Tuesday, May 31, 2011

Let the Voting Begin

This may indeed be our longest show ever - right at three hours of HD video. We had a number of interesting developments and breakdowns in the shop, and then we introduce the TEN finalists in the EVTV "Build Your Dream" EV components contest.

Each of the sponsors selected two of the 955 entries received, and I selected two as well. In itself, that you had a one in a thousand chance at $20,000 in EV stuff is kind of a remarkable thing. You won't find that often or everywhere.

In this episode, we attempted to summarize all of them with a "reading" of their entry and a bit of discussion as to why the sponsor may have found that an attractive notion.

In any event, we will running the balloting from now until August 31, 2011 and each of our viewers will have a chance to make a selection. Help me by making your best considered decision as to where to ship this pile of EV components to have maximum effect. You may vote HERE.

We are designing a drive train for the Aptima Motors eCobra. As a bit of an add-on, Paul at Aptima wanted a J1772 charge station like our Texaco gas pump. I completed that work this week. Paul selected a Chevrolet Corvette pump style and we've been wiring it for J1772 charging. We show the result in this episode.

I have often said I've never met a connector that I liked. I may have finally met my match. The Yazaki J1772 connector has much to recommend it. We are gradually converting all of our cars to work with this connector, and I have gas pumps at the shop and at home. The one at home does not yet feature the J1772 connector, but it will soon. I have fallen in love with them.

Why? They have a very solid feel. You can insert them VERY easily and they lock securely in place with a faint "schnick" as the switch locks in place. We have long eschewed the proper 240 amp connectors on our cars because of the desire to be compatible with 120vac opportunity charging at grandma's house. So we've used Marinco marine covered NEMA 5-15 connectors. They hold up pretty well, but many of our extension cords do not. Over time they become loose and are a bit flimsy. QUITE a bit flimsy compared to the solid cast plastic J1772 connectors.

We had previously published a diagram of how you could wire your car to accept a J1772 charge using a switch, a diode, and about two resistors.

I've recently received a number of e-mail questions seeking this information. We now have an index thanks to Christopher Fischer, but it seems a large percentage of our viewership would expect me to know precisely what episode and minute any particular discussion could be found. Google has indeed spoiled us.

The communications between the J1772 Electric Vehicle Support Equipment (EVSE - we call it a gas pump) and the car is really very simple. The EVSE generates a 12v square wave alternating between +12v and -12vdc and applies it to pin 4 of the connector via a 1 kilohm resistor (1000 ohms). The junction of pin 4 (copilot) and the 1k resistor is then monitored to determine its voltage level.

The duty cycle of this square wave indicates how much current the EVSE is able to supply. The charger in the car is then supposed to limit itself to that amount of current. We don't know of any available charger that can do this. As the nature of such specifications are, we would expect this to be ignored for several years, and at some point in the future, someone will "discover" this feature and use it to communicate something else. It is possible the OEM cars actually look for this waveform, but we would be astounded.

Why? Onboard chargers are rarely of a capacity above 3000 or 3600 watts. They just become to physically large and too physically heavy to fit in a car at much above that. You might see a few 5000 watt units in some select vehicles.

The EVSE seems to have broken into two levels of Level II charging - 32 amp and 70 amp. The first is 7640 watts more or less and the second is 16800. Almost everyone is opting for 70 amp cables and connectors. Either way, its much more power than the car can use in almost every case. And so the recognition of available power is rather moot. And will likely become more so in the future until Level III DC charging becomes a thing.

So the "spoof" we've come up with on the car side is simply a diode, to eliminate the negative portion of the waveform, and two resistors that tie back to the neutral pin of the dual phase supply. This supply, identical to your home 240vac of course, consists of two phase lines of 120vac each, and a neutral return which is tied directly to ground in your electrical panel.

When you plug in the car, the diode and first resistor form a voltage divider with the resistor in the EVSE. It is designed to drop the positive portion of the waveform to +9v. The EVSE would detect this as a "vehicle present" signal. When you throw the switch to engage the other resistor in parallel, this drops the vehicle side resistance to about 877 ohms and the voltage to about +6v signaling the EVSE to provide power. There is another level lower yet for signalling for ventilation. This would be for flooded lead acid batteries that produce hydrogen when charging. We don't expect that to be used at all.

We rather thought someone would develop a little PCB to implement all this on the EVSE side quite properly. David Kerzel threatened to, but never completed the project. And someone should. One viewer sent us a device from Menneckes that purported to do that. It was quite complicated and limited to 32 amps. And it's somewhat expensive.

So we genned up a horribly simple little spoof circuit to accomplish basically the same thing, but without the square wave. It just uses a simple 12v to energize two Kilovac ANA200 contactors. Each contactor carries a single phase of the power. And so when they are NOT energized, there is NO power in the cable at all beyond the 12v signaling voltage which is provided by a 3 amp 12v supply inside the pump that feeds off a single phase of the input power.

The circuit is shown below.

This circuit is so blatantly simplified and NOT SAE J1772 2010 that I refer to it as J1771 1/2. The entire circuit is mounted on a 1 inch square of PCBoard material. The heart of it is an LM1458 operational amplifier we use as comparator. 12v is applied across a 5k potentiometer to ground. The wiper is adjusted for 6v and applied to the + (non inverting) input of the opamp. The same 12v is applied across a 1k resistor to the - or inverting input. The junction of the 1k resistor and the opamp input is tied to pin 4 of the J1772 connector - the copilot signal.

The output of the opamp switches a fairly sturdy MOSFET which is connected in series with the coil of a small relay. When energized, the relay switches 12v to the two phase contactors, energizing them to apply power to the car. It is somewhat important to tie the neutral pin of J1772 to the common or ground side of the 12v power supply.

In this way, the switch on the car, when closed, drops the voltage to 6v, the comparator output goes positive, the MOSFET turns on, and the relay switches on the contactors. We simply adjust the pot to trip at that point.

So we wind up with much of the safety and sturdiness of the J1772 spec, without the complexity. Eventually, someone WILL do a proper circuit for this with ground fault sensing. Better, use GFI circuit breakers in the panel.

A team of five or six engineering students from Imperial College in London, devised a plan to build an electric car and drive 26,000 kilometers down the Pan American highway from Anchorage Alaska to Urushaia at the southern tip of South America. They filmed this journey and we had ordered a copy when they completed the thing last November. We received it just last week and had actually forgotten we had ordered it as it was six months late in arriving. http://www.racinggreenendurance.com

The thing winds up as eight episodes on two DVD's. It's 19.99 British Pounds so about $30 before shipping but you have to see this. It is too comical to believe.

First, they built an open top SRO race car with a massive 50kw Thundersky battery pack and a 3.5 inch ground clearance to drive across the Pan American highway. Idiotic in concept from the get go. The tiny racing shock absorbers were sporting 1100 additional pounds of batteries. Engineering school apparently isn't what it once was.

But their "mission" was to drive from one press event to another, touting how electric cars were finally here and practical for every day driving. Instead, they rather proved both graphically and dramatically that it takes a half a dozen recent engineering school grads to even keep an electric car rolling at all.

Never at any point in the journey did they pick up on the irony. They wrecked the car twice. It burst into flames several times. And they failed every piece of equipment in the car with the exception of the very sexy EVO motors they had. All four shock absorbers broke at various times in the journey. They burned up their chargers. They burned up their BMS quite frightfully and the DC-DC converter at one point. Almost everything we've ever destroyed here in the shop, they managed to tick off the list on the road. No matter how rain bedraggled and discouraged, there they would be at the next press event with the canned speech about how great the electric car was and how well adapted it was to all of this. I watched every minute enthralled.

They never really went into any useful detail on all these failures during the whole series. They would show some discouragement, vaguely describe what they think happened, and then seek the assistance of some Mexican or South American farmer to make the repairs.

One of things I did pick up very early in the first episode was that they used a pair of Rinehart Motion Systems PMX100 controllers to drive the pair of EVO motors, which were direct drive to each rear wheel. And they complained of severe Electro Magnetic Interference (EMI) problems which they initially solved by wrapping the UVW phase cables from the controller to the motor in aluminum foil. Eventually, they replaced the foil with shielded cable.

We had been suffering for months with some gnawing problems with the Mini Cooper. I drive it every day. But there is always a slight shudder on takeoff at low RPMS. Worse, Rinehart has finally implemented a proper brake potentiometer input to control regenerative braking. We use a hydraulic pressure transducer to produce this 0-5v signal and it appears to work quite well. But whenever we had it CONNECTED to the controller, whether using this braking mode or not, we get an over current or over voltage fault at about 3500 rpm and the controller shuts down. So we are stuck with brake light signal regen and have to leave the transducer disconnected to drive the car.

We had actually sent Rinehart one of these transducers. Their car ran fine and they were unable to reproduce the faults. Very puzzling.

After watching the episode, I somewhat loosely wrapped the phase cables in aluminum foil and tied them to ground. The faults disappeared and the shudder as well. Elated, I had Karl remove the aluminum foil and install some braided steel tubing over the cables. To do so, he had to disconnect the cables from the controller.

Each cable was terminated in a ferrule crimped around the strands of the cable and then this is locked in place with some hex bolts in the controller. Unfortunately, when you lock them in, it deforms the ferrule. When you disconnect them, the cable then comes out but the ferrule does not.

To reassemble, he simply stuck the cable back in and tightened it down. But the old ferrules were in the controller. And a strand of the center cable made ground. When we applied power, we blew up the controller.

While this is discouraging, everyone is a bit elated. Rinehart is happy to repair it, and we're happy to send it off, because this nagging question seems to be solved.

The actual solution is of course the shielded high voltage cable. Unfortunately, this is horrendously expensive. But we think EMI may be the cause of a lot of things - like blown DC-DC converters, instrumentation ghosts, etc. It is particuarlly an issue with AC systems. But we are going to start using this cable anyway on ALL installations. Both battery and motor connections. It is manufactured by Champlain Cable Corporation EXRAD XLX shielded cable specifically for electric and hybrid vehicles. And basically after THIS months long drama, we are simply going to the use of that for everything regardless of expense.

Per usual, Chris Brune of Rinehart Motion Systems had brought up "noise" several times as a cause. He was sufficiently vague about it that I bullheadedly dismissed it each time. Just shoot me. Again, the main value YOU derive from these videos is from me screwing things up - AGAIN. I have on NUMEROUS occasions mentioned the EMI problem, particularly with regards to delicate BMS and instrumentation circuitry. I probably need to watch my own videos a bit. Too soon old. Too late smart. Spend the money on shielded cable and it won't be a problem. No more welding cable at EVTV.

We received our Tremec TKO600 transmission from Mike Fortes of Fortes Parts Connection this week, along with a Ram 11 inch ceramic clutch and pressure plate, a lightweight aluminum flywheel with steel face plate insert, and hydraulic clutch slave cylinder. Mike has a lot of Cobra experience and has provided some serious adult supervision here on components for the Aptima Motors drive train. This tranny/clutch should handle up to 600 ft-lbs of torque on the eCobra.

Matt Hauber has returned to San Diego to begin his life as electric vehicle converter extraordinaire. I understand he's starting out with an S10 pickup conversion there. He was a very talented and useful young man here at EVTV. His youthful enthusiasm made us all edgy. Or maybe envious. I'm uncertain which.

We welcome Carl Skircheck to the team. Carl actually had an early career in race cars in Detroit where he grew up and has managed a Chrysler parts department in recent years. Much of the mechanicals we find mysterious here at EVTV seem to come easily to Carl. With a couple of phone calls he has located some Silicon Oxide ceramic bearings for Speedster Part Duh which we are going to try. He is going to mount the Tremec in the Cobra this next week. He immediately brings to mind how embarassingly little Brain and I really know about automobiles and particuarly parts and mechanical assemblies. This may make the 3 lbs hammer entirely unnecessary here at the Motor Verks.

Jack RIckard


  1. Podcast http://feeds.feedburner.com/EvtvMotorVerks - best viewed with iTunes.

  2. Hi jack,

    " Onboard chargers are rarely of a capacity above 3000 or 3600 watts. They just become to physically large and too physically heavy to fit in a car at much above that. You might see a few 5000 watt units in some select vehicles"

    I have a Manzanita Micro PFC-30. Up to 30amps in at 240v. Isn't that 7200 watts? Plus, the size is quite small. The PFC-40 is also the same size factor.


  3. This will go down as one of the greatest quotes ever - Monty Python's got nothing on you, Jack:

    "Let me put this in terms that even a PhD can understand: Joel, you're as full of shit as a Christmas goose."

    Now that was funny :-)


  4. Regarding shielding. In theory, the three phases of the cable harness should cancel each other out if they are tied tightly together or, perhaps, braided -- may not be possible. I would not suggest shielding the wires individually, as per the inductive losses mentioned, but shielding the entire cable might be necessary. My own car is waiting for me to get another job so that I can go back to complaining about vendors. I am also working on the charger and instrumentation angle. I'm contemplating buying a PFC-50, ripping the control board out, reverse engineering said board, and then upgrading it to j1772. I'm also looking at the ST8916 ( stackltd.com) as a possible starting point for dashboard instrumentation.

  5. +1 on shielding around all three wires. Don't waste effort on putting individual shield wires on the three phases.


  6. Jack,
    I wondered if you were hinting Matt done a runner but wasn't too sure. A shock to see him go. Lets hope he does well. He's mechanically bright and will make great contributions to the EV world. He'll leave a trail of very happy customers. Just a pity he had to go with such a bang.

    Your Mini..
    I do believe shielding does not necessarily absorb magnetic flux over unscreened cable. Electrostatics returning off the motor, oh yes!

    I'm a great believer in keeping the motor leads as short as possible and the same length if only to ease the inductive loads against the controller. You might find the Earth shield does not need to be heavily earthed as the resultant of the three cables should hopefully almost equal zero volts. Like 3 capacitors to a common rail.

    Sorry to get on your case more. I previously mentioned about your mini's wiring scattered around the bay and said some lengths of spirwrap should do wonders for electromechanical reliability.

    One step at a time!

  7. This comment has been removed by the author.

  8. The PacTracker SUCKS. I wouldnt even use it on lead acid. It takes the power needed to run it from a single battery. So over time that batt will go out of balance.

  9. BMS mandator: "Jack, you just don't get it! My mind is made up. Don't confuse me with the facts!"

    LiFePO4 'Expert': "Lithium batteries drift. Because I said so! Quit confusing me with the facts!"

    Wow. That was a ridiculous response on why a BMS was required. I wonder what he would have to say about my data which contradicts his presumptions? Oh yes, "Don't confuse me with the facts!"

    David D. Nelson

  10. As several pointed out, if you put a Paktrakr on em, they WILL drift. I'm convinced these people see what they want to see. That's not science. And its certainly not education. You see what's in front of you. And you devise theories that might explain that, and then you test that. While you're testing that, you see something else in front of you. That's why I go in a circle most of the time and don't know anything. I know some ABOUT some things. But I never really get to know anything.

    Ok, so I HAVE learned to tell when someone is as full of shit as a Christmas Goose.

    Jack Rickard

  11. Those that can't do teach. Those that can't teach become department heads and make policy. Let this serve as a warning to all dumb asses that write letters under university letterhead. You will be famous for your stupidity! Jack, I love the "passion" in your exposure of this pretender.

  12. I shot Dr. Pepper out my nose. That was really good.

  13. Jack,
    What is the cost of the "EXRAD XLE 4/0 shielded cable"?

    Maybe we can do the Group Buy? :)

    Thank you.

  14. No idea. The 2AWG I hear is $10 per foot.


  15. After seeing this notable video I just have to share.

    Infrastructure and public perception will be the game changer. Its here in spades.

  16. Have you had any noise issues with the Curtis system? I have not, nor have I heard of other Curtis users having problems, so the question might be why is the Rinehart more susceptible? I also crossed my the phase leads instead of running them parallel, don't know if that made any difference, you might try it on the Rinehart with no shielding as an experiment.

  17. The cables were twisted for a substantial portion of their length.

    No, I'm unaware of any noise issues with the Curtis system, and yes, I think Rinehart should address this more in design than in prevention.

    That said, we do have anomalies observed in all of our cars at various times. I've had the Curtis throw faults and shut down. It is quite rare, but it happens. We recycle it and bring it back up, and drive on without further mishap. Since we cannot reliably reproduce this, and it is quite rare, it is anomalous. The Rinehart kind of points up the problem. I simply did not think noise could be the issue with the symptoms we were observing. But it appears it was and that others have had them.

    We've had other issues. DC-DC converters that operate fine for a period and then simply blow up - quite a smoke check at that, and more than one. I've become increasingly suspicious of noise on the pack input. EVnetics for example now has a special inductor for this specific problem that you put on the input to your DC-DC converter.

    The TIMS600 was apparently unaffected, but I can tell you that all of our instrumentation was just impossible with this controller. It totally destroyed an EVision and an Arduino.

    I have always thought this was the source of the problems the Battery Management Systems were having. They more or less work, and then they more or less burn the car to the ground. Anomalous. But with a real penalty on the occasional fault.

    We're just going to adopt shielded cable as a matter of course at this point.

    Jack Rickard

  18. Just got off the phone with IEWC

    They stock and sell the CHamplain Cable

    2 AWG IEWC P/N EXRAD02-XLEOBS-5 $4.21 foot
    2/0 IEWC P/N EXRAD2/0-XLEOBS-5 $8.64 foot
    4/0 IEWC P/N EXRAD4/0-XLEOBS-5 $12.70 foot

    Their web site is pretty stupid. You won't find it there or be able to order. Order at mailto:customer.care@iewc.com

    Jack Rickard

  19. Jack, does the Rinehart have any form of EMC certification?

    EMC testing and certification has become a major expense for any form of automotive electrical product. Even passive devices do sometimes experience problems let alone an AC controller.

    Question, to where would you ground the shield?


  20. Hi Jack,

    another source of braided metallic braid is www.techflex.com.


  21. $12.70 a foot? Aluminum foil and heat shrink it is. LOL!

  22. $12.70 per foot would be for the Cadillac for 4/0. We'll use 2AWG for this at $4.21 and I'll need about 4 feet actually.

    No I want to try it.

    We cannot get to the terminals inside the motor. So we are making a small metal box with a large gland nut that will attach to the frame of the car right next to the motor. The shielded cables will connect to terminals on the box and the shields to the box itself.

    Jack Rickard

  23. Remember that you should only land the shield at one end of the wire. If it is the box end, don't terminate the other end of the shield.

  24. Interesting. Rinehart recommends landing it at the box. Why is it important to only connect the shield at one end?


  25. > Why is it important to only connect the shield at one end?

    Because you don't want to create a ground loop (see: http://en.wikipedia.org/wiki/Ground_loop_%28electricity%29)

  26. Mad is correct. The link he supplied explains it. Ground loop will defeat the shielding. I've ran into this problem before where there is noise on a shielded cable and it almost always ends up being that the shield is connected at both ends. Shield should drain to ground at only one place. If a wire terminates and then another piece of wire continues the signal, the second wire should only have one end terminated as well and usually it is at the point where the two wires both connect at the terminal strip.

  27. How will ground loop occur here? Ground loop is a problem when your using the ground as return for the current. In this case all the current is in the 3 wires. In this case the shield is not part of ANY circuit. Analyze your circuit first before you declare "Ground loop".

  28. As I understand it the shield will be come a circuit if both ends are connected to a conducting medium which completes the other side of the "shield" circuit. Any induced current in the shield then will have a circuit to go through defeating the purpose of the shield in this case and why it may have been called a ground loop. Only connecting one end of the shield makes sure this cannot happen leaving at most eddy currents in the shield.

    Here is an excerpt from an email Lee Heart posted on the EVDL about the same thing, "If the coupling problem is severe, do both. Pair all the wires, and twist the pairs together. Put grounded shields over them, and ground the
    shields at only one point (so there is no current in it, which would upset the balanced currents in the wires)."

    He also mentioned some problems with noise and computers. The permalink to the email is http://electric-vehicle-discussion-list.413529.n4.nabble.com/High-voltage-wiring-next-to-low-voltage-wiring-tp3169935p3170667.html

    David D. Nelson

  29. Re Evo motors


    A turgid start but there is a diagram at around 3:30 which illustrates how EVO motors produce such a high specific power.

    Maybe worth considering for the Cobra?

  30. Jack, you mentioned in the video about reversing a controller as a battery charger. I've been playing with this idea of an integrated system on a "H bridge" controller for a few weeks.
    I'm going to be all ears on this matter.
    Lots of chatter on shielding. I'm sure nobody is actually arguing. They just need a common connection...........
    Mr. Hardy
    A tin of beans or the more squashed shape of a Heinz sponge pudding tin holds the maximum volume for a minimum amount of "tin" on diameter vs. length. Something to avoid for an electric motor I suppose.
    Agni (and copies) used this rule to great effect for a small motor with a high output.

    I must say, that motor/controller package seems really nice. I wonder if they exist? The 140 or maybe the 240 package would be lovely in this:

  31. Re EVO - I don't know how advanced they are. They gave me a quote for a 140 motor controller package at around £10k for motor and controller a year or so back. They are a spin out from Imperial College London (one our better engineering schools) with funding from several sources.

    I suspect that one advantage of the "bean tin" form factor and the axial field is better cooling.

  32. Hey Jack, could you post some detailed pictures of your EVSE board ? I'd like to build my own - I'm not an EE, and looking at diagrams gives me a big headache :)

  33. The EVO motor has several advantages. Most have been known for many years. They also have a big problem. RPM. The centrifugal forces tear it apart. I'm told by the good professor that they have some secret sauce with regards to materials, that makes it hold together. Apparently so. The EVO motor was the only thing that did NOT blow up on Racing Green's Pan American highway trek.

    The torque of a motor is a function of current its true. But it's also a function of distance - the distance from the magnetic interaction and the shaft. The further away from the shaft this occurs, the greater the torque for the same current. It's like leverage. The longer the torque arm, the better.

    They further claim that they have a much greater swept area of magnetic interaction. It apparently runs all the way from the shaft to the outer edge. Those two factors give you greater torque in a smaller package. But the shape causes immense centrifugal forces at the edge as RPM goes up. It's great - if you can hold it all together.


  34. I guess the key question is whether or not there is a market for a $15,000 AC motor? If you add $5000-$6000 for an AC controller, you can approach $21,000 or more for an AC drive train. What's that about?

    I can get a pretty tricked out V8 for $6500. I mean pretty deluxe.

    At some economiic level, a solution simply isn't a solution.

    Jack Rickard

  35. Not to mention the crazy voltages these EVO motors require to eek out the performances required.

    Whats 650V about?

  36. 650V is about putting maximum power into play with minimum conductor diameters. Pure and simple. Higher voltages, lower currents, and so finer windings - more windings in less physical space. Lighter motors. There are some huge advantages to doing motors at 650 v rather than 150v. Unfortunately, the same is not true for the overall car and many of the peripheral items that go into making it operational.

    But the common thread in higher powered polyphase induction motors seems to be higher voltages and somewhere around 600-650v seems to be the sweet spot.

    The forklift is simply not a bad model. We have had electric forklifts in continuous production for over 100 years. They tend to be 48v, but a top speed above 15 mph is simply not necessary. So EV's have mostly evolved toward 150v, which uses the heavy forklift motor design but ups the voltage to account for highway RPM's and consequently speeds.

    I see a further movement toward 250v coming in our future pretty broadly. But anything above 300 volts starts to become awkward with long battery strings, problems with peripherals, some increased danger of electrocution, etc.

    Jack Rickard

  37. Not so sure about that, Jack. Motor torque is proportional to amp-turns. Lower current means more turns and higher current means fewer turns for the same static torque. In the end they weigh pretty much the same and the power in is pretty much the same for a given static torque.

    I don't know why higher voltages are better but I'd guess that the higher voltages make it easier to push the desired amps through the reactive components of the motor & controller at high R's. -Klaus

  38. You wrote: "The duty cycle of this square wave indicates how much current the EVSE is able to supply. The charger in the car is then supposed to limit itself to that amount of current. We don't know of any available charger that can do this. As the nature of such specifications are, we would expect this to be ignored for several years, and at some point in the future, someone will "discover" this feature and use it to communicate something else. It is possible the OEM cars actually look for this waveform, but we would be astounded. "

    Well, then prepare to "be astounded" because most of the OEM chargers do pay attention to the Pilot signal. A Tesla Roadster has a charger capable of ~19kW (240V@70A), yet can plug into outlets with 120V@12A, 240V@24A, etc. The Nissan Leaf has a wimpy 3.3kW charger on-board, and will charge at 16A if it sees a high enough pilot signal else it can drop back to 12A. (I think even 10A for Europe.)

    This Pilot signal has been around in the J-spec for a long time. If the vehicle uses an old J-spec socket ("Avcon") or new J1772-2010 it is supposed to pay attention to the pilot signal, and most OEMs do. Home brew EVs not so much.