening" any time your foot is off the accelerator seems like a terrible idea to me. Why would you want to dump perfectly good kinetic energy back through a lossy system? You have a marvelous indicator to let you know when to do that, right? The break peddle? What am I missing?

  42. Just did a bit of reading about Dr. Nourbakhsh and his work. He started working with electric cars in 1988 and he is now particularly focused on gas to electric conversions. He started a "crowd sourcing" effort that has collected data for over 8700 actual miles driven by private individuals on american roads.

    Among other interesting observations, the data show an average regen gain in the area of 20%. And yes, it varies all over the map - literally! It seems inconceivable that in reality there is actually no gain and in fact a loss and all of these "average Joes" driving these thousands of miles are deluded into thinking regen works, when in the real world when it is actually a penalty.

    After rereading almost every post on this topic I realize that Jack and I (and some others) have been arguing 2 different points. I have been saying that regen works technically, and Jack is saying it doesn't work practically for him in his driving. I'm still not convinced, but at least I understand his position better.

  43. Dan, glad to hear you finally understand. It was frustrating to watch you arguing the wrong thing.

    I'm still eager to see if we can get more people to confirm this result by performing the test on their own cars. Are any of the AC powered EV owners going to try doing this? There have been at least two posters who have stated that they have such a vehicle. I really hope you decide to try this. You really have nothing to loose by performing the test. You either confirm Jacks result or you add conflicting information. Either way we learn something new.

  44. DAW:

    Dr. Nourbaksh work pretty much is centered on measuring AH out and AH in and in fact I've corresponded with him about it. I didn't realize there was any confusion here.

    We DO expend AH in driving teh car. We DO recapture AH from regen. It didn't disappear and we get it.

    Now, what isn't being done is do the same drive with the regen entirely shut off and compare teh net results. I'm reporting some surprising results in the comparison. It would appear that there is no gain in reality across a real drive. That is NO GAIN while you are sitting there with AH data in your hand.

    We've done this on two cars. But you're all correct. It's 5 drives. It occurs to me that they both had the same meter model, and maybe there is something in the meter.

    Tonight I'm charging the Mini. Tommorrow I intend to make the same long drive and then plug it into our Texaco Gas Pump to charge. This will give me a kWh reading from the wall and we'll IGNORE for teh moment the meter in the car.

    The following day, I'll drive it without regen. And again charge it from the Sky CHief pump.

    This will give me two new data points. The Brusa charger is pretty good, but relying on a charge voltage curve to set our "full" mark is a little squishyh. We're dirtying our data with the charger inefficiency.

    But on the other hand, I've been wondering what wall energy/expense per mile were anyway with this car. And it gives us a different take and cuts the LinkPro out of the test.

    Perhaps it IS our meter.

    Jack Rickard

  45. By the way, Gizmodo suggested the kWh measurement from the wall a couple of days ago and I dismissed it a little brusquely. My apologies. On second thought, I think its a pretty good idea.

    Jack Rickard

  46. Gerald:

    I kind of like this idea. TWO brake pedals. Why not? It's a little out of the box. When I was a kid I drove a tractor on the farm. It had a right brake and a left brake and you used them often to turn. How we wound up with one brake pedal on a car anyway?

    So what level of range extension or efficiency do YOU measure with your setup?

    Jack Rickard

  47. Gerald:

    Further, when I was in high school a number of my friends were very involved in dirt biking. These are small motorcycles that as I understood it, wound up parked in trees and creeks alot after the crash. You ride them on dirt trails in the woods around here.

    They featured a "compression release". When navigating down a STEEP hill, you could kind of brake from the back by hitting the compression relesae. This cased the motor to quit idling and instead deployed a lot of back pressure through the tranny to the rear wheel to slow you down.

    It was like a little trigger on the righ handle bar that you would hit with your index finger.

    Jack Rickard

  48. Just a thought but why not try a bench test with say a 12v pack , xantrex , psu and one of your constant current loads? Take say 1ah out (by actual calculation not the xantrex) then put back .1ah and see if the results match. I suspect the battery chemistry may well be playing a part here. I'll try and do this myself this week if i can get some time.

  49. No regen post, sorry guys :)
    Jack, the parts of telling us the practical side of the components is very helpful.
    I would like to see much more of that, these "try and see what happens next" are such a fundamantal help for converters like me.
    If you get some time during your regen-tests or you don't know what topic to talk at the following friday shows, please tell us more about the components and how they work in daily usage (like the MEA DEA heater).
    So, "stay on the ball" and many thanks :-)

  50. Jack, a compression release should do just that, release compression in the cylinder to make turning over the engine easier, and if used when running wouldn't it give less engine braking?

    Back on topic, even if I accept your oscillation theory I can see how that might come into play with regen on the A pedal, I don't see how it can be a problem with regen on the brake pedal. You may need a greater pedal travel for strong regen before the brakes come into play, or a separate pedal as was suggested, though that would be a bit cumbersome. But with regen on the brake pedal the car should drive exactly like a non regen vehicle, until you hit the pedal. With my setup I'll see 200 amps of regen on strong deceleration, what's the max you've seen?
    I do like the idea of charging so you can measure kwh from the wall to eliminate any potential issues with how the Xantrex measures things. That's how I did my measurement, though it took a while as I had to do it at 120volts through a kill a watt.


  51. JRP.

    A compression release kills the idling motor and causes a HUGE deceleration on a dirt bike. I'm not guessing. If you're not ready for it, it will throw you over the handlebars. The useful part is that it applies only to the REAR wheel, an advantage on a steep forward downhill.

    As to the braking, yes everyone can get with it on the accelerator but it is conceptually more difficult on the brake. It is actually the same thing.

    You ARE changing the drive characteristics of the car. The whole point of regen is to pre-empt the existing brakes. Since we can't make them come on "later" in the braking process, we have to have regen deploy "earlier" than the friction brakes. So much lighter brake pressure deploys regen, either by pressure transducer or by signal light. How do yours come on?

    In doing so, we have changed the drive characteristics of the car and you will drive it differently. This is an hour and a half drive and small differences in how you drive the car are very cummulative. This is why I have ZERO interest in doing a 1 mile lap and seeing what happens. I don't care what happens. I already know the regen works and produces current. What I want to know is what is the effect when driving the car in the real world on real drives. It's not that I don't understand teh theory of regenerative braking guys. The components are all more or less working. I could be measuring it incorrectly, but the batteries, motor, and controller on both cars are doing their job superbly. The question is, does regen return enough efficiency gain to be worthwhile in EV projects commensurate with the additional expense and complication it causes?

    But since I'm so "real world" I like your idea and Gizmodo's idea of measuring kWh at the wall. It has a couple of problems:

    I don't know the accuracy of the kWh meter in my Trusty Texaco Sky Chief charging station. And I have nothing to calibrate it against.

    We are dirtying our data set with charge efficiency - reducing the dynamic of any differences.

    We are relying on the charger to cutoff the charge at precisely the same point - a dubious assumption.

    The advantage of course is triangulation. We are cutting the Xantrex out of the loop and using an entirely separate and different measurement device.

    I will probably do a DIFFERENT drive this time. First, I'm tired of the existing one. And second, you don't make the SAME drive every day. Actually many of us do, but it gives us another data point.

    Jack Rickard

  52. So DON'T TOUCH THAT DIAL. Underdog will be back with MORE Underdog adventures....

    Jack Rickard

  53. Jack,

    I like you dual brake pedal idea. I think that will provide more pure/confident regen data.

    I like the "fill'er up" approach too. It of course depends how squishy the "full" indication is. The complementary approach to consider is to drain the rest of the "gas" at the end of the run and see what is left. I would suggest using a constant (controlled) discharge of maybe 10 amps or more and measure the energy removed to empty. I know the "empty" indication is squishy too, and has the added issue of working close to the edge of battery damage. As you plan to do on the fill up approach, you really should measure WH to empty.

    You have a fun job!


  54. For once I agree with you Dan. Not on draining the cells until they are empty, that's even more squishy than full. On the part about me having a fun job.

    We are having a LOT of fun, and we're learning a lot. How much better does it get? Tell you what, since most of you can't give up your life and go do it too, I'll share.

    And here's a little tidbit. If you DO spend nearly a week making a little test set and you DO wire your entire battery pack with fire causing spaghetti, DON"T go drink whiskey and have a great idea to speed things up by charging at 42 amps through that test sets 16 AWG wiring and set your test set on fire - as well as some of the car wiring. Some of the briliant midnight ideas I get.....sheesh...

    I'm sure I'll get the smoke cleared by this morning's drive....


  55. Jack,

    I don't change my driving characteristics, I don't have as much regen amperage capability as you do with your AC setup and I currently actually have it set to a 10kw limit, which is about half of what it can do, because I have my wife driving the car sometimes when she goes into the city and she likes to use the regen but she doesn't like how it feels with too much power and it's easier to ramp the power when it is set with 10kw at full travel with a linear regen amp curve but the car is light and 10kw will get you stopped at a plenty acceptable rate. When it pulls more power back into the batteries than that brushless DC motors get a bit noisy or at least this Groma does. On a side note, the AC motor you have, I cannot hear, brushless DC I can if I'm pushing more than 50% amperage over what I normally use at a steady speed, most lighter hills I don't hear it and not while cruising but step it down all the way when below 3000rpm and a fairly sporty but mostly pleasant sound that is similar to how the Honda hybrids sound with their BLDC motors but those are nearly silent since they push 100amps, mine makes about twice the sound.

    To get back on topic, when driving in the city where I might travel a block or two and stop at a light, I probably gain about 20% of my energy back in the in/out if I'm 'being nice to traffic' and doing what they do which provides for more required stopping power, but if I coast as much as I can but not in a way that takes me forever to get somewhere and travel at lower speeds and regen when getting closer to the light, I'm getting roughly 10%-15% better. ...but this is stopping at every other light. I've done the comparison on and off on my normal commute crossing one side of the city to the other, in the end it is really at 10% when it is all averages when in the city because I can time the lights fairly well and coast and there is part of the drive that is in town and on the highway so the gains aren't as much.

    The plot thickens however, my commute changed, I no longer work in the city, I commute from my suburb to another suburb using the highway, I usually go 60mph or whatever the speed of traffic is. I'm losing most of my energy through wind resistance traveling at that speed. I have about 3 stop lights and a stop sign on average to stop at and the off ramp where I try to coast as much as traffic will allow and then swing around the clover and regen from 30 to 0. Do I have many opportunities for regen, no, I used to but not anymore. What do I gain now that I'm driving on the highway and barely ever stop? About 200wh out of the whole one-way trip, so per charge I could probably get 4 miles if I drove it to work and back twice, a whopping 800wh out of the 20kwh that I'd use for two trips, in reality a savings of under 5% which isn't even saving a dime a day for me. I get an extra mile. I don't have hills that would require me to use any method of slowing down either.

    I'm not upset about it, I got my motor, controller, and battery-vco(on a serial comm) setup as a business deal because I provided that company with many connections to get things done so I didn't pay extra like others could. I actually like the AC50 setup and my wife likes the idea of having a sliding pot on the stick shift to control the regen so if I did it again for her, I'd probably go that route.

    To put it simply, my drive, an extra 3 miles out of an 80 mile drive. My wife, since she doesn't like to bug other drivers and coast more than they do, she could probably get 8 extra miles out of it or about 10%. ...but at a lower efficiency than just driving efficiently, including the drawbacks that come with being efficient.

    Gerald Williams

  56. Gerald,

    Correct me if I am wrong, but it seems like you are also comparing the drive to itself and not to a drive without regen.

    Is it possible for you to perform the test as outlined by Jack? Can you run a loop of 25-50 miles that you believe has "a lot" of regen. Calculate wH or AH used and then take the exact same drive at the same speeds with the regen turned off. Calculate the wH or AH used and then provide us with a comparison.

    I really am interested in seeing a 3rd party perform this test.

  57. Jack, I'm glad to hear you are going to use the "gas" pump to fill-er up and measure KWhrs. I saw a Kill-a-Watt type meter at westlakes for about 30.00 dollars. You just plug it in like a small extension cord. You could plug this in series with the Sky Chief pump to get two KWhr readings for comparison. I say the more data the better.

    All the best JMS

  58. "The whole point of regen is to pre-empt the existing brakes. Since we can't make them come on "later" in the braking process, we have to have regen deploy "earlier" than the friction brakes. So much lighter brake pressure deploys regen, either by pressure transducer or by signal light. How do yours come on?"

    My regen is only on the A pedal, and set to give me about a max of 200 amps. With your regen on a pressure transducer even low pressure is also applying the brake pads, which reduces your regen potential. That's why I asked what the max regen you saw on the brake pedal regen, specifically on the AC50, not the Mini since it's a higher voltage system.
    Other than a panic stop I don't see how braking characteristics is really going to change the dynamics of the drive. Power brakes, manual brakes, regen, a parachute, whatever, if you stop in the same distance, after you stop it's the same energy to get you moving again. Only one of those choices actually puts some energy back into the system. Something isn't adding up.

    Looking at some of my cell log 8 graphs and zooming in on 10 second sections I saw no oscillations more pronounced than the resolution of the image, so if it's there I didn't pick it up.

    The PFC end of charge is "squishy", but if temps are close it should be fairly accurate. You might want to monitor amperage and voltage and stop the cycle manually at the same level each time to be sure.


  59. Food for thought...

    In 2008 the EPA revamped their rules to provide fuel mileage estimates that more closely reflect "real world" driving, and the "real world" was hard on all vehicle mileage stickers, especially hybrids. Under the new rules the Prius is rated at 48 city and 45 highway. To my knowledge this is the first production car to achieve better mileage in the city than the highway. Is this the result of real world regen working?

    I have to give my buddy Tom credit for beating this tidbit into my thick scull. If you drive a plugin electric vehicle, why would you even CARE if regen works. You are already enjoying a weeks worth of commuting for the change in your couch! How long will it take you to pay for all the fancy regen gear at $0.03 per mile GAINED!! The one caveat is that if you are range limited or have range anxiety regen might be soothing to the soul. Of course if you are major "greenie" you will sleep better if you save a kilowatt (or AH).

    I still say regen works, but I'm not sure I care anymore!


  60. "Is this the result of real world regen working?".
    No, the electric motor stops working at speeds above 40 mph in the Prius. The only time you are under electric power is in the city.

  61. Jack said: "I'm sure I'll get the smoke cleared by this morning's drive...."

    is a 24/7 webcam in your shop an idea?

    I know I'm a lazy reader, but am I correct you have some second thoughts now about the accuracy of the xantrax in case of quick current direction switching?

    Regards, Jan

  62. "No, the electric motor stops working at speeds above 40 mph in the Prius."

    The problem with this argument is that 100% of the energy needed to move the Prius comes from gasoline, so at ANY speed gas is ultimately the supply. The batteries are used only to store some of the energy the ICE produces (the argument is very different for PEV). Any gains must come from either reducing energy used or recapturing some that would be otherwise lost. City driving has the advantage (over highway) of less wind resistance, but this has historically been more than negated by the inefficiency of frequent stops (heating up your brakes) and periods of idling (all energy produced during idling is lost). The hybrid makes up the ground by recapturing some of the energy using regen braking and reducing consumption by shutting down the ICE when stopped. I don't know the percentage contribution of each, but I am confident a significant portion of the gain is from regen.

    Does anyone how to find the EPA driving profile they use to evaluate gas mileage?


  63. Hmmm.. Didn't realize how far behind we are here. Gents ALL electtic cars and I would assume all hybrids get better mileage in the city than they do on highway driving. And it is usually MUCH greater difference for an electric than listed in the Prius numbers.

    The answer is very simple and quite universal. Wind resistance. Most people simply do not accord it the value it has. Try this non-thought experiment. Remove your windshield and go drive 70 mph.

    This was why the national 55mph speed limit during the "gas crises" of a decade or so ago. I think Jimmie Carter....

    So why do most cars have BETTER mileage on the highway than they do in the city? It has to do with how truly horrible the efficiency of the ICE engine is. As bad as it is, it is a little better at a steady RPM and truly a mess when RPM and power loads vary. It completely inverts the MPG numbers and in fact led to having two numbers - in town and on the freeway.

    If you remove the drive train factor, all cars take much more energy to move at 70 mph than at 35. The Wind resistance is pretty negligable up to about 35-40. After that it is a squared function.

    My cars will do 100 miles on a charge and they will do 100 mph. The one thing they will NOT do is both at the same time. I would guesstimate 100 miles at 35-45 and 60 miles at 70 and 35 miles at 85.

    Jack Rickard

  64. "My regen is only on the A pedal, and set to give me about a max of 200 amps. With your regen on a pressure transducer even low pressure is also applying the brake pads, which reduces your regen potential. That's why I asked what the max regen you saw on the brake pedal regen, specifically on the AC50, not the Mini since it's a higher voltage system."

    JRP. This is a very bad habit you picked up on that DIY forum mess. If you don't HAVE it how do you know so much ABOUT it? We're not putting any pressure on the pads on light pressure on the brakes. But we maximuze the regen in this zone. As you step through the regen, you do indeed get friction brake. This is MORE exacerbated on the Mini as we use the brake signal light and the regen BUILDS over about 2 seconds and it takes a LOT of pressure to get any friction brakes at all.

    So however we have it, we have it two different ways. And in any case, you have it on your accelerator only. You make intuitive leaps in logic based on sentence fragments you excise out of context - the DIY Forum hallmark.

    As to the oscillations, I was describing something a bit more broad than an 8 or 10 second segment. Slow accells and decels across a 5 mile level drive was what I described. How many of those would show up in a 10 second segment? Picture six or eight of those in the 6 minutes it would take to make the drive. And they would be pretty gentle changes in cell voltage.

    Jack Rickard

  65. "is a 24/7 webcam in your shop an idea?

    I know I'm a lazy reader, but am I correct you have some second thoughts now about the accuracy of the xantrax in case of quick current direction switching?"

    The webcam is an idea. What would be its purpose?

    I'm not having second thoughts about the Xantrax in case of quick current direction switching. It samples hundreds of times per second and I cannot imagine that being a factor.

    But yes, the one thing in common between the two vehicles is the Xantrax and so if all of this is just as impossible as you all claim, I have to look there.

    Jack Rickard

  66. Jack, can you clear something up for me about measuring Ahs? I can't quite convince myself if an Ah In equals a little more energy than an Ah Out or not. For example, if driving the Mini at 40A and 375V, then that's 15,000 W. But if you're regenning at the same 40A, the pack voltage will be a little higher than 375V. So is the same measured 40A putting more energy back into the pack than it took out?



  67. Jack said: "The webcam is an idea. What would be its purpose?"

    For our entertainment, of course!

    Regards, Jan

  68. Bill,

    I understand the voltage differences between discharging (using it) and charging are the battery losses.

    Regards, Jan

  69. Jack/Bill,

    "But if you're regenning at the same 40A, the pack voltage will be a little higher than 375V. So is the same measured 40A putting more energy back into the pack than it took out?"

    That has been my concern and point about AH vs WH this whole time and why I asked for voltage readings during steady state, hard accel, and hard regen. This is how my though process goes in concept. You drive along at steady state and the battery voltage is Vs, your current is As, so your power draw is Vs * As = Ps. You accel hard and the voltage drops to Va, current goes up to Aa, with a resulting power of Pa. You decel hard and regen raises the battery voltage yielding Vd * Ad = Pd. When you just look at Amps you miss the fact that during accel your power draw is overstated because the voltage dropped and when you regen your power (energy recovery) is UNDERSTATED because the voltage increased. For run to run comparison purposes the accel and steady state power draw numbers don't really matter even though they are inaccurate because you have the same conditions with and without regen. The same cannot be said about the regen cycles because now you are making a fundamental change between runs and inaccuracy in readings on the regen run will give you inaccurate comparison to the run without regen. I hope I have explained this well enough. If you van provide voltage reading during the 3 conditions of driving we can calculate the impact of the inaccuracy. Note that as battery SOC changes you have more inaccuracies (as you aptly point out). But unless you can show they are trivial, they must be dealt with.

    To me it is clear that we need to look at power, not current.

  70. Figuratively, the "value" of an AMP is different in these 3 phases of driving, so if we add them/ subtract them we are mixing fruits! Power calculation eliminate this problem.

  71. "Jack, can you clear something up for me about measuring Ahs? I can't quite convince myself if an Ah In equals a little more energy than an Ah Out or not. For example, if driving the Mini at 40A and 375V, then that's 15,000 W. But if you're regenning at the same 40A, the pack voltage will be a little higher than 375V. So is the same measured 40A putting more energy back into the pack than it took out?"

    Ok. I'll try.

    When you discharge the cell, the voltage at the cell terminals will decrease in proportion to the amount of current you take out of the cell. We CALL this "internal resistance" although it isn't really resistance in the normal conception of such things. For example, if my pack is 375 volts open circuit, and I put a 100 amp load on it. Let's say it decreases to 370 volts. If we divide the 5 volt difference by the 100 amp load we see that we have an internal resistance of 0.05 ohms. These cells are more like 1 milliohm, but you see what I mean.

    Similarly when charging a cell, we have an open circuit cell terminal voltage and this varies by SOC. Our pack at a low state of charge might be 366 v and at a full state it might be 375v. As it so happens, at 3.35v per cell, which is what I typically see in fully charged cells, the Mini pack is 375.2 volts.

    At any SOC, in order to reverse the process and "charge" the cell, we have to force current into backwards. The amount of charge is strictly a function of current, the quantity of electrons forced onto the anode and back into the crystalline structure of the carbon on the anode, to recombine with a lithium ion and form a whole lithium atom. The deintercalation process requires the release of two lithium atoms from the crystalline structure and this requires 2 electrons to accomplish - the pairs are about the way lithium ions share electrons with the carbon in the crystalline structure.

    So the only "recharge" we accomplish is by forcing electrons onto the anode current collector and INTO the crystalline structure formed from carbon and lithium ions. If we "count" the electrons, we would normally do this in a quantity termed a coulomb. And that count is by definition current. If we count the number of amperes of current over time, we have ampere hours - 1 amp for 1 hour being an ampere hour AH.

  72. In order to force ANY current onto the anode, we have to apply a voltage potential LARGER than the open circuit terminal voltage of the cell. The difference in potential defines the level of current.

    You won't actually see the applied voltage if you measure it because it is virtually all being dropped across the current into the cell. So you could apply 10 volts or 12 volts or 100 volts to a cell with an open circuit voltage of 3.1 volts say, and you would still read 3.1 volts. As the cell takes on charge, this would slowly rise - based on the rate of current into the cell, which is in turn a result of difference in potential between the applied voltage and the terminal voltage.

    The bottom line is that indeed, you will need a higher voltage to create current into the cell. But this is part of the losses cummulative to produce the AH of current into the cell. Similarly, you have losses in the cell to produce the AH OUT of the cell. They heat up.

    Because the voltage is variable, both in and out, we normally just count the AH. Total power is of course E X I. But in talking about regen, we normally by convention talk about the number of AH out and the number of AH in - the result. And the losses are the losses. You and Dan want to count them AGAIN.

    Estimates of regen efficiency are BASED on AH counts. So many AH out , so many AH recovered. Using this metric, I don't see any 21% efficiency gain here. But I DID change the game. Instead of measuring AH out and AH in and comparing them, I sum the AH out and the AH in to get a NET AH per drive end to end. When I compare the NET, between two drives, I don't get 21%. I don't get any percent.

    Changing the measurement won't get the energy back. The losses are the losses. That you need more voltage to produce the AH doesn't matter, and in fact it varies, depending on SOC and what you have available. I can put AH back in at 1 AH or 100 amps per hour.

    So changing to kWh is a useless exercise and explains nothing. Now we are measuring apples, and the rest of the world is measuring oranges that already include the losses.

    The only reason to change to kWh is to use a different meter that I have available at the gas pump charging station. It adds a lot of other things as well that tend to mask the results. But ultimately, we are interested in the cost of energy and the amount of energy we are billed from at the wall. So I'm not averse to the concept. If anything, it is MORE "end to end" and "real world."

    Jack Rickard

  73. One final note, more on charging. We use CC/CV charge curves to charge the cells to an apparent terminal voltage defined by the manufacturer. This voltage is NOT the open circuit terminal voltage of the cell. It's the charge voltage.

    When we disconnect the charger, we have a large number of electrons that have been forced onto the anode current collector, but not reabsorbed. This is commonly referred to as surface charge. As they are fully reabsorbed into the crystalline structure, this surface charge dissipates and the cell gradually returns to it's fully charged potential. This is nominally 3.4 volts in LiFePo4 cells and varies in batteries depending on their cell chemistry. Lead acid cells have a voltage of 2.1 per cell. And LiFePo4 cells have a cell voltage of 3.4. We'll commonly see 3.35 volts on an LiFePo4 cell that we term fully charged. This is because we do NOT want to overcharge LiFePo4 cells, and we commonly DO overcharge Pb cells.

    Jack Rickard

  74. You are still getting caught up in comparing a run with regen to itself. From what I saw in the video, the voltage when discharging was about 119v and when regenerating it was about 121. That is only about 2%.

    I thought the same thing myself when I first saw the original video, but it does not explain why both runs had the same number of amp-hours. If you were recapturing amp-hours (or watt-hours) you should have had less in the run with regen. We know we were capturing amp-hours because we can see it on the meter. Given that the NET AH was the same for both runs, then the car with regen must have used more AH to accelerate and go steady speed. In fact the amount of extra for acceleration and steady speed was equal to the amount captured by the regen when you slow down.

    I posted on the original blog this example of how it can't be because of wH. In order for you to use less wH and the same AH the pack voltage must drop by the percent of benefit. If you have a 7% benefit from regen then the average (over the course of the run) pack voltage must drop by 7%. Since there is no way we saw the pack voltage drop in any significant amount, then the difference in wH could not have been the answer.

    ---below was copied from old post---

    The original run was 78*120=9360wH. For conversation purposes, if the second run had a 15% increase (use any number you like, it all works the same) then you are saying that the car actually took only 7956wH. If this is true then the car would have to had an average pack voltage over the course of the trip of 102v (78AH*102=7956). There is just no way that was true.

  75. It looks like i posted the same time as Jacks first post. I got an error, and then my post arrived after jack finished. Sorry for duplicating information you posted Jack. Thanks for your explanation, it is much better than mine.

  76. "The answer is very simple and quite universal. Wind resistance. Most people simply do not accord it the value it has. Try this non-thought experiment. Remove your windshield and go drive 70 mph"

    So true - The vind resistance scales with the speed squared, but the power we use to maintain the speed scales with the power of 3 ! That is the main reason for me to drive at 50 mph than 55 mph even though the general speed limit here in Denmark is 50 mph. And highway speeds 68 mph/82 mph are completely insane ;-)


  77. Hi Jack
    I just wanted to comment that it seems most everyone is missing entirely what you're actually saying, in both this post, and the previous one - I 'get' what you are saying

    You're not measuring/noting "Amps" used, as so many keep mistakenly saying, but AmpHOURS
    While yes they seem, based on the name, to be the same, they arent
    Amps, is the current flow, but AmpHOURS is storage capacity
    For those that seem confused on the concept

    AmpHours is Storage Capacity
    "the electric charge transferred by a steady current of one ampere for one hour" (irrespective of voltage)
    "The ampere-hour is frequently used in measurements of electrochemical systems such as electroplating and electrical batteries."
    WattHours is Consumption
    "Energy in watt hours is the multiplication of power in watts and time in hours"
    Jack is measuring how much of his total storage he used, the amount of [K]WH used to feed into that storage tank, is irrelevant

    - Think of AmpHours as Gallons of fuel, and Amps as the flow rate

    An analogy using a Petrol (or as you USA people call it - Gasoline) car
    Regenerative Braking, is being touted as a 'fuel saving device' - what Jack has done is done a drive WITH the device, and measured the end result of how many gallons of fuel his car used
    Then he did the drive again, with the fuel saving device disabled/bypassed, and again measured the end result of how many gallons of fuel his car used
    What he found, is that the supposed 'fuel saver' was actually causing him to use more fuel *OVERALL*
    ok - sure, during the drive with the 'fuel saver' (regen) on, some fuel (AMPS/current) was flowing back into the tank - but the overall fuel consumption appears to be MORE *with* the 'fuel saver' (regen) enabled, than without it
    - in 2 different vehicles no less!
    What he's been asking for, is for others with the same 'fuel saver' (regen) system - is to try the same experiment, one long (50km/30mile+) drive, with the 'fuel saver' enabled as per normal/usual - and one exact same drive (or the exact same route anyway - the specific drive will be *slightly* different, just due natural variances in the traffic flow, hence the longish distance being required) but this time *COMPLETELY disabling*/bypassing the 'fuel-saver' (regen) - not just 'driving around it'

    whew, sorry about that, that was a lot longer than I expected it would be

    An Avid Watcher (but unfortunately, not in a position to build or buy my own electric car)

  78. "We're not putting any pressure on the pads on light pressure on the brakes."

    Please explain to me how light pressure on the brakes, which triggers a brake pressure transducer and thus must increase brake line pressure to work, does not transfer the brake fluid pressure to the pads.

    Regardless, how ever it may work, it still does not address the fact that a car with regen on the brake pedal only cannot somehow be less efficient when not braking, and you have not explained how it's less efficient when sending energy back into the pack as opposed to sending that energy into the air as heat. Do you not see the problem here?


  79. Found this on the Aptera forum:

    So one of the main reasons for moving the drive on the Aptera forward is the implimentation of regenerative braking. Turns out even on "heavy" cars This is may not be worth anything. I already know for a fact regen on a moped adds nothing to the range, using AC hubmotors. It does damage stability however. The moped back wheel wants to lock up. I took it as - the bike is too light for the energy recovered to be make a difference. The Aptera is much heavier at 1800# estimated.

    So, I found a very thought provoking series of tests on Jack Rickard's blog.

    Unlike most of us he is actually building cars and testing them. I can't dispute his findings.

    They lead me to question the "Dogma of regen".

    Ya got to love it! I no longer believe regen to be worth a damn.

    Won't change that view now, until I see compelling hard comparative data.

    How much easier is it to engineer a DC -no regen solution - and cheaper!

    Copied and Pasted from

    I'm searching for real world test results for verification of results.

    Regards JMS

  80. "Regardless, how ever it may work, it still does not address the fact that a car with regen on the brake pedal only cannot somehow be less efficient when not braking, and you have not explained how it's less efficient when sending energy back into the pack as opposed to sending that energy into the air as heat. Do you not see the problem here?"

    In the first place JRP, it's not really incumbent on me to EXPLAIN anything to you. I have made some tests and reported some results. And I've drawn some conclusions. IN an effort to accommodate your sniping, I have offered several theories, that would indeed account for all the observations. I don't think you even read them.

    The most cogent is that regen changes the drive characteristics of the car, and consequently how you drive it, to be LESS efficient overall, while recapturing some energy, apparently insufficient to get you back to where you were with no regen at all.

    Could be the meter. Could be the weather. Could indeed be what's playing a the movies. And I of course could be wrong.

    Jack Rickard

  81. Jack,
    I fully believe that you have faithfully reported your results. I even believe your explanation that "...regen changes the drive characteristics of the car, and consequently how you drive it, to be LESS efficient overall..." This makes sense to me.

    But at the same time I am a true believer in regen. Not in the religious sense. I understand how regen works, and in a properly configured system it should be possible to get an efficiency improvement. The decline in efficiency you are seeing implies there is something wrong with your configuration/implementation.

    In the discussion that follows I will give my honest understanding of your position. If I am wrong it is not because I haven't read what you have said (above). In fact I have read it several times.

    I think you have hit the nail on the head with your explanation of the problem with "50% of max regen neural braking off the accelerator" mode. A situation where you are either braking or accelerating is obviously not good. You need some workable method of coasting too. So to make this implementation workable you need a button (or something) which can easily be used to turn off regen when you want to coast.

    I think I understand your explanation of the effect in "regenerative braking on the brakes only" mode. But it seems to me there must be a problem with implementation here also. I think that in an effort to get substantial regen effort in, before the friction brakes begin to have substantial effect, you have made the regen activation too sensitive. And this leads to the oscillation problem you note. If you are happy with this then there are two easy solutions which are not exclusive.

    1. Reduce the amount of regen produced by the brake pedal.

    2. Use some other way (another "brake" pedal) to control regen.

    I'm sure you are well able to come up with some simple mods to the Speedster-duh though obviously not till its on the road again. I don't know enough about the mini to comment on how easy it would be to change, but i'd be surprised if it was very difficult.

    Don't give up on regen just yet. I'm quite sure we can get past these teething problems. And keep up the good work


  82. I can't see how you could ever get 10% or even 5 % regen, you dont use you brakes that much!

    How many people would do 1km of hard/continuous braking in a 100km drive?? I doubt very few.....thats only 1% regen if you could recapture it at 100% effiency!

  83. Wow, extra pedals and switches, altered driving style. Who knew regen was so much work?

  84. Yup, that's the net of it. It isn't a good ratio to equate 1% of hard braking to 1% of a 100km drive, but the underlying point is vital, and now we're finally getting down to being able to say it correctly:

    There just isn't that much braking energy available from which to regenerate charge energy.

    Jack's breakthrough is that this fact is compounded when compared to normal (non-regenerative) braking, where there's apparently a synergistic effect that actually boosts efficiency when the electric drive is only used for acceleration and not also for braking.

    Consider another data point about how little brakes can be used: Until effective disc brakes arrived at the Indy 500 on Jimmy Clark's Lotus 25 in 1963, the conventional wisdom was that the brakes on an Indy car were good for just two or three hard stops during the race. They were only used to bring the car to rest in the pits, and held completely in reserve on the track. The cars demanded to be driven that way.

    Jack's Mini is a little like that- the brakes are hard to unsatisfying to use, so the driver does the job largely without them.

    This reparte is all getting very interesting around the paddock, but the horse is long out of the barn: regen isn't worth the effort or expense for most driving.

    Quite astonishing, really, given all that investment in it. That's what happens when OEMs dabble in fits and starts for years on something without ever committing to really understanding it...

    Thanks, Jack.


  85. Jack, I read and listen to all your explanations, just not sure I agree with them yet. Not sure you're 100% convinced by them yet either.


  86. Jack,

    I hope your recovery from the flame out is proceeding well. We anxiously await new data from your test using a different sensing technique. In the mean time I would like to review the bidding.
    1) You made multiple test drives, attempting to make them as identical as possible for valid comparisons
    2) You observed recovered current during regen braking events on the meter and felt the effect of it (slowing the vehicles)
    3) Your AH data gathering device shows the drives with regen used the most energy.
    Unfortunately these three occurring simultaneously is a logical impossibility. Item 2 is direct observation so we're pretty confident in it. You suspect item 1 is the problem; the drives must not be identical. For me, (and probably others) I trust your driving technique and think it is more likely an issue with item 3. While I am convinced the problem lies here, I am not sure if I can account for it with the AH vs WH concerns (I'm still working on that).

    Thanks for sharing the results of your research with the world! Without a doubt your conclusions will influence PEV conversions. No pressure!


    BTW - Are you taking a DC3 to the Oshkosh gathering?

  87. Jack,

    I did the calculations below based on the batteries you have in Speedster #2. What it illustrates is the difference between amps and watts at 100 amp draw or regen. I calculated the voltage drop/rise at 100 amps using internal R per cell of .9mOhm. The calculations show a 5.4% difference in power at 100 amps in vs 100 amps out. And the result is linear so at 200 amps it's 10.8% different. I don't know if this accounts for all of the discrepancy, but I think it is part of it.

    What say you?


    Speedster #2
    Motor AC-50 HPEV
    Controller 1238-7501 curtis 3 ph
    batteries - capacity 180 AH Calb
    batteries - voltage 120 VDC
    volts per cell 3.4 volts
    Cell in series 35 cells
    internal R est. 0.0009 Ohm
    Current In/Out 100 Amps
    drop 3.15 volts
    V at 100 A draw 116.85 volts
    V at 100 A regen 123.15 volts
    W at 100 A draw 11685 watts
    W at 100 A regen 12315 watts
    difference 630 watts
    difference 5.4% %

  88. One more thought on the "fill'er up" or "drain'er" out method of detecting energy left in the battery. The discharge graph of the SkyEnergy 180AH shows a dramatic drop in voltage vs AH when you get below 2.8 volts, while charging is asymptotic to 3.2 volts. There is only about 5 AH from 2.8V to 2.6V, while there is 160AH from 3V to 3.2V. All this to say the discharge transition is much (32 times?) sharper than the charge. You know much more than I about the danger zone for discharge, but if you dare get down to 2.8V or even 2.6V I think the draining to empty should be less "squishy" than the fill to full.

    Here is the link to the discharge graph I for your pack.



  89. Jack, I have greatly enjoyed your work and videos, and I have watched them all from the beginning. I really appreciate your approach of actually "doing" and "measuring" instead of just "talking about" something. But as an Electrical Engineer with 30+ years of experience, I decided I should comment on this regen topic. I think the purpose of regen is to reduce the amount of energy (electrical in this case) used to get from point A to point B. As you know, for electrical energy, energy is usually measured in watt hours or kWhs, and this is the unit that the power company uses to bill us. As you also know, 100A at 10V for 1 hour would be 1000Wh or 1kWh. My problem is that when we only measure Amp Hours (AH), we are not measuring energy unless the voltage is known (and constant). Since the batteries are know to drop in voltage under load, and rise in voltage when being charged, the voltage will not be constant.

    Perhaps a conceptual example might help. Suppose we had an ideal voltage source of 100v (D.C.) connected to a lossless DC-DC converter which output 10v, connected to a .1 ohm resistive load. Suppose we measure this circuit for 1 hour. The current flowing into the load will be V/R = 10v / .1 ohm = 100A. The power being dissipated will be 10v x 100A = 1000W. The current flowing into the DC-DC converter will be 1000W / 100V = 10A. The energy dissipated over the 1 hour will be 10v x 100A x 1 hour = 1000Wh. If we measure the AH just before the load, we will measure 100A x 1 hour = 100AH. If we measure the AH just before the DC-DC converter, we will measure 10A x 1 hour = 10AH. Clearly the voltage makes a difference.

    If when being discharged the battery pack voltage fell to 10v while still putting out 100A for 6 minutes, but rose to 100v while putting in 10A for 6 minutes from regen, then the amount of energy out (10v * 100A * .1h = 100Wh) would be the same as the energy in (100v * 10A * .1h = 100Wh) even though the AH in (100A * .1h = 10Ah) DOES NOT equal the AH out (10A * .1h = 1Ah). Real batteries probably do not have this magnitude of voltage difference, but as you have demonstrated in your most recent video, the battery voltage does vary quite a bit under load and charging.

    In summary, measuring AH is clearly very important when working with these batteries since the specifications and characteristics are given in AH. But when considering the benefit of regen, I think you really need to measure the amount of energy used on the trip, and just measuring AHs in/out of the battery pack does not give you that information.


  90. (Part I)


    I have greatly enjoyed your work and videos, and I have watched them from the first one on. I especially appreciate your approach of "doing" and "measuring" rather than just talking about a topic. I have learned a lot!

    But as an Electrical Engineer with 30+ years of experience, I thought I should comment on this topic of regenerative braking (regen). I think the purpose of regen on an EV is to reduce the amount of (electrial) energy used to travel from point A to point B. (This should also increase range.) As you know, electrical energy (current x voltage x time) is typically measured in Watt Hours (Wh) or kiloWatt Hours (kWh) (where 1A x 1v x 1Hr = 1Wh), and we are billed by our power company in units of kWh. For example, If a circuit provides 100A at 10v for 1 hour, then it will have provided 100A x 10v x 1hr = 1000Wh = 1kWh of energy. If we only measure amp hours (AH), then we are not measuring energy unless the voltage is known (and constant). Perhaps an example would help.

    Suppose we have a circuit with an ideal voltage source providing 100v (D.C.), connected to a lossless DC-DC converter providing 10v (D.C.), connected to a .1 ohm resistive load. The current going into the .1 ohm resistive load will be v/r = 10v / .1 ohm = 100A. The power being dissipated will be 10v x 100A = 1000W. The current going into the DC-DC converter will be 1000W / 100v = 10A. If we measure the energy dissipated in 1 hour, it will be 10v x 100A x 1h = 1000Wh. That is also how much energy the ideal voltage source provided. If we measure the AH just before the .1 ohm load, it will be 100A x 1h = 100AH. If we measure the AH just before the DC-DC converter, it will be 10A x 1h = 10AH. 100AH does not equal 10AH because the voltage matters when measuring energy.

  91. (Sorry for duplicate comments. I received an error when attempting to post the first one, but it seems it was posted anyway. I did not realize that until I tried to reconstruct the comment and break it up into smaller pieces.)


  92. Dan and Steve. There is a greater voltage required to charge batteries, but the batteries don't benefit from it. You have to have it to force the current. And we measure that in AH. Now if you want to break it out, you're measuring that power AGAIN.

    It's all true, but it wouldn't account for any of this.

    We're kind of going over it now, but we've changed things a bit and gotten quite a different result.

    In my mind, the URBAN part of the drive was about half. That's because it takes about half the time. As it turns out, it comprised 11 of the 48 miles. The rest were more freeway like (about another 10 or 11) and rolling country blacktop county road (26 miles}. PLenty of hills and curves on the county road, but really one of the reasons I like to drive that is the Porsche handles it so well. That' s probably not good for regen, and probably IS good for free coasting.

    We did a very different drive out US 61 through Cape to Jackson, the urban part, and then the urban part back. Though 21 miles, it takes almost the same amount of time as the 48 mile drive we were doing. And we're getting very different results on both the kWh meter and the AH meter - and showing really quite impressive gains in fact.

    So I'm probably going to have to fall on my sword on this one. I was measuring the percentage of road style based on time instead of distance, and probably mischaracterized the drive pretty badly.

    The latest results are based on precisely TWO drives, so it's still pretty preliminary. I'm learnign that this stuff is quite a bit more variable than I had thought. So before going into further detail, we're going to do a series of drives in the MIni Cooper.

    The transaxle of the Speedster has been sent to California for rebuild. It should be back toward the end of next week. Then we will repeat a greater SERIES of tests on the Speedster.

    But from what I'm seeing, I would predict if we stay in stop and go, which is hard work actually in this heat, we're going to see a GAIN from regen.

    Jack Rickard

  93. Dan:

    Actually we're taking TWO DC-3s to Oshkosh, N25641 and N341A. We're leaving for a DC-3 gathering in Illinois on Saturday the 24th and we'll arrive at Oshkosh on Monday.

    Jack Rickard

  94. Quite an unexpected turn of events. Will the next episode dedicated to this (likely to expected) change in results?

    Regards Jan

  95. Interesting development on regen! I'd like to throw another thought out regarding WH. Sorry to keep on but it is important to this subject. Forget about the battery for the moment and just consider the motor/generator. The power flow through it is volts time current, regardless of which way the current is flowing. Conservation of energy requires all that energy go somewhere. When acting as a motor all of it goes to the wheels except for losses. When in generator mode, all of it must go to the batteries except again for losses. Since we know that the voltage varies significantly it must be included for accurate use and recovery to be calculated.

    I enjoyed the battery discharge experiment. These batteries are getting to be pretty amazing. My belief is the advances being made in batteries will bring electric vehicles to common use mire quickly than expected. We shall see. I hope to see you in Oshkosh.



  96. "But from what I'm seeing, I would predict if we stay in stop and go, which is hard work actually in this heat, we're going to see a GAIN from regen."

    Who could have predicted that?

    Sorry, couldn't resist. This is why I wasn't rushing out to buy or rent a controller to do my own testing. I figured at some point something would turn up to change your mind.


  97. Before a conclusion is drawn, I suggest that you get others to drive the same cars and the same route. If possible also drive the same route in a Tesla Roadster which have quite strong regen.

    What you have proven by now is that driving style affects range more than regenerative braking.

    I just don't understand why you become a more aggressive driver with regen than without.

    Tom R Simenstad
    Th!nk City EV driver since July 2003

    PS. sorry that my first comment ended up under the wrong heading. TRS

  98. Seems like nobody has answered Simon's questions of July 12 yet:
    "do the motor speed controllers reconvert 3 phase AC into DC during the regen process? does it have some sort of retification circuit before it back feeds into the batteries?"
    Yes to both questions.

    An advantage of AC is that the speed controller can double up as a charger. Two EVs I positively know do this is Tesla Roadster and the new Th!nk.

    Tom R Simenstad
    Th!nk City EV driver since July 2003

  99. "TWO brake pedals. Why not? It's a little out of the box."

    Car manufactures will probably follow the established standard. There are enough old folks stepping on the wrong pedal already!

    European big rigs have an electric/hydraulic brake that is activated by a lever on the right hand side of the steering wheel. That may be a standard worth copying.

    For me the two stage, brake light switch activated, regen on my EV works OK. Perhaps Th!nk should have kept it that way instead of changing it to gas pedal activated like other manufactureres. That said I have no problem adapting to gas pedal activated regen, either.

    Tom R Simenstad
    Th!nk City EV driver since June 2003

  100. "To my knowledge this is the first production car to achieve better mileage in the city than the highway."
    All EVs that I have driven, both with and without regen, get better mileage in the city (0-30 mph) than on the highway (50 mph) with my gentle driving style.

    Aggressive drivers (hard acceleation and braking) will get better mileage on the highway than in the city, even with EVs.

  101. "Please explain to me how light pressure on the brakes, which triggers a brake pressure transducer and thus must increase brake line pressure to work, does not transfer the brake fluid pressure to the pads."
    There is a considerable amount of friction in the brake piston-to-caliper housing seal and between the pad and the caliper. Everyone who has replaced brake pads have fought this friction.

  102. Did a test drive with regen off: 71.1 km, 20% energy remaining used 78 Ah and the duration was 1 hour 52 minutes. I'll drive the same route later with regen on.

    Tom R Simenstad
    2000 Th!nk City driver

  103. "2. Use some other way (another "brake" pedal) to control regen."

    When I made this comment I wasn't strongly advocating it. I'm sure the are other equally good or better approaches Jack could use. This is just one. My intention was to offer "an" alternative which would make Jack's Brake-pedal based system less dysfunctional. I say dysfunctional because at the time Jack was reporting that it actually reduced his overall efficiency.

    One idea I do strongly suggest is a method of turning off regen when you want to coast. To be worth while, this would have to be very easy to get at. Some of the drag racers use a nitrous oxide button mounted on the gear stick. This is a normally open push-button. It lets them control their NOS so that they can turn it off and on very quickly and easily so as to maximise performance without destroying their clutch. Jack has actually mentioned several times that regen makes it more difficult to change gears. This would give him the best of both worlds. The ability to use regen /or/ coast as he wishes to maximise efficiency and controllability of the car.

    The only real down-side is it may be difficult to do without looking tacky.



  104. "One idea I do strongly suggest is a method of turning off regen when you want to coast. To be worth while, this would have to be very easy to get at."

    On the Mini COoper, I do have a simple dash mounted switch to shut regen off, and it is very useful.

    For one thing, we have six gears in the Mini. These tests have been done entirely in 3rd gear, and indeed, that is the most comfortable gear for most driving. But with regen, we do have some oddities in shifting gears if we are in a regen situation at the same time.

    More importantly, the regen leverage is very different in 3rd from reverse. As a result, backing the car with regen on is an exercise in frustration. If it is set nicely for 3rd gear, it is less aggressive in 4-6 which is fine for highway driving. But when you put it in reverse, it is very low gearing and the regen becomes very strong, so the car jerks something awful. So I just switch it off when backing up.

    There is of course another way of dealing with this. I could just leave it in 3rd and reverse it electrically.

    The other situation is ice and snow. Regen could actualdly be dangerous in this situation. So having the switch to turn it off is a big plus.

    Believe it or not, I've not figured out a way to disable regen on the Curtis with a hardware signal. There must be one. BUt I can't find it. I guess I could route the brake signal through a switch, but that would not affect regen on the accelerator.

    Jack Rickard

  105. "An advantage of AC is that the speed controller can double up as a charger. Two EVs I positively know do this is Tesla Roadster and the new Th!nk."

    AC Propulsion seems to have come up with this idea first and in fact they patented some of it. THis seems to have evolved into the Tesla system.

    I would very much like to explore this on Speedster Part Trois. But it requires some software changes in the controller. Basically, you use some contactors to connect AC to the controller through the motor windings. That part's pretty good. But I think we'll have to do some WinVCL work to get the controller to actually go into regen when doing this. Unfortunately, I can't find the WinVCL programming environment that Curtis refers to anywhere yet.

    The other problem is that the Curtis controller is limited to 130v input. I'm not sure how this would work with 240 - perhaps just fine. It is 120v per phase. I could put one phase on each motor winding with the third connection to the neutral and disconnected from the motor.

    Jack Rickard

  106. "I'd like to throw another thought out regarding WH. Sorry to keep on but it is important to this subject. "

    Dan, please do NOT keep on with it. I assure you it is NOT important to the subject. I know you believe it is, but you're manfully attempting to count this TWICE. It's already accounted for as I've explained, and as I noted in the beginning that I did NOT want to get into this. I know exactly where it leads and it is OFF INTO THE WEEDS. It has NOTHING to do with whether regen provides efficiency gains or not.

    Jack Rickard

  107. "Quite an unexpected turn of events. Will the next episode dedicated to this (likely to expected) change in results?"

    I don't know. I want to do more testing on the Mini this week. And we should have the transaxle back by the end of the week. But we're going to go play airplane on Saturday and will be gone till the following Wednesday. So it could be a week or so until we have a transmission back in the Speedster.

    If the results are swamped by what kind of a drive, that makes this all a little more complicated. In my mind, HALF of the drive we were doing was very urban. It was half the TIME of the drive, but not nearly half of the distance. In fact, the truly urban part turns out to have been less than a QUARTER of the drive by distance. What would regen efficiency care about time? Nada.

    In arrears this seems obvious. And the difference, JRP, is that I'm quite willing to change my mind. In fact, I was very uncomfortable with our original results but could not seem to find a handy cause.

    Yes, we'll probably have Brian replicate the results. He drives actually much more aggressively than I do, so that will be a difference. And it will be interesting to see just how much noise in all this comes from the driver. It's true that I take delight in just rolling and rolling and rolling, in some situations for a mile or more, using nothing by way of energy and actually passing cars along the way.

    So I want to do more test drives in the MIni and more test drives in the Speedster and be a little less quick to come to any "conclusions."

    Jack Rickard

  108. Jack, thanks for the testing and publishing your results.

    How long had the car sat before the test? If the car sat for a few days, I could see the oil being stiffer and the batteries less responsive for the regen test. Then the batteries would be a bit stronger for the non-regen test the next day since they have been recently exercised.

    Ideally you'd do "ABABAB..." type testing -- regen one day, nonregen the next, multiple times. This would accomplish 2 things: One is to look for system drifts, like batteries working better due to recent use. The other is to give you some statistics -- if the numbers are jumping around you might not have a statistically significant difference (or sameness).

    I have a car with regen, I'll have to try your test myself. One thing I have noticed: If I do strong regens (like occasional 200+ Amps), my lead acid batteries stay very well balanced with no BMS (all within about 0.03 Volts). With weak regens (50 Amps), the batteries quickly drift to be 0.5+ Volts apart. Jack have you seen anything like that with your cars and lithiums?

  109. Jack, have fun at Oshkosh (one of these days I have to get there; there's a Tiger Moth in my family.) Just a comment on regen and everyday driving, it's not only to increase range but provides an analog to "engine braking" to keep speed under control going downhill. There's a couple of roads around here that I avoid with my pickup conversion (DC series motor) precisely for that reason. Fairly steep hills with tight turns and/or stops at the bottom. There have been some very bad accidents at these spots.

    Also, a technical note about Prius electric motors. At speeds greater than ~41 mph the ICE is spun up as MG2 will overspeed if the ICE is not turning. This is because of how the CVT works. It's very common to use electric power exclusively at highway speeds (I see it all the time on ours); the ICE is turning but MPG is 99.9 reflecting no fuel usage. It's a mechanical thing (at least that's the way I understand it.)

  110. Hi Jack,

    I think your analysis makes sense if you change your driving style without regen in order to use the kinetic energy better. This requires watching traffic in front of you closely and leaving a little more following distance. If you do this, you can make significant differences in energy usage.

    If you can coast so that you are just arriving at the light as it turns green and cars start to move in front of you, you've successfully used all the kinetic energy and haven't paid any regen losses.

    Using these techniques, I've increased my fuel mileage in an ICE by 30%. This is on a very familiar, commuting route where I know all the places the traffic slows down.

    Done well, I think this will always do better than regen unless you have situations (such as long hills) where you just can't use the kinetic energy.

    A driver that doesn't like to coast (many don't) may see much better results from regen. They aren't going to use the kinetic energy anyway (just turn it to heat) so it is useful to get some of it back in the battery, even with losses.

  111. I drive a Solectria force, AC drive w/regen, to work and back quite often. The force uses the accelerator pedal method of regen with a rather large neutral spot in the middle between the accelerate and regen ranges. My drive is very hilly and (round trip) has about 50 traffic lights.
    Couple of things I'd like to note. One being that my AH useage day to day can be +/- about 3ah (say 6-7%) depending upon traffic, which lights I hit, etc. The second being that, on the few occasions I've tried it, driving without regen drastically cut into my range.
    I have hundreds of trips clocked in (although not many without regen) and my own real world experience seems to bear out the it works very well.
    Also as one on the side benifit it greatly reduces the wear and tear on the cars friction brakes. I may never need to buy replacement brake pads for this car.

  112. Hi Jack,

    I'm not sure how to explain your results, but I just tried this on my Tesla Roadster. I ran a three-legged route (with an uphill section, downhill section, and flat section) both with regen on (as normal--I just let off the accelerator to slow down) and off (whenever I started to slow down, I would put the car in Neutral and hit the brake pedal instead).

    The uphill leg had 5 stops. It took 10% more energy with regen off.

    The downhill leg had one stop. It took 105% more energy with regen off.

    The flat leg had 4 stops. It took 17% more energy with regen off.

    For the whole route, it took 25% more energy with regen off. I guess a better way to say it is that when I turned regen off, I lost the opportunity to reclaim 20% of my energy, which was turned in to brake heat instead.

    It seems to me that regen is not "good" in that you should not *try* to regen while you are driving. It is only "better than braking". So the rule should be that if you are going to brake, regen instead. But if you don't have to brake, don't. If you can do your whole trip without slowing down, you'll get the best range of all. Regen doesn't add range; it only saves some that you might otherwise waste.

  113. Here's another way to describe how to get the best range:

    1. Maintaining your speed is best, if you can.
    2. If you can't do that, then regen to slow down.
    3. Only use the brakes to slow down if regen isn't doing the job.

  114. Maybe this has already been said and I just missed it - but (assuming you don't have a mountain halfway along your journey) regen braking is only fixing a problem implicit in a particular driving style. If you hold a constant speed right up til the last minute and then brake, then regen will save some of the energy lost to braking, but energy-conscious drivers will try to coast whenever they know they will be stopping shortly. This means the kinetic energy stored in the vehicle is directly used to overcome friction, and so doesn't have to go through the energy loop of generator -> controller -> battery -> controller -> generator.

    Regen braking would probably show more significant gains on a racetrack or in very hilly / mountainous terrain. For the above reason, however, it will probably never provide a compelling advantage when commuting in a light vehicle.

  115. Are you saying 107 miles range with a DC motor (no regen)? And also 107 miles range with an AC motor (with regen)?

    If so, then there's your problem. DC to AC controller might consume 10-25% of the energy that passes through it. DC to DC controller should be much more efficient.

    Do the test again with the AC motor with regen off. And compare it to the test with the AC motor with regen on. In order to measure the effects of regen, keep all other variable the same.

    By the way, it's nice to see someone putting a DC motor in an electric car. Those DC to AC controllers waste so much energy.