The Power of Creative Failure and Separating the Players from the Poseurs

I’m delighted to report the first shipment of the first Tesla Model S drive unit with our Alset control kit. It is going to a startup in Oakland California that has big plans for electric boat propulsion – ergo the shipment. They have no need of a Quaif ATB Limited Slip Differential. That program is now well under way and drawings completed with a first manufacturing run scheduled for May of this year.

But you don’t need one for a boat. A boat poses some OTHER challenges, including a LOCKED differential so power is transmitted out one side of the gearbox. It remains to be seen what longterm success that poses for the gearbox. It wasn’t designed to work that way of course.

Boats pose some other challenges which I have outlined carefully for the purchaser. The primary one is that while cars accelerate and decelerate, boats typically run at quite high power levels more or less continuously. This takes them into kind of a different world of thermal management, and as many of you have heard me say numerous times, some days its seems the entire EV task revolves more around thermal management than electrical power.

Fortunately, boating provides its own solution to this problem to some degree. A LAKE makes a fantastic HEAT SINK. And so I have pointed him toward some through-hull heat exchangers where he can pump a lot of coolant through a heat exchanger that actually hangs in the lake water. And so instead of a liquid coolant to air heat exchanger transfer, you have liquid coolant to cool liquid heat exchange which is several ORDERS OF MAGNITUDE more efficient. Probably need a largish fill tank, two through hull heat exchangers and pumps, etc to actually do this. But I think it can be done. The problem of course is that Tesla’s drive unit has the motor and inverter on the SAME coolant loop. This is why Tesla Model S owners who have taken their cars to the track have suffered such universal disappointment. Two laps on a nice track and you’re done for an hour or so. The inverter goes into current limit, and you limp back to the staging area in humiliation. The system works perfectly well for street driving, but when you go into the rigours of track racing, it’s a fail. You just can’t get rid of the motor heat quickly enough to keep the inverter below its 85C threshold.

I don’t actually pretend to know how all that works out on a boat. But I think he’s got a shot with a couple of largish through hull heat exchangers and a largish fill tank. I would use the tank to hold and have THREE pumps. One to pump it through the drive unit. And one for each of two heat exchanger loops. The object is to keep the fluid in the tank below 55C for the inverter.

In any event, we shined up one of the drive units with some aluminum polish and was delighted to see it polish up like a new dime. And as of this morning, it is scheduled for pickup on a pallet and shipment to California.

Along the way, the guy ordering all this wanted to provide his OWN drive unit purchased elsewhere. I don’t really know WHY he wanted that. I suppose he thought he could get it for less money. But he actually bought it from a guy who I buy most of mine from.
I explained that there were some interesting things going on in the Tesla space, and I had no confidence our system would work with all drive units out there. But if he wanted to ship it here, we could work with his drive train, but we had to test it here if we were to be involved at all.

FOR..TUNE….ATE….LY. It’s good to be me. Our system would not make the new drive system peep. It talked CAN. I can communicate with it. But it does not want to go into drive and turn the axles. And with one of ours right next to it, with the same harness and production compoenents, that one DID turn.

So our customer immediately begged to purchase the one I had wanted him to take in the first place. Leaving a bit of a mess. He had gotten a 14 day warranty against arriving DOA. But the seller insists it’s not really DOA. We just don’t know how to work it. Which is a bit circular in that we are the only ones that know how to work ANY of them. The only one that has made a single Tesla Drive Unit turn except us is Michal Elias, by REPLACING the entire control board in the inverter – not a solution we are at all attracted to.

But the supplier takes the bizarre position that it WOULD work in a Tesla so it cannot be DOA. Which is conveniently extremely difficult to prove one way or another. Making the DOA warranty not worth the e-mail its printed on. He knew full well what the customer was doing with the motor and it is indeed DOA for any practical purpose. But that’s rather between them.

The rest is again more or less absurd. There is ZERO market for salvaged Tesla drive units in Teslas. Due to some problems had on early versions of the model, Tesla has had to replace a number of the drive units. This reached such a level of concern, they extended the 8 year 100,000 mile battery warranty to cover the entire drive train. And to shut down the chatter, if you just allude to noises from the back of the car, they now change out the entire drive unit without question really. This ploy has been effective, but for some this has resulted in three or even four drive units in the same car. We have an early 2013 with 10,000 miles on it and so far, not a peep. But as a result, NO ONE would ever put a salvaged drive unit in a Tesla because all operational Teslas are still covered under the drive train warranty. And if they weren’t you would still be better off buying a brand new one from Tesla. So we have this bizarre circular conversation over a DOA warranty, on a unit that is DOA at the moment. I never got such a warranty from the same guy after buying six or eight of these. Of course, it never occurred to me to ask for one. And if he HAD given me one, I’m not sure what it would mean.

It does kind of prove the policy of “we build em here, we test em here, and no you can’t have anything until I SAY it says Haines.” Imagine this guy with an unknown motor, our kit, it not working, and we’re trying to fix this by telephone and e-mail between Cape Girardeau and Oakland. 100 telephone calls and 3000 e-mails won’t fix it, but with $15-$20k invested, he IS going to want to talk on the phone and send e-mail until he can find a bad guy to blame. I’m not signing up for ANY of that. It would be an impossible situation.

Incredibly, the seller contacted me to ask if I would GIVE the guy MY motor, previously bought and paid for from him, and take the lame one in its place. ????? I told him I had already sold the guy the other motor and he was enraged that I “stole” his customer. And so you can see what this all gets into. Minimum of two parties unhappy over what I’m trying to do to ENABLE better vehicle builds in the future. And at this time, salvaged Tesla Model S drive units are worth about $400 for scrap metal if you can’t make them turn. And so far, we are the only entity that has made one turn in public.

And I remain totally convinced that if we can enable YOU to build better, prettier, more effective vehicles, that you and that car will influence many many others to view electric vehicles more positively leading ultimately to a transformation entirely away from fossile fuels and into magnetic drive for ALL personal transportation and really in the grand scheme of things quite soon. Probably another six or seven years but to me that is soon. And the Tesla MOdel S drive unit is the crown jewel of all EV components as far as I am concerned. There’s a couple of things I would do different. And I believe they will. For example, the smaller 170kw Front Drive has TWO cooling loops, one for the motor and one for the inverter. But it IS the best available on-planet.

Which rather begs why I do all this to begin with. Despite the rumors of a cash-crazed EVTV attempting to rape all vehicle builders everywhere for the maximum possible ducats, it has really been no secret that I have more than enough resource to live out all my remaining days playing bridge at the club when the weather doesn’t support a golf game and sipping gin and tonic until I am permanently inured to the vicissitudes of malarial Anopheles gambiae. We price things based on repeatable sustainability and use the reaction as a vote to either redouble the efforts or a pretty good indication that nobody really values or wants it and not to waste time with it. But dealing with those who truly are all about the money can be tiresome. Everyone, it seems, has an axe to grind and how it is viewed becomes entirely a self-serving version of whose ox got gored.

My mission with the Tesla Model S drive train for me personally is to minimize product support issues and avoid disappointing customers to the point that our reputation and trust is diminished. And yes, I’ve seen some potential problems from the beginning.

But I was FLOORED by the reaction to the last video. With three drive units performing flawlessly in front of me, and one that we have never before encountered but does not work RIGHT OUT OF THE BOX, there was an ERUPTION of cheering fans that would do Donald Trump proud GLEEFULLY giggling like school girls that they knew we would fail all along. They were actually hoping and praying for our failure all along and were DELIRIOUSLY happy that “we had failed.”.

Now its true I’m not talking about the EVTV cognoscenti. Not at all. But the dilettante newbie wannabe poseurs on Tesla Motors Club again. I know. I swore I wouldn’t peek again. But the theories, along with the claims, are totally preposterous. Yep, the immobilizer got him. Secret codes. Info from the BMS. High Voltage Interlocks. Limp Mode. The theories rage. The one that finally caused me to understand what I was dealing with in the case of one of these guys that I had kind of accepted to work with a bit, was the insurmountable claim that in fact, the ones I DID have working were working in the of course well known 80kW LIMP MODE and so didn’t REALLY work at all, they just appeared to. Not only did I have one I would NEVER be able to turn, but indeed the ones I was turning WEREN’T REALLY. The code doesn’t lie and in a miraculous 45 minute disassembly of the actual machine language FROM the inverter, he could state unequivocally that I NEVER HAD MADE ONE TURN. It only APPEARED to have turned and you can take it to the bank Rubio, THE CODE DOESN’T LIE.

Unfortunately the poseurs do. Unashamedly.

As it so happens, I HAVE done a little bit of machine code reverse disassembly. Sufficiently successfully that the owner of the code actually took me under his wing, gave me his source code, and indeed we became partners. But it is heroically ugly. It doesn’t happen in forty five minutes or even a day. And you of course DO have to know what processor it runs on or their is no point starting the process. These guys know JUST ENOUGH to convince themselves of their own self aggrandizement, and perhaps fool a few of the truly lost. And I can report they become absolutely ENRAGED if you call them on it.

In any event, reversing machine code back into assembly and making any sense of it IS doable but it is kind of a last resort. And while I DID probably clear a million dollars on my PIMP program for TBBS in its day, being the first mail port using UUCP ever done in assembly language (and the last as far as I know) a million probably isn’t enough to go through all that again. I DID actually meet my wife at a PIMP (Personal Internet Mail Processor) party in the presidential suite of the Weston in Atlanta, so I guess it was worth doing it. But that was what I did to decompress and have FUN when I was publishing Boardwatch Magazine.

And if you can do it between e-mail messages on a processor that you don’t even know what it is or who makes it, you’re beyond a genius. Your a delusional wannabe poseur and really really need some medical supervision and probably ongoing for some time.

But it also laid bare how woefully ignorant of the BASICS of electric vehicles these guys are. With a little game coding skill, they think the car is easy. As you all know, not exactly. First an IMMOBILIZER is a particular reference to a specific thing in automobiles. It refers to the use of an RFID or radio fob to enable you to START the car. I think they believe that it somehow “disables” the “engine”. For the driver, that appears to be the effect, but of course it is a conversation with a receiver that then is linked to the Engine Control Unit (a multicontroller) which then just doesn’t do any of things it must do to START the engine. The engine doesn’t have any part of it. Or in an electric car the inverter. You COULD disable the inverter. It just never would occur to anyone designing such a thing to do so. You tend to hold transactions of that sort in the ECU or in the case of the Tesla, the MCU (the computer with the main console) because that’s where the processing power and acres of memory are located. It simply does not ENABLE the engine or in this case of inverter until it is good and ready to do so and that is suspiciously close to about the time it gets a valid RFID. So you would infer that it does IMMOBILIZE your car. But the engine, the inverter, the tires, the windshield wipers, and the seat heaters all don’t know very much about all that.

Similarly, the concept of an 80kW LIMP MODE is absurd. Of course, like DOA, it is a bit difficult to prove or disprove – which is exactly why it was proposed. I can, if I at all needed to, put a load on a Tesla Drive Unit and test it for power. The problem is testing it for 350kw. I don’t have a 350kw generator. Indeed it is just a dangerous amount of power outside of a car. Indeed, it’s actually a dangerous amount of power IN the car. But even cranking 80kw to 100kw starts to make steam and whistles and the building to shake. So it would take quite a bit to set that up (needlessly) and comfortably it would not disprove the assertion. Of course, you have to question where the assertion came from in the first place.

But the telltale had nothing to do with the code lying nor not lying or there being any code to read. It was the gleeful assertion and the lack of knowledge as to what “limp mode” might refer to.

In an inverter, there is almost always a temperature sensor, and almost all inverters use the same temperature to safety itself. 85C. And at that temperture, the software simply limits the current output of the inverter to prevent the temperature from exceeding 85C. And it will do so right down to NO current if it has to. Properly referred to as current-limit, it is often termed “limp mode” because you can still drive just enough, maybe, to get back to the shop if you take it easy.

We have actually simulated limp modes based on state of charge, to save our batteries. We usually set it up where the car accelerates very slowly to 20 or 25 mph and not beyond. You could “limp” off the freeway and into a gas station or maybe home if you were close enough but even your “daughter” would know something is badly wrong.

Most of the cars we have built have been 80 KW FULL POWER on a good day. And while it is true that the Tesla is a bit heavy at 4600 lbs, an 80kW LIMP MODE would be about an 11 second 0-60mph, equivalent of the production acceleration of the 1957 Porsche Speedster, and with obviously no RPM limit a top speed of 130 mph. In other words, that ain’t limpin. Indeed, I think I could take a Tesla car and put an 80kw motor in it and a 100 mile per charge battery and it would be marketable – at maybe $25 to $30,000.
It would be sedate. But fully functional.

Actually with a 9.73 ratio, I would say 0-60 in nine seconds. So the whole concept is comically unschooled.

Let’s see. IT has to have voltage and current from the BMS. I hope so. We have voltage and current from the ISA and putting it into a BMS message is trivial. I kind of doubt it, but easy enough to do.

Oh yes, HVIL. One guy actually believes he has proven this one with a charger I think it was a wiring coincidence. But I’ve actually BITTEN on this one now TWICE. Otmar proposed it one time and I laughed out loud, but he IS Otmar and so for historical reasons I have a natural tendency to defer to him. So I had to try it. Made no difference at all. I tried it again this week. And again no difference.

The theory is that the device has a 60 ohm resistor in it and the BMS passes a voltage that causes exactly 20 ma of current to pass through a “loop” of all the major units. The charger, or inverter, in theory can then monitor the voltage drop over this resistor and either operate or not operate depending on what it sees.

Totally preposterous. No one would design such a thing. What they DO do often, is safety the high voltage system. In the Coda, they use a Delphi connector that is specifically designed for this. It has a separate little HVIL connector riding on top. And so the battery pack monitors this loop through the chargers, dc to dc converters, inverter, etc to make sure they are all in place and connected before energizing the contactors and connecting high voltage to a cable that might be dangling or otherwise not secured.

Tesla does the same thing, and in the same place. The battery module has contactors in it and it runs an HVIL loop out the battery and through about everything. Some things have this 60 ohm resistor and some not and some have lid switches and some not, but you have to make it entirely around the loop winding up with a known resistance value (beyond a simple short) before it will close the contactors and apply high voltage.

We have a couple pins on the inverter, and a 60 ohm resistor inside. But the inverter doesn’t care anything about it. THE BATTERY DOES. And since we don’t have a battery or contactors there’s no point hooking them up. In any event, on our original motor and again this week, I hooked up a 180 ohm resistor in series with 5volts for 20 ma and of course it has no effect on anything. These guys are reading messages from God in cloud formations, and KNOWING better I go for it anyway. MOstly because it’s simple to check. But it’s just a preposterous concept. It would be designing failure IN to a perfectly operational device on purpose and to no end I can conceive of. High Voltage Interlocks are COMMON PRACTICE and done more or less the same way every time.

So my faith in the better angels of human nature takes another bit of a blow. But beyond that, it is just silly noise. ANd I DO know not to upset the animals in the menagerie or even visit there.

Will we get the other motor working? And what of our other motors and controllers?

I don’t know. First, I’m not sure who the other motor and inverter belong to. I know we’ve changed the seals, the fluids, and polished it up. It isn’t leaving here until that is addressed.

Here’s the working theory. Tesla has indeed updated the software – the firmware – in the inverter to provide some sort of versioning handshake. All units removed from cars wrecked BEFORE that update should be mine. All removed from cars wrecked AFTER that update would appear to have some additional features so far unknown.

If that’s the case, I already have the software. Should be in my car. When we first start the obnoxious motor and inverter, it spews forth 356 frames of 5F4 message id. Encoded in that is an ASCII string describing some sort of four digit code that looks a bit like a month/day thing. It then quits sending that. It sends normal CAN after that. But we cannot make it spin.

On my car, it sends ONE 5F4 frame when FIRST STARTING UP and then a few seconds later ONE 5F6 frame with a similar number. I’m assuming the car returns one of the intervening messages to satisfy this handshake request.

To make it work, we either have to:

1. Figure out what that response is.
2 Figure out how to reflash an earlier version of the firmware into the inverter.

My preference, even if 1 is easy and it looks to be, is still 2. This would protect us from ANY future updates they might want to perform. Some automakers do indeed serialize their major components mostly to defeat your right to repair. While I would be disappointed if Tesla did this too, I would be a little surprised as well But its possible.

If we can download the old version and use UDS or whatever to reflash it into any drive unit, we’d be good.

And what of the D version – the front motor. It uses torque commands – no throttle input.

The answer is I don’t know. And that’s a good thing. You see, when all of this is easy, or when I do know enough about electric vehicles, to satisfy my own curiosity, I’m outta here. Just as I left the Internet. I’m on loan to electric vehicles until it’s future is assured and my intellectual curiosity is satisfied.

But the gleeful cheering and chortling is comical. Up to and including 128-bit RSA, we can break it. It’s about time and resource and how many times we want to reset it and do it again. It really is not unrealistic to set one up and let it chew for 3 months around here in automated fashion. Like mining Bitcoin.

But it won’t need all that. There’s only a certain amount they are going to go through before it causes them more problems than it solves. So I can absolutely count on a pretty meager level of effort to deploy and a similar level of effort to work around. The DMOC645 was a puzzle. The Coda was a puzzle. And this has already been a puzzle. It’s a bit more of one if this theory holds. And there is still the possibility that I was standing on a connector or had a pin pushed back anyway.

As I said, if I knew how it was all going to come out, I wouldn’t bother with it.

If it turns out to be REALLY HARD, I have six or eight drive units laying around here pulled some months back that are going to be REALLY valuable.

This week I have kind of concluded negotiations to have a specific thing built in China that we’ve talked about before. A DC CHAdeMO charge station. It will be under $25,000. Instead of a 3-phase 380v ac power input, it will use DC 150-500 volt power from a battery bank. It will do the full CHAdeMO 50-500v spec output at up to 125 amps (50kw) AND it will include a Maximum Power Point Tracking (MPPT) solar charge function to charge the mother battery bank from solar panels. Would you believe I just have not found anybody who did this and indeed not found anybody that would even consider making me one, that just bypassed the AC input so I could use DC.

These guys have agreed to a very wide input (150-500vdc) and even wider output (50-500vdc). I suppose I could be completely wrong about this and biased because of the $55,000 AmerenUE wants to put 3-phase in my shop. But I think 3phase power is evil.

1. It’s not allowed by code almost anywhere in the United States in a residence. This includes apartment buildings and complexes. You just cannot install it at any price. So those who want fast charge at home just can’t have it. Sorry.

2. When and where you do get 3-phase in the U.S, almost ALL of the utilities are running a special on 3-phase power. A special gottcha that is. While we normally price electricity at a fixed rate applied over your total usage, 3-phase is magically priced at the PEAK rate of your maximum usage, and that is applied to ALL your usage. Fast charging is almost the anti-thesis of this. We charge at a HIGH rate for a short period of time, and then don’t use any at all for hours on end. But whatever we use, will be priced at the rate applying to the peak current load we consume. It makes it quite a bit more expensive – almost prohibitively so.

3. It’s hard to get an installation in rural areas, farms, or out on the Interstate between cities. Guess what? That’s where we NEED fast charging. In town it doesn’t matter so much.

Beyond that, there is an underlying current across EV ville that I’ve been aware of for several years. “I want to drive on sunshine.” Explaining that their grid tied solar feeds INTO the grid, and then it comes back out to run their normal charger and charge their car, and so it really really is driving on sunshine is one level of abstraction too far removed. No, they want to SEE it charging from solar directly. Ok, through a battery bank then.

We are NOT going to kit this up into a total solar/battery charge station. I get it. YOU want to do that part. Ok. We’re going to source and offer the charge station part, full CHadeMO specification. YOU provide the battery bank and solar panels. And, yes, you will be able to generate your own fast charge station anywhere.

And while $25k is dirt cheap for a fast charge station, it’s still a pile of ducats. So I’m not sure there is a huge market for such a thing. But I hear about it regularly and we of course need it here at EVTV. It will let us complete the other end, the CHAdeMO kit for conversion cars. But it WOULD allow the well heeled Tesla owner with solar plant to spin his own home fast charge system.

So I’m kind of excited about it. The company uses the usual stack of AC powered chargers in their normal offering. Which is why I usually can’t get anybody to talk about bypassing the AC rectifier. They can’t. It’s in the modules. So this is kind of a total redesign. But it’s a company in China that does these things anyway, does solar battery charging, and has decided I might have put them onto a possible seam in the zone here. So it is going well so far. As it is developmental, I always have the option of snatching defeat from the jaws of victory. But I’m holding a good thought on this one. We have some pretty good communication going here.

With three drive units in front of me running, and one that doesn’t, I suppose I can accept that we have failed. But that is just a very different thing for me. Whevever I succeed at something I always wonder what parts i could have left out and still succeeded. In failure, sooner or later I always know exactly why. For 50 years I’ve kept a book of sex education – every time I got screwed I wrote it down. Guess which experience provides the best material for the book?

So creative failure is the source of knowledge for me, and success, while vaguely pleasant, offers little for the future and is inevitably short lived. This has gone on for so long that on the prospect of failure, I go into a kind of excited stiffie, and faced with the imminent prospect of success, my mind starts to wander toward other projects.

Similarly, we are anxious for any help from the gallery. Its not so much we need the help, but we enjoy the company of talented people in common cause. A reassurance that we are not the only savages on the buffalo hunt.

But if you are looking for self aggrandizement, have a secret project that is going to rain all the riches of Croatius on you and your descendants for a thousand years if you can just keep it a secret (while getting it exposed on a video publication) I can’t help you. And you bore the shit out of me almost on contact. Wannabe poseur hackers need not apply.

If you are just interested, curious, and technically inclined, I’ll actually stop what I’m doing and its all about you as far as I’m concerned. What part of all this do you want to talk about?

Did I tell you the part about changing the world? Forever?

Jack Rickard