Tuesday, February 8, 2011

On The Road Again....

Well, we kinda/sorta have the Speedster Redux back on the road again. We're a little ginger with the new motor/controller combination with the brush seating issue. But even at a 2000A/sec slew rate we got a 10 second 0-60 time. Hardly anything to brag about I'm afraid. I think we can improve it, but the much touted performance hoped for is probably not to be.

I'd guess because we've become a tad bit obese - probably a little blood pressure issue as well. I don't know if that's me or the car.

In any event, we are a portly 2385 lbs with the additional 21 cells in the car. Our front/rear distribution has moved a tad to 38/62 from 40/60 in Speedster Part Duh.

The heavier weight makes it a smooth ride and the center of gravity is sufficiently lower you can indeed feel it. The electric speedster never did "roll" in a turn nearly to the degree of the ICE version, but now it just doesn't lean at all.






The Chennic chargers we've used very successfully at lower voltages in the past didn't last the week. They just crumbled in front of us with an audible and alarming snap crackle pop sound eminating from the rear engine compartment for 15 minutes. No smoke or fire. But no 12v either.

We've replaced the Chennic with a homebrew - three Vicor bricks obtained on eBay. Tom Alvary has pointed out that these come from Troy Gaud who in fact works for Azure Dynamics. The legend goes that they ordered a brazillion of these from Vicor for an Azure Dynamics DC-DC converter and are remaindering hundreds of them on eBay at $24.95 each. Such a brick direct from Vicor would be about $172.

The VI-251 series is supposed to be a 150v input 12v output module. It should handle 85 to 215 volts. But according to Tom, these are 'special" made for Azure and indeed they are marked for 200v input. He tells me that he has it on authority that they are good from 130v to about 250v. They put out 200 watts each.

So we cobbled three of them together on a heat sink for a steady 12v output. Unfortunately, with the headlights on we're doing about 11.4 volts and the EVnetics Soliton1 errors out and shuts down. Jeffrey Jenkins notes 11v is the trip point in actuality and indeed, our HID headlamps might kick it down that low when coming on. So we had a problem.

Fortunately, the Vicor bricks can be trimmed for different output voltages - all the way DOWN to about 2.5 v but up as well to about 13.2v maximum. We put a 470K resistor between the T terminal (trim) and S+ terminal (sense positive) on each of the three bricks. The result is 600 watts right at 13 volts.

We are further advised that anything we can do to stabilize that 12v rail would be a great boon to the Soliton1 and indeed some concern of DESTROYING the Soliton1 on noisy 12v spikes. I rather think that should be EVnetics problem, and point out that the automotive spec indeed is to operate down to 9.6v. But it gives me a good jumping off place next week to start talking about capacitors and ultracapacitors.

I've put this off for some time. And I'd like to put it off further. I fear my views on supercaps are going to seriously disappoint many of our viewers. And it appears its kind of like the regenerative braking thing. I've got Siemens and Maxwell and several universities all publishing reports of 30% gains in efficiency. Iv'e been playing around with it a bit, and I guess at this point I think it is ENTIRELY BULLSHIT. And i mean ENTIRELY. No such gains. And they leave clues in every report - mostly noting that even HIGHER gains can be obtained with regenerative braking and with the "proper algorithm" to predict regen periods and driving cycles.

We seem to paint ourselves into these corners with great regularity. I read all the typing I can. Get all excited. But if I can't make the meter move, I can't make the meter move. And all the kings horses, and all the kings men, cannot show me HOW to make the meter move, except to suggest longer and longer strings of tasks I can endure that somehow with JUST ONE MORE THING will get it there.

So I take on the dragon. And for months get screamed at by everyone who's ever been in an online forum and just KNOWS how it is and how its supposed to be and cannot imagine WHY Siemens or Maxwell or whomever, with ALL their brilliant and educated engineers who ALL eat right and exercise regularlly and are much trimmer than me and have LOWER blood pressure to boot - people who clearly are NOT little and ugly and whose wife dresses them funny.

And I don't know, my knees hurt and my elbows hurt and I just don't know if I want to charge in on another fairy tale deflating confrontation with the entire population of the planet. So I've put this off and put it off. Maybe I can put it off some more.

But ultracaps DO make marvelous filters. And I think we have an astounding insensitivity to just how much electronic noise these cars make and how that impacts everything in the car. Worse, it is very difficult to deal with in that it goes to longevity - ie DC-DC converters for example. But also controllers and batteries in general. And so it's hard to make a case for something that preserves the life of your components for 20 years when the car is two weeks old. How do I explain it and worse, how can I prove it and finally, how can I show THAT on video. We struggle to communicate the BMS issue with CARS bursting into FLAMES all around us.

So its not a topic I really relish.

We do hear from a lot of viewers about the Lee Hart Batt Bridge concept. It's easier. I still kind of hosed up the explanation on the video, but close enough for government work. The issue here is we don't use them. Actually we do. We just don't do it that way.

Speedster Redux has an old version of Victor Tichonov's EVision and it performs the batt-bridge function quite well actually - displaying an LED bar graph that comparies one half of the pack to the other. I myself am going a different way with a Roving Networks sensor and some Objective C software on a Macintosh that will give me the same dislpay.

But neither of those are really available. And the Lee Hart Batt-Bridge is both easy to implement, and costs about a dozen dollars. It accomplishes most of what the BMS crowd is trying to save us from - sudden infant cell syndrome where a cell mysteriously fails of its own volition. We haven't had much of that, so we're not quite as tuned into the need as they would like us to be. But the concept has bare merit. And for $12 you could implement this and it would probably work to accomplish that.

In any event, I've had so many requests for a detailed explanation that I finally did one. The original writeup and diagram has always been online and really covers it as well as it needs to be covered.

During our ride, we went out to the airport and happened to catch my daughter Jennifer in the act of committing aviation in our little MD-500C helicopter. This was another one of my little obsessions where I spent a million dollars on a $300,000 helicopter to wind up in the end with a very solid little $300,000 helicopter.

Probably spent that on the daughter as well. IN any event, after a VERY rough adolescence, she has emerged as a remarkable young lady on the Dean's List here at Southeast University, majoring in anthropology and she's been taking helicopter and scuba diving lessons. She likes the field work and presumes to displace Indiana Jones in the search for the Holy Grail and so needs to learn to scuba dive, fly helicopters and the rest of the James Bond repertoire to adequately compete. I'm actually very proud of her and might advise Harrison Ford to keep his grungy little deHavilland DHC-2 Beaver out of harms way if he knows what's good for him. She's already got a pretty heady reputation locally for bar brawling and sport fighting - but I'm hoping Southeast University can put a little polish on that. I do digress. My own fault. She really WAS cute at the Tai Kwon Do gig at age four. I just had no idea it would take....

Enjoy the show. 105 minutes as I recall...

I understand that the Jimerico dual 11 motor ships this morning... Hopefully we'll have it in hand by the end of the week.

Jack Rickard

29 comments:

  1. Jack,

    Once again, an excellent show you all put on this week. Yup, a kudos to all of you who put this show together. I like the meter gauge batt bridge. Until your software one is done I will try my hand at the simple little project. I like the solid state little dc dc bricks too. Talk about saving space. Might just have to try that as well. Nice drive but on video it's kinda long. I now know your streets better than mine. :)

    Pete

    ReplyDelete
  2. Jack,

    Every time I saw you write about using the ultracaps I kept saying to my self, why not just put in more batteries instead? These LiFePO4 cells are so efficient compared to lead acid I couldn't imagine ultracaps helping much at all. I understand how a small set might help reduce the ripple the batteries see but not much else. It sounds like you are coming to the same conclusion.

    David

    ReplyDelete
  3. This is the second time I've written a brilliant reply today, and Google just tells me it "cannot comply with my request" and tosses the whole thing away. DOes this little short one make it through?

    ReplyDelete
  4. David:

    In a nutshell, yes. I've reviewed all the literature and indeed built a $10,000 47 Farad monstrosity that weighs 100 lbs and takes up quite a bit of room. We've played with every sort of IGBT switch/coil setup. And it appears to be yet another myth that sounds very, very good. In practice, it's a total loser with respect to range or performance.

    ALWAYS you can get more range and better performance with more cells occupying the same space, weighing the same, and costing about the same.

    I hesitate to take on this beast because it once again places me, standing their with a few meters in my hand, against the entire free world, all insisting that I'm just doing it wrong and if I would just try ONE MORE THING it would work like magic to make magic electric vehicle that would go 1000 miles on a charge at 720 mph. They want magic dust SO BADLY and if you take it away from them, they are deeply deeply butt hurt and express that as screeching anger, and of course I'm the focal point because I happened to point it out - viewed as peeing on their campfire.

    So I don't know how I became the duty naysayer and wet blanket at the EV party. But I am from Missouri, and you have to "show me." Better is if you show me how and I can duplicate it. If I can't see it on some sort of meter, at this point in life I have to assume it is more magic dust bullshit.

    I guess I could keep my mouth shut and avoid the confrontations, the heat, and the anger. Not a bad stroke.

    All that said, we're going to install both a 500F Ultracap on the Escalade 12v and a 172V 47Farad ultracap system on the Escalade. Since it won't get us ANY more range, and it won't get us ANY better performance, I'm sure you're going to wonder why.

    If you be nice, I might try to explain it.....at some point. Or maybe not....

    Jack Rickard

    ReplyDelete
  5. Hi Jack,

    Please try to explain :) I will agree with you on the range (no significant effect) - but what about life time on the battery?

    Martin.

    ReplyDelete
  6. Jack and Brian, Thanks for another great show. You can get VI-251-09 dc/dc's on ebay for $10.00 ea. when you buy 20
    http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=190498459191&ssPageName=STRK:MEWNX:IT

    Mark urban

    ReplyDelete
  7. Yes. What about battery life? How about one guy in Australia who put over 100,000 miles on LEAD ACID batteries by buffering them with Ultracapacitors. We can certainly smooth the demand on the cells AND send those transient noise spikes to the back of the pack where they belong pretty pronto.

    Jack Rickard

    ReplyDelete
  8. The snake oil salesman is alive and well and so is the crowd daisyeyed by his sermon. I have yet to see good proof of a shunting, or otherwise, BMS that works. Here I've seen good proof that they are not needed.

    I'm also seeing prismatic cells laid on their side in Speedster Redux, something not advised buy their makers. But then I've seen prismatic cell manufactures advise to use a "good BMS" as well. I assume, Jack, that you have had no problems putting cells on their sides as long as they are charged and discharged properly.

    Okay, time to poke my nose into the magic kingdom and see what's burning. -Klaus

    ReplyDelete
  9. I never knew that there was a supposed benefit in distance or performance, I thought the ONLY benefit was increased longevity of the battery pack. That's the only thing I was hoping to get out of caps anyhow.
    In any case, I think this one's a game changer:

    http://news.yahoo.com/s/yblog_thelookout/20110209/ts_yblog_thelookout/wikileaks-saudis-running-out-of-oil

    ReplyDelete
  10. Hey Jack,
    Has anyone tried to sell you a BMS for all those ultra capacitors?

    A point of interest. All those caps in series. They self level the charge voltage in series don't they. Could they be tapped every so often back into the batteries? Presuming the correct number in the string.

    Or am I mistaken in some way.

    ReplyDelete
  11. Andyj:

    Maxwell indeed provides us with little shunting levelers that can do a couple hundred milliamps.

    As to interleaving the caps to the cells - don't go there girlfriend. Just don't get me started.

    Jack

    ReplyDelete
  12. Firstly, on your lastly. Shhh! People might find out.

    And the BMS.. I *was* joking. Shunting levellers to limit uncontrolled voltage gain through EMI spikes? Too weeny for anything else.

    We need lots of wires across those cells to ensure reliability. Everyone here knows that!!!

    Well... I've baited Jack and got baited breath for the low-down on these things. >:-)))

    ReplyDelete
  13. I'd be very interested in what you find out with ultracapacitors. I always understood that they were a band-aid to cope with the lousy C rating (hence poor power density) of early Lithium ion cells? Now a non-problem as 1000 Amps in Speedster Redux illustrates

    ReplyDelete
  14. How much does the pack voltage sag when you are drawing 1000 amps?

    ReplyDelete
  15. It depends of course on how long we run at 1000 amps. The log file shows a minimum of 153 volts at 1000 amps from a high of 188v to start that run. This is fully 35 volts or 18.6% of our pack voltage.

    Jack RIckard

    ReplyDelete
  16. If you search for "100.0" in the log file that will take you to every entry where duty cycle is at 100%. Look for all the entries where motor current is at 1000A (or very close to it) and note the pack voltage and RPM. Find the highest RPM for the lowest (sagged) pack voltage and there you will have the Back EMF limited RPM. I estimated that 4200 RPM was the highest you could get with your pack delivering 1000A, but it appears that the actual maximum is 3913 RPM (this is with the pack sagging to 155V).

    More amperage would get you faster acceleration in the first second or two of the run, but pack voltage appears to be the real limiting factor for the rest of the time (indeed, in some of the runs recorded in the log file the controller is at 100% duty cycle for much of the time, with motor amperage falling as RPM continues rising).

    Your pack isn't sagging all that badly, actually. You can determine the resistance of your entire pack from the difference in pack voltage when running at 1000A motor current between 100% duty and 90% duty (for a difference in battery current of 100A). One pair of entries shows a difference of 3V (158V at 90% and 155V at 100%) so 3V/100A = 30 milliohms. This is exceptionally good for a prismatic LFP pack. I am duly impressed. The estimated short circuit current for your pack is in the 6kA+ range, or about twice what the much-abused AGM pack on our dyno can deliver.

    ReplyDelete
  17. Hello Jack and the amazing crew;

    I was thinking about the use of automatic transmissions with electric motors and especially CV transmissions.
    Feel free to tell me that I'm out of my tree, but here's my thought:

    In an ICE vehicle, when the driver demands additional torque, for passing or hill climbing, depressing the throttle causes the transmission to shift down, in order to increase the engine's RPM's, to reach for the higher torque point.
    However, with an electric motor, the higher torque point is found at a lower RPM than the transmission might be causing the motor to use.
    Would that not cause problems with mapping the shift points in the transmission controller, and the technique for achieving greater torque for acceleration?
    Or, am I nuts?

    J.G. Joseph (Joe) Jackman
    Vancouver, Canada.

    ReplyDelete
  18. I think it boils down to power and efficiency, for both electric and ICE. Power is the product of RPM and torque as far as either motor is concerned. If you want torque, gear it down. If you want speed, gear it up. Both can be had, but not at the same time with the same power. Then for cruising you want efficiency and both ICE and electric will have a point of maximum efficiency. I think a good system would take all three under consideration for all driving requirements. The transmission should probably be reprogramed to match the new motor but nothing drastic, I would think.

    ReplyDelete
  19. Joe:

    Well it certainly could. Inevitably, the shift points on an automatic transmission for an ICE engine are going to be at the wrong point for an electric motor.

    Fortunately, things are a little more complicated than that. Pressing down on the throttle doesn't precisely cause that. In today's automatics it is determined by a "controller" that uses engine speed, road speed, vacuum, throttle, etc as inputs and shifts the transmission electronically. Usually, there are "tuner" kits for these controllers that allow you to revise the parameters.

    The 6L80E transmission in the Cadillac gets many inputs from the ECU and the controller itself is inside the transmission case. I think it is an exceptional transmission. Unfortunately, no one has developed a tuner kit for it. I STILL think we might have used it as it has provisions for manual shifting in the car already. But everyone tells me it cannot be made to work.

    TCI has a 4L60E transmission WITH their own programmable controller AND paddle shifter inputs. And they do some interesting things to add a couple of actuators to make this 4 speed also a six speed just like the 6L80E. FInally, they "harden" some of the components allowing them to use higher torque values - up to 850ft lbs.

    This transmission was $6800. It still irks me to pull a perfectly good 6L80E out of the car and spend that for $6800. Particularly when I suspect that I will find the manual shift to be the normal order of the day. BUt Matt and TCI assure me they can tune up this controller to shift quite automatically and perfectly for an alectric motor. And if it can be made to feel good, we'll use it that way.

    Again, in practice, I think we'll start out in second, double click the paddle to fourth, and double click again to sixth during most accelerations. But predicting the "feel" of transmission gearing is not something I do well. It always surprises me. And in the MIni Cooper we learned there's a reason. SImply changing the controller caused us to favor 2nd gear most of the time, instead of third.

    I remain a believer in transmissions over direct drive. You need about half the power with a transmission as you do without. In this case, we are going to need so much power to take off, that I think the hydraulic buffering in the torque converter is actually going to be quite important.

    I could be wrong.

    Jack Rickard

    ReplyDelete
  20. Jeff:

    The limiting factor on voltage, aside from the number of cells, is the brushes and commutator arcing. Previously, the unofficial "rule" for Netgain motors was 170v. George Hamstra reports that HELWIG feels their new brushes can do 192v while minimizing this arcing. We are doing that max voltage and so far haven't noticed any arcing or evidence on the commutator of arcing.


    Very interesting. I was thinking of going to a LOWER voltage for the Cadillac and doubling the batteries. But the new grey CALB cells are showing 0.6 milliohms for internal resistance, even better than these blue CALBS - and these set on the floor for 15 months.

    So it might be better to show a higher voltage with single cells as they appear capable of providing the entire 2000 amp maximum.

    From what you are indicating, we should be able to decrease our 0 to 60 times by simply shifting earlier and keeping the motor at a lower RPM. That's a little counterintuitive, but we'll try it.

    Jack Rickard

    ReplyDelete
  21. Well actually it doesn't quite boil down to power and efficiency. To me, it boils down to feel. And while we can mathematically model power and efficiency, my experience is that this fails generally in feel.

    Different cars feel different when you drive them. You can model gear ratios, and power levels, and calculating torque and power for any given RPM is trivial. Making it feel right is a little more difficult.

    And simply more power often doesn't do it.

    Jack Rickard

    ReplyDelete
  22. Jack: Yeah, the WarP-9 really doesn't like to see more than 170V - especially at high current - but that's what the maximum motor voltage limit function in the Soliton1 is for... just set it to 170V and don't worry about it anymore.

    And yeah, it is a bit counter-intuitive that with an electric motor you need to shift a little bit sooner the harder you are accelerating. However, even monstrous amounts of torque doesn't add up to much horsepower at low RPMs, so you don't want to start off in too tall a gear, either (which you already noted yourself in arguing against direct drive above - a sentiment we all share at Evnetics/Rebirth Auto, btw).

    So you may not be able to shave off too much more time from your 0-60 runs, but how does the car feel?

    ReplyDelete
  23. Oh, one other thing... don't forget that the WarP-9 is about 65% efficient at 1000A.

    ReplyDelete
  24. Jeffrey:

    It feels very good to me. I guess with the much higher voltage and the signficantly higher power over the Kelley, I expected it to be a bigger difference - not a lot really. But Brian says I'm crazy and it is much more responsive.

    It feels VERY good - I was out today in wine country, albeit a little early this time of year and doing some preliminary testing. I'm kind of into the winding blacktop road thing at 45-55 frankly in this type of car and the NEDRA gig really doesn't hold much for me frankly. Brain and Matt like the performance, or so they claim. I like tuned elegance and sure feel.

    We are heavier, there is just no way around it. But it doesn't feel bad, just a little different. And it does have its compensations. I did a couple of different style runs under different conditions and was sporting just barely over an amp hour per mile - like 1.04 on average. I found that a bit shocking actually. I was hoping for a 150 mile run. It would appear we'll do 175 with an outside shot at 180 if we want to hurt em a little.

    When they're new like this, we generally see something in the 184-185 amp hour range on these cells. I'll probably want a read on cell voltages at a mostly discharged state to make sure there isn't anything amiss, which there won't be frankly, but then we'll make a serious run at 170 or 180 miles on a charge.

    So I'm pretty jazzed about that. Stopped at Hunter Valley Winery this afternoon and the car of course made a huge splash.

    They always do.

    We're still fighting the Reva2/fuelgage/tachometer situation. Brain found our RPM off in RPM mode and changed a setting on the tachometer, now in current mode it reads twice the current and so I have to change the setting on the Reva2 to match it. And I don't actually know what he did. AND I'm not sure he does either. But he has RPM with a little light gun within 30rpm now. So we have SOMETHING reading correctly.

    And the full sweep fuel gage seems to be tracking nicely at this point, I'm very pleased to report.

    We removed one of the three heater PTC elements and I can now bear the heat when the heater is on, it was burning my legs previously.

    So it's coming together. Ironing a few things out. The new 3.44 R&P gives us taller gearing and so all four gears mostly make sense now, although I can easily pull away starting in second gear.

    As we noticed on the Mini, their is a finicky relationship between gear ratios and available power. The Mini now feels better in 2nd in town where before it was 3rd. But we lost 30kw in power going to the RInehart Motion Systems controller from the TIMS 600.

    So it all IS about feel. We haven't had this much power with a 3.44 before, and so it is kind of new ground. But overall I would say it feels very good, a little ponderous at times from the higher weight, but you would have had to driven this car a LOT to notice that. Very low COG of course. So your butt is very nailed to the pavement. Hard to describe.

    Jack RIckard

    ReplyDelete
  25. Hey Jack,

    Have you done any testing with LIFeP04 in float conditions?

    Specifically UPS/Solar array. What float voltage should we setup I was thinking 3.3 - 3.4V/cell but I was wondering what your thoughts were. The way these battery based grid tie solar inverters work you simply set the sell voltage which basically sets the float voltage.


    Thanks, Jim

    ReplyDelete
  26. Jim,

    The CALB (SE) spec sheet here http://lithiumstorage.com/index.php?main_page=page&id=18&zenid=ffqtqacnjf16t8g6dmu7vplu42 lists 3.4V. I read of someone overcharging a TS cell by leaving it at 3.65V for an extended period of time. I wonder what the theoretical full voltage is.

    David

    ReplyDelete
  27. Thanks for the spec sheet David. I guess I don't really care where "full" is. I just want to stay well away from overcharge territory. I don't see any benefit in keeping them at full charge. Even at the expense of a few less AH. The 80-90% full area I think would be best for longevity of the cells and that is the goal.

    My laptop is a perfect example. When I leave it plugged in to the charger it shortens the life span of the battery. Not sure how high they are charging that but you get the idea.

    The reason I am looking at this is the huge losses in charging lead and keeping it charged or overcharged "equalized" to to keep it from loosing capacity due to sulfation. lead really doesn't like sitting undercharged and sometimes the sun just doesn't shine for weeks and the inverter is always drawing some power.

    Where as lifep04 doesn't really care as long as its not under or overcharged.

    Thanks, Jim

    ReplyDelete
  28. You're welcome Jim. The reason I was wondering where the theoretical full voltage is because then it would be easy to stay under it. By my measurements on a 100Ah TS cell I found at most 1Ah between 3.45V and 3.65V. I charge my pack to 3.485vpc but would charge to less than 3.4V if floating for the reasons you mentioned.

    David

    ReplyDelete
  29. There appears to be further support for this. I read a recent article that said Nissan Leaf and Toyota Prius only charge to 90% of full charge. GM Volt only charges to 80%. And even Tesla has an economy charging mode that charges to 80%.

    Although I'm not automatically an OEM fanboyz, it would appear if you are warranteeing cells, undercharging is one of the preferred techniques.

    Jack Rickard

    ReplyDelete