All Your Drive Unit Are Belong to Us.

rickard

The whole is often greater than the sum of the parts. Like many technoids, I tend to live a bit isolated and removed from other humanoids. Yet over and over learn the same lesson. We are what those around us make of us.

Just a few weeks ago, Mark Weisheimer of Columbus Ohio and Byron Izenbaard of Kalamazoo Michigan very generously travelled to Cape Girardeau to address the tragic neglect leading to the demise of our Azure Dynamics eTransit Connect. Byron was particularly fascinated with this vehicle and rather talked Mark into launching a rescue mission.

With the attention span of a four-year-old, I was intently focused on other things, but ok, come on down. How fortuitous.

They were able to quickly address the problem and even performed conformal/potting compound surgery on the wake-on-charge module of the vehicle to get it up and running. But there was quite a bit of further fallout to come of it.

First, Mark on his return sent me a link to a Tesla Drive Unit offered for sale on eBay. After 30 seconds of considering the eggregious cost, endless difficulties, and total uselessness of a drive train that no one could use for any imagined purpose beyond boat anchor, I grabbed it anyway. Incredibly, within a couple of days I stumbled on another and grabbed that one too. Two totally useless 400 pound boat anchors. Jackpot.

warrantytape

Byron, having now not only fixed an eTransit Connect but driven it, returned home with eTransit lust in his heart. I told him I had found several Ford dealers who had had these things for several years without a sale, and then through the same cunning neglect technique I had used to disable mine, disabled theirs. Now what is a Ford dealer to DO with a disabled eTransit Connect? First, Azure, who was supposed to support the warranty work, had gone bankrupt. The dealer typically had $50,000 invested in a vehicle that was supposed to retail for $58750. Now it wouldn’t work at all. How could they sell it, with the Ford logo emblazoned across the front, as a new vehicle, while explaining to the customer that it didn’t run and the Ford dealer couldn’t fix it, and couldn’t tell them where or how to fix it?

This kind of outcome is NOT how to get auto dealerships across the land to become enthusiastic over electric cars.

In any event, within hours, Byron had located one and arranged a purchase at $6500 – somewhat off sticker price. And after a bit of fritzing around with it, got it on the road.

Byron is young and was completely flushed with the concept of the “hack team” and indeed we had gotten the errant Chevy Volt charger and the Coda UQM drive train to work. Even the Ebenspacher heater more or less heating. So he began talking up a “hack team meeting.” He actually created an online survey of what dates might work for whom. And without really any input from us they decided to just have one on the weekend of June 12-14.

I’m normally a bit jealous of my weekends. You know, boating, camping, hiking, time with the family…. err…ok…in your dreams. Actually none of those. We shoot on Friday, edit all day Saturday, and upload on Sunday. A “weekend” is time off work for Byron. For EVTV, that IS the main part of the workweek and indeed there is no EVTV without it. But ok, cool, we will reroute all packets around the damage.

Then Collin Kidder made a very strange, actually an outrageous pronouncement. He believed he could get the Tesla Drive Unit to turn in a single day.

Collin has actually joined EVTV as a kind of part time remote staffer and we pay him a modest salary to fix my coding screwups. This has apparently swollen his confidence beyond all earthly bounds or so it sounded. Actually, he had been working on another project – a CAN analysis tool he calls SavvyCAN. SavvyCAN is interesting from a number of angles. First, he’s used a development platform termed QT to compile the program separately for Windoze, Linux, and Mac OSX. Better, it seems to work. But the program allows you to capture and send CAN traffic in bulk. More better, it allows you to do some analysis on the messages that just goes way beyond what I can do with my admittedly rudimentary Excel skills.

We’ve also added a new full time staffer to replace Brian Noto. Bob Wilson drove from San Jose to join us. As we wanted to play with it anyway, we set out to mount one of the drive units on a brand new maple test bench. We used one of the better place battery packs as a power supply, and wired up all the connectors on the drive unit. Game on.

We quickly found from the service.teslamotors.com web site schematics that the inverter on the drive unit actually had a hardwired connection to a dual output accelerator. One of the hack team members offered that the part number of his Tesla throttle. We searched our friend eBay and found that that part number ALSO corresponded directly to a 2012 Ford Fusion hall effect accelerator. So we got one of those.

Connectors have been the bane of my existence for forty years now. But incredibly, the control input connector was a 23 pin Ampseal that looked oddly like the ones we use with the Brusa chargers. Actually, its the SAME one. The six pin on the Ford fusion was a little harder, but showed up by part number on the Tesla web site connector cross index. The four pin connector on the motor encoder did as well, but wherever we searched for this connector part number, it was either unobtainium, or in a couple cases we COULD get it with a 13 week lead time and a minimum order quantity of 20,000 units.

Byron again came to the rescue. He had a four pin connector on his desk at work that looked remarkably similar. He made some measurements, and offered that it might work. I ordered some, and it fit as if it were designed for that. It wasn’t. Entirely different manufacturer and part number. It’s still better to be lucky than good…again.

[jwplayer streamer=”rtmp://s3einxnpkaij93.cloudfront.net/cfx/st/” provider=”rtmp” file=”news061215-iPhone.mp4″ hd.file=”news061215-1280.mp4″ image=”http://media3.ev-tv.me/news061215.jpg”  width=”850″ height=”522″ html5_file=”http://media3.ev-tv.me/news061215-1280.mp4″]

So we kind of raced through building a wiring harness for the drive unit, and a control panel to do the precharge and 12v and so forth to light up the drive unit.

We DID have one bit of an advantage. I had a Tesla Model S you can’t see in the picture because it is right BEHIND the camera in the shop. And indeed, in ANOTHER bit of connector luck, we had located the mating connector to the Tesla diagnostics port, located under the center console. And so we had developed a kind of finished version of our CANDue CAN product specifically for the Tesla Model S. http://store.evtv.me/proddetail.php?prod=TeslaCANKit

And so we were able rather easily to capture traffic from the “Drive Train CAN Bus” on a live car.

As it turns out, yes, the throttle signal is hardwired into the inverter. And the brake signal is as well. But all the driving modes of the vehicle are really funcitons of the inverter. You select them on a touch screen, and that is sent as CAN control messages to the inverter. Sounds simple, but there are a lot of modes. CREEP ON and OFF, TRACTION CONTROL ON and OFF. STANDARD and RANGE modes to alter the power curve. STANDARD and LOW modes for regenerative breaking. There is even a PARKING BRAKE on/off selection in the vehicle. And of course the usual PARK, REVERSE, NEUTRAL, and DRIVE.

Broadly, the way it works is that when you move the gear shift from PARK to DRIVE, a CAN signal is sent to the inverter. It puts itself into DRIVE mode, but to confirm, it also sends out a CAN message that effectively changes the instrument display so that P goes out and D lights up.

But it didn’t give up easily. First, in all our SavvyCAN testing we had been kind of preoccupied with how fast the program can capture and record CAN messages. The Tesla spits out some 1500 frames per second. And we can’t drop ANY frames or all is chaos. But we had rather failed to really ever test how fast it SENT frames. SavvyCAN has a function to send a whole captured or written file out on the port. But alas, when it came time to do it, we were pumping a heart pounding 7 frames per second OUT.

Fortunately Collin was here for the event. He quickly kluged a blunt force fix and we able to do the 1500 frames per second. Still no joy.

We all hold Tesla somewhat in awe. Some of the written descriptions on online forums such as Tesla Motors Club actually become bizarre. Because of the cellular network linkage to the car, they ascribe a god-like omnipotence to Tesla. And that all of Tesla’s technology is a kind of other-world level of sophistication. There are many TEsla owners who wouldn’t dare to connect anything to the diagnostics port as they are certain Tesla is “watching” them and would instantly notify them that their warranty is void if they took a peek into the workings of the car.

It is not in Tesla’s commercial interest, obviously, to dissuade anyone from this view. But it does vary somewhat from reality. In even discussing using the Tesla drive unit, we received dozens of e-mails from well intended but largely unwashed viewers of all the reasons it wouldn’t work and all the evil that would befall us. You would think the drive unit was some sort of Pandora’s box and they feared not only consequences for EVTV, but indeed the entire world might be punished for religious infractions. The wrath of Tesla.

Even within the hack team members on sight, there was discussion of serial number verification, encryption, and various strategies Tesla might have used to prevent us from using their electric motor.

One knowledgeable offering was that the high voltage interlock circuit used some sort of magic 20 ma current loop and the inverter monitored this precisely. If it varied at all the inverter would not work. With some familiarity with the source, I proclaimed this bullshit but at the same time made provisions for providing it. Indeed, like ALL the OEM cars, there is an HVIL loop that basically removes high voltage if anyone has a cover off or a unit removed during maintenance. This is to prevent technicians from shocking themselves while working on the car actually. But the 20ma internal measurement within the Inverter was a bizarre concept. And while we carefully provided it, it had no effect at all, on or off. Its simply a loop through the inverter, monitored by some circuitry in the contactor box, to make sure it was physically connected.

Philisophically, it just makes no logical sense to me that Tesla would do ANYTHING to prevent us from using a drive unit out of a wrecked Tesla. What Tesla doesn’t want is any expense to support that. It’s an uncontrolled situation with too many questions. They can’t provide technical support for such a thing. But I cannot imagine them actually going to any expense to prevent it either. What happens to your trash after it goes to the dump? You don’t want to keep paying for it. But beyond that….

I believe their SECURITY concerns are properly focused on the car/planet interface – the cellular gateway. You don’t want OTHER people to access your Tesla car. But you want to be able too.

Indeed, just like the Chevy Volt charger, at some point where production constraints go away, I see Tesla just selling the drive units over the counter. Get online and order one. Batta boom batta bing.

Of course, I could be wrong. Once an organziation reaches a certain size, their logical line of reasoning engaged in their own self interest becomes flawed. But there is no magic sauce, no mystical technology, no magic. It’s all just software.

No the stopper was much more mundane and made a lot more sense. In my dotage, I have already pulled two J1772 charge stations OFF the wall and into the floor, and one J1772 charge port OUT of a car by the roots. I was moved to rig the yellow VW Thing where it couldn’t drive while plugged in. Pity I didn’t do the same thing to the Escalade.

Tesla does have a hardwired input from the chargers to the inverter. When you plug your Tesla into the wall, the inverter is aware of this and will not allow you to turn the motor under any cirmumstances. This is an input and with 12v applied to the inverter, we measured 3.3v on this pin with nothing connected to it. That looks like/sounds like the Vcc for a multiprocessor. And quite commonly, you will put a “pullup resistor” from such an input to the 3.3v power supply. This draws basically zero current, but the 3.3v is felt on the pin.

Let’s say that this 3.3v represents a logic 1. How do we make it go to zero. If we attach a ground to the pin, we have completed the circuit from the power supply, through the pullup resistor to ground. But the input pin reads 0v. And so with no ground, we weren’t turning. That tells me that an open would be read as an INHIBIT from the chargers. And a ground, conversely, would be an ENABLE. We grounded the pin, and the motor jumped to life.

This is not precisely the end of the story. We can turn the motor by playing back recordings of Tesla Model S CAN traffic. Ok, kind of limited known recordings. But there remains much work to work out what each message does. We start by tossing out messages one at a time until the motor won’t turn. Those messages are obviously not needed. And then we have to decode the data in the messages that DO count. Non trivial.

But all that is doable. What was proven at this point is that it can be done. There are no “show stoppers”.

Our Generalized Electric Vehicle Control Unit is of course our weapon of choice. But it was kind of designed to take accelerator inputs and convert them to CAN torque commands. In Tesla drive unit case, the throttle is hardwired. The CAN just represents modes. And those are driver selectable. So what we REALLY need is a human interface to enable the driver to select the mode, and the proper logic to prevent one mode from eliminating another (we can’t be in DRIVE and REVERSE at the same time). And then some actually pretty simple CAN to convey those modes to the inverter. That looks like/sounds like a different device entirely. A touch screen interface much like that in the Tesla Model S. Only perhaps 5.5 inches across rather than 17 inches.

And so our goal is pretty simple, we want to develop and interace/controller and wiring harness anyone can connect to a Tesla drive train, and then install the whole into any vehicle they want to drive.

It’s an exciting project. Why? Tesla is simply the top of the pecking order. They have the best cars and the highest performance. The P85, which both of the drive trains we have are and which is in my personal Model S, is a 416 hp (310 kW) and 443 ft·lb (600 N·m) drive train capable of 16,000 rpm. It includes a 9.73:1 reduction through the gear drive/differential. Our UQM Powerphase 100 and Siemens/DMOC645 offerings, by contrast are just over 100 kW, or 1/3 this power.

In Byron’s eTransit Connect, the Siemens/DMOC645 system really seems perky – nice drive feel. That is a 5005 pound vehicle. Understand that when you put that in a 2400 lb VW Thing, with a multispeed transmission, the perceived performance is MORE THAN DOUBLED. And so you can see why I have been so enthusiastic about these drive trains. I know what they feel like.

Damien has put this same power plant in his heavier BMW E31 Der Panzer. And indeed, as it so happens, in this video and on the same weekend that we made the Tesla Drive Unit spin, Damien did his FIRST rollout of Der Panzer under its own power. This car is listed as 1830 kg or 4030 pounds at the curb. Still less than the eTransit Connect by 1000 lbs. So while I don’t think you’ll see Damien on the drag strip with this, I think he will find it MOST pleasant and responsive to drive. Again, he has a multispeed automatic transmission to leverage that power to good effect. The eTransit Connect is limited with its single speed gearbox.

But as Tim Allen was wont to say, – we need more power. We’ll rewire that sucker. A Tesla drive train then becomes the highwater mark in what you could power a custom vehicle with if we can find a way to make it available.

It appears we can find a way. In my world, that means bigger badder more gorgeous custom electric cars.

We were not the only ones working on this problem as it turns out. We were just the first to pull it off. And not by much it appears.

Michal Elias, originally of Pardubice in the Czech Republic, currently resides in Geneva Switzerland and works at CERN as an RF design engineer. He sports a master’s degree (electronic engineering) in robotics and cybernetics, with a bachelor degree in measurement, from the Czech Technical University in Prague. He announced first spin of the Tesla Drive Unit ONE DAY later on Sunday. Probably would have won the race except for a parking pawl issue he didn’t take into account.

Elias took a different approach that I don’t actually think will be very useful for using Tesla Drive Trains. Instead of controlling it with CAN mode commands, he bypasses the inverter control logic entirely. But it becomes interesting in and of itself for other reasons as you will see.

We’ve been corresponding with Elias for about six months. He’s basically developed an AC motor controller board that can be married with almost any power switching device to drive ANY AC motor. His claim is that it can LEARN the important motor parameters without much intervention from the operator. It runs a series of automated tests carefully measuring votlage and current reactions on each phase. To demonstrate this, he wired it into a Chevrolet Volt inverter. He literally wires it into the gatedriver circuitry of the power electronics, bypassing the Volt control circuitry, and so can turn a Volt drive train.

He assured us he was waiting delivery of the latest version of his printed circuit board design and would send us one as soon as they came in. We’ve actually procured a Volt inverter and were sitting here waiting on it. I guess other things became more important as we never received our board. (This happens a LOT at EVTV. We’ve found the quickest way to get rid of somebody is to tell them YES).

We had originally located a Mercedes Benz B Electric prototype in Germany where it was for sale quite expensively. We pointed it out to Damien Maguire who talked about buying it. Or, I don’t quite recall, all that MIGHT be entirely in reverse. Maybe Damien told us about it. IN any event, it is actually a different motor, different firmware, and we wanted nothing to do with it – shipping costs transatlantic being a big part of that. In any event, it appears Michel bought it and he claims VERY inexpensively (Take that Damien – lash/lash). It is definitely NOT a Tesla Model S motor, but was of course manufactured by Tesla specifically for Mercedes and IS quite similar.

In any event, Michal got it and actually disassembled the Inverter side anyway, which is educational. We learned, for example, that the Inverter is liquid cooled as well as the motor in kind of a cunning coolant routing scheme. The power electronics are quite compact. And indeed, that is our main problem wiht the whoel approach. The Tesla Model S drive train is remarkably compact. The entire unit is 32/12 inches long. Ok 34 if you count all the mounting tabs. In any event, what we admire most is that much power in that compact and tightly integrated a package. Disassembling and reconfiguring it somehow otherwise was not on our agenda.

But Michal’s approach is really almost exactly what Damien did to the DMOC645. He replaced the usual controller board with an open source one. In this way, you do not have to reverse engineer the control interface. Just make up your own. And this is what Michal did to the Tesla Drive Unit.

Rather a bunch of wiring, but you get the idea. And so Sunday morning, less than a full day after we turned our Drive Unit, Michal turned his – slowly but in a very controlled fashion.

Long term, we are actually hoping to carry the Elias controller board as one of our products. Assuming of course that it can “learn” motors. We were actually more impressed that he got the Volt inverter to run the Tesla motor, than that he got the TEsla inverter to do so. That means a Volt inverter could conceivably turn a Siemens motor, or a UQM motor. More impressively, the Volt actually has TWO 110kw inverters in it. How about one very compact power unit to drive TWO UQM motors or Siemens motors.

Why not such a good idea with the Tesla Drive Unit? The Tesla unit is a marvel of a LOT of power in a very cunningly designed very compact integrated design. That is its main feature. Marrying in a largish controller board and/or breaking it into component pieces just isn’t on our radar screen.

And again, one of our themes going forward is to use OEM components specifically so you can just drop in a new Tesla Drive Unit in place of the failed one, and obtain one anywhere they are available. Which hopefully will be anywhere. We want to minimize and isolate the additional hardware and software to marry it into the vehicle. Little low voltage CAN control boxes don’t actually burn up. Power electronics do. CAN provides great isolation between the two.

So, different approach but a VERY interesting development and we will continue to follow this in the future. I would be tempted to put a pair of them on a dual Volt inverter and drive to motors from the same box.

We also had a bit of a brainstorming session about future projects for the EVTV Hack Team. One of the things Elon Musk announced at his Annual Shareholders meeting, other than the fact that he would certainly consider eliminating leather seats in order to save the cows, was the ongoing buildout of the Tesla Supercharger network. And part of that in 2016 includes a Supercharge station in Cape Girardeau Missouri it would appear. We may very well have a Supercharger here before we get a CHAdeMO fast charge station.

Wouldn’t it be nice to know how to drive coast to coast on the Tesla Supercharger highway – using your 1960 Studebaker Champ pickup truck????

And what of our EVTV hack team. What motivates all these people from such diverse geographies to freely give of their time and treasure to these little projects with no hope of real financial gain?

They are a really good group of guys, thats what. I would feel deep guilt in taking advantage of them except for a little something I’ve learned over the years….

Elon Musk was talking in one of his many interviews on the topic of Tesla employees. He was actually at MIT and of course the MIT interviewer, knowing Elon himself was a Stanford Alumni, was trying to evoke an endorsement that if you wanted to work at Tesla, a degree from such a top tier engineering powerhouse like MIT or Stanford would be just the ticket. He was visibly shaken and indeed angered at Musk’s reply. “We not only don’t care if you graduated from MIT with an engineering degree – we don’t care if you’ve ever attended university at all. We look for people who can do things. And we can tell pretty quickly if they were the ones doing it, or one of those watching it be done.”

Our society and our economy is quickly moving to a two tier system. The haves and the have nots. But it isn’t quite what the liberal elite would have you believe. It has nothing to do with Wall Street, Bankers, or who HAS the wealth and who doesn’t have the wealth. We are quickly moving to a society delineated by technical competence. If you go to university for a degree in history or English literature, you’re going to wind up having difficulty paying off your student loan with a minimum wage job at Sunglass Hut, no matter what they determine the minimum wage should actually be.

And if you actually CAN do stuff technically, it’s kind of hard to hold a job. That’s because within just a few years your stock options normally cause it to make little sense to keep going to work every day for a mere salary. Even if it is in the top 1%. In Tesla’s Annual Shareholder Meeting, they also announced that their Chief Financial Officer, Deepak Ahuja, was leaving the firm. No, the company isn’t in trouble and no, they haven’t had a falling out. But as a multimillionaire now many times over, Mr. Ahuja does have a bucket list to work off and he isnt’ getting any younger at age 51. A kind of once removed relative of mine had the same thing happen at Facebook – at age 31.

Bill Gates dropped out of Harvard almost immediately. Steve Jobs audited a single caligraphy class at a school his parents sent him to, but which he never actually attended. Zuckenberg dropped out of Harvard. Indeed the list of people DOING things in the tech world, starts to look like an ANTIDOTE to a university education. And in the land of Venture Capital and Silicon Valley, they have quickly learned not to rely on University credentials when investing money. Ok. It’s true a PhD doesn’t HURT your chances. They don’t actually hold it against you – for the most part. But it doesn’t do what you would think.

Mark Wiesheimer and I and several others are, ahem….well lets just say “past our sell-by date.” But as we have ongoing pressing projects at EVTV, PLEASE don’t mention to young people such as Collin Kidder, or Paulo Alameida or Byron Izenbaard, that if they ever did show up at Tesla Motors explaining that they reverse engineered the Tesla Drive Unit using hardware they built by hand and a CAN software library they coded from scratch for an Atmel SAM3X Arduino processor, I happen to know they would not be allowed to leave the property alive without accepting a job offer and signing a non-disclosure non-compete. They’ll just send you down hall after hall, turn after turn, looking for the exit out of the building until you give up exhausted and agree to serve. It’s just how the Valley works these days. And as to your degree certificate, “Ah yes. Pretty. Very nice. I’m sure your parents are very proud.”

Moving in to the future, you are either in the tech industry or unemployed. I’m FASCINATED at the moment by Walmart. After years of criticism for their paying low wages, and providign LOTS of jobs to people just starting out, as well as geriatric pensioners looking just to augment their income and maybe get out of the house and mix with people in society, Walmart is throwing in the towel. Faced with an unknown and apparently unknowable obligation mandated by Obammacare, the very real threat of increases in the minimum wage, constant criticism in the press, they have loudly announced that they have had a change of heart and are going to pay EVERYBODY $16 per hour.

Isn’t that great. I was in a new NEIGHBORHOOD Walmart this past month and was amazed find NINE SELF CHECKOUT LANES and a SINGLE checkout lane manned by a checkout boy (supervisor actually). I went to SAMS club for my Saturday lunch visiting the 12 wizened hags who man the little plug in fryers with their bacon bits and sausage samples. I was shocked to see 10 left but two BRAND NEW machines that offer free samples for a swipe of your SAMS CLUB card. No greeters either.

And then my brother reports that FIVE supercenters across the country all closed down simultaneously at exactly 2:00 PM one afternoon last month with NO notice to employees. They all have IDENTICAL “plumbing problems” and intend to reopen in two to three months. Care to wager how many “employees” get called back? Within two years you will select your products and put them into the bag on your cart, swiping them with your smart phone, and wave your smart phone at a swiping station on your way out the door. There will be about six $16 “supervisors” in the whole store to handle any “problems.”

The unholy triumvirate of a government gone crazed with power to regulate businesses, allied with employees drunk with a sense of entitlement, is a powerful motivator to replace them all with unregulated robotics and technology. This is REALLY what’s happening to “jobs” in America. The Chinese havent’ taken all our manufacturing. We are manufacturing more than we ever have manufactured. There are just no people actually IN the buildings where that goes on. If you produce technology that feeds that need, you prosper mightily. And if you don’t you are resigned to an ever shrinking labor force of people who take in each other’s wash – at minimum wage.

The other strategy is outsourcing. In the past 40 years we have grown from 5 million to 30 million “companies”. Each have a tiny handful of employees who make very low wages, and one owner who does pretty well. And there is plenty of business from large corporations who used to do that in house, but now outsource it.

Both trends are simply long term solutions to dealing with a government power crazed and tax drunk on making “employers” do the dirty work government doesn’t want to do, while keeping the cash flow the goverhment wants to have.

In any event, we are inordinately pleased with the outcome this weekend. My many thanks to the Hack TEam members who went to the distress of attending and to those contributing valuable ideas who did not. Congratulations to both Damien Maguire and Michal Elias on their simultaneous breakthrough events.

Exciting times. Stay with us.

jack rickard

hackteam

Infinite Mile Warranty

The Tesla Model S drive unit warranty has been increased to match that of the battery pack. That means the 85 kWh Model S, our most popular model by far, now has an 8 year, infinite mile warranty on both the battery pack and drive unit. There is also no limit on the number of owners during the warranty period.

Moreover, the warranty extension will apply retroactively to all Model S vehicles ever produced. In hindsight, this should have been our policy from the beginning of the Model S program. If we truly believe that electric motors are fundamentally more reliable than gasoline engines, with far fewer moving parts and no oily residue or combustion byproducts to gum up the works, then our warranty policy should reflect that.

To investors in Tesla, I must acknowledge that this will have a moderately negative effect on Tesla earnings in the short term, as our warranty reserves will necessarily have to increase above current levels. This is amplified by the fact that we are doing so retroactively, not just for new customers. However, by doing the right thing for Tesla vehicle owners at this early stage of our company, I am confident that it will work out well in the long term.

Does this mean my drive unit is still under warranty? Even though I bought it out of a wreck? I AM still a MOdel S owner.

FINAL NOTE

The two VEGAN Shareholder proposals brought by “shareholders” at the Tesla Annual Shareholder Meeting both lost.

Mr. Mark Peters Proposal Number 1 went down 70,258,210 against to 614,771 FOR.

Ms. Elizabeth Farrel Peters Proposal 2 went down 66,998,279 AGAINST to 531,462 FOR.

As it turns out the two are married. And an even GREAT coincidence, they are married to EACH OTHER. Imagine their surprise then when they EACH found the OTHER at the event with an almost identical proposal. Surprise surprise! If only they had known. They might have found a way to resolve the differences in their proposals to present a single unified front. Or was that the point? Maximum face time. Maximum self aggrandizement. Maximum self importance.

Hillerary Clinton has vowed to get to the bottom of it and find out why Tesla paid MR. Peters 614,771 votes while MRS. Peters only got 531,462 votes – for essentially the same work. That’s just 86 cents on the dollar – vote wise.

Of course MR. Peters is the majority shareholder – in his house in Hurst Texas at least – with 1540 shares. His wife had a scant 250 shares to vote.

And the lesson here is that if you are going to advocate something, be sure and do it with OTHER PEOPLE’S MONEY.

89 thoughts on “All Your Drive Unit Are Belong to Us.”

  1. Re: Tesla First Turn… I am in awe.

    Re: Michal Elias… His approach is achieving one of two goals the EV aftermarket community still needs: A standardized controller that manages existing drive units. (The second major need is the ability to better mold an arbitrary battery pack to the voltage/current needs of a particular drive unit).

    1. Thus our interest in the device. Normally, an inverter and motor more or less have to be designed for each other. The MES-DEA system in the Mini Coooper did have some automatic tuning procedures, but you were very much involved in running them. Elias claims his device does this rather automatically.

      Actually, there are numerous power electronics modules available on the market now pretty much in finished condition. The problem has always been an FOC control for them. Imagine a day when you can choose from a variety of motors, a variety of power switches and use them in any combination with a single control board. That would be cool.

      Because of its tightly integrated mechanical design, I don’t think this is a good solution for the Tesla drive unit. I suppose it would offer some flexibility in “hot rodding it” a bit powerwise if you can generate the current (already 1200 Arms for the P85). But I don’t quite see how to put all that together mechanically.

      But for the Volt/Leaf/BMW/VW cornucopia, this might really have some use – if it does all he says it does and assuming you don’t need a Masters degree in cybernetics to install it.

      Jack Rickard

  2. Is this the wrong time to try to crowdsource for a plane ticket to Palo Alto? I hear there are some cool places to visit there. 😉

    Interestingly, while I didn’t expect Tesla to put a supercharger near me they most certainly are going to do just that and even this year yet (supposedly… Look at their 2015 super charger map). It might be because there is a Tesla company presence in Grand Rapids now (a stamping plant.) So, if we get to hacking the super charger network I’ll be able to go test it directly here in town. By 2016 there are supposed to be superchargers all over Michigan.

    Indeed, it was an exciting weekend – lots of breakthroughs and progress. The future is looking to be very exciting.

    1. During the hack team toast all that was missing was the passing out of cheap-sun-glasses first…….cause the future’s so bright we gotta wear shades.

    2. This runs in to a grey area for me. I have absolutely no problem hacking the components in the car, but the supercharger is another matter. Tesla owners pay to gain access to the network ($2000 or $2500 I can’t remember which since it is now included in all cars). Tesla has offered to make it available to other manufacturers for that same fee per car. If we hack the protocol, it should be something that is kept private and we (the EVTV team) should see if we can take up his offer to license the access to the network. It would be awesome to supercharge my DIY EV, but not if I’m stealing the electricity which is what it would amount to without paying their access fee.

      1. That is a very interesting twist on the “politics” of charging protocols. By becoming a licensed SC car (or alternately a licensed charge controller) the adoption of the SC protocol is extended, which may be very attractive to Tesla. Create an add-on charge controller (licensed) that has a SC “in” and (J1772 | CHAdeMO | etc.) “out” and you then have access to a much larger set of charging stations. With this new, more capable energy source, the charge controller can deliver higher currents to the batteries (which as Damien encountered, doesn’t always happen with all charge stations). Improved charging performance, increased station availability, extended charging compatibility.

        Tesla then somewhat co-ops charging of their competitor’s autos, and shows the world that the SC protocol should be the manufacturer preference.

    3. If I got it correct, you have to buy a ticket to be allowed using the supercharger. This could be an option when buying your Tesla or for other cars they could make a deal with Tesla.

      Tesla has excluded Wrecklas from the supercharger until they have bought a red ribbon.

      This could be done in the car or the charger. Most likely it has to do with the VIN number. Of coarse we could fake that, but would it be polite?

      EVTV is Teslas most prominent public relations team. Maybe there is a red ribbon for free? Maybe this is the right time to ask.

      They who have opened the CANBUS and the drive train are able to open the supercharger network as well. There is no doubt

      CHAdeMO or Tesla, that is the question here.

      Cheers
      Peter and Karin

      1. Hehe, I have been to Vegas twice in the past but sadly it doesn’t appear I’ll be there this August. I knew that there will be a Model S hacking presentation. I assume that it’ll center around the usual access vectors – WiFi, ethernet, and the center console itself. CANBus attacks are pretty rare because the average security hacker type doesn’t know anything about canbus. It’d be dangerous if I went to DefCon anyway. They’re looking to hire security people you know. 😉

        I fear that all this publicity about security in cars will lead Tesla and others to make some actual attempt to lock out their canbus networks. Really, I have no idea why they don’t. I could easily secure the whole car within a couple of months and nobody would ever figure out the comm protocol again. But yet, car after car is produced with the network details pretty much out in the open. My best guess is that they want to be able to use Vector Canalyzer to debug the car and it’d be a pain in the ass if the comm protocol weren’t easy to define with a DBC file. This might be a bit of myopia brought on by access to the big name commercial tools. They could write a custom program to do the canbus debugging but they likely don’t because CANalyzer exists. So, I guess that’s to our benefit.

        1. Very noble of you to avoid temptation for our continued benefit 😉 No doubt somthing from the OEMs will come along eventually and dick things up in the name of “safety” (read: limiting liability), but at least we can make hay while the sun’s shining with the components out there today thanks to you whitehats.

  3. Jarkko Santala

    To get SuperCharger working you’d probably need to fake a known good VIN. We know that Tesla can ban wrecked Teslas from charging by VIN. Or some other unique identifier. This would probably make the VIN/ID banned pretty fast. And an annoyed owner somewhere. Unless it’s Jack’s Tesla that gets the cold shoulder. Or maybe it will just work and Elon will applaud you. Who knows.

    1. I have no evidence that Tesla has ever locked anyone out of a Supercharger or has the convenient ability to do so. You kind of have to buy in when you buy the car to get the hardware in the car to do this, and the 60kW originally didn’t have it. The “lock out” encountered by the gentleman who restored a salvage car himself was to run the car at all. And Tesla does have access to the car through the cellular network and I think it a heinous crime that they did so or even that they CAN do so. This should be illegal.

      The Supercharger protocol does send a VIN to the Supercharger. Not sure what the impact of that is. But I would assume you have to have a good one to charge. I do.

      Beyond that, the dynamics of all this are a little tricky. If I approach Tesla and want to sell a “kit” to allow other cars to charge on the Supercharger network, they are going to be very gunshy of the whole idea even if they didn’t already have their own asshole so puckered up that you couldn’t drive a nail into it with a sledgehammer. How do you do a supercharger interface for unknown cars with unknown batteries? This makes no sense in their world if they WANTED to do it. The concept is untenable on the face of it.

      On the other hand, if we reverse engineered it and got it working on a couple of cars, that certainly isn’t Tesla’s fault. If we approach them then with an offer to include a license fee in OUR already working kit, representing the $2000 or whatever they estimate their costs for electricity is, most of the risk is ours and most of the money is theirs. A pseudo VIN and a piece of inlet hardware is litttle enough to ask.

      They would probably want some sort of technical review of what we had done to prove we couldn’t dick up a supercharger.

      What I would imagine they will really do is cheerfully agree to talk about all that, review what we did, and then tell us it just doesn’t fit in their program. They will then modify the system where it wouldn’t work any more with more encryption. It has gradually evolved to be a bitter vicious world in corporate America. I’m not really new to that actually.

      But you don’t get if you don’t ask. I choose to believe in and expect the best of people. This leads to occasional disappointment, but is a huge improvement over going around bitter and vicious all day every day. In the case of the hack team, I have to point out that I believe in AND got the best of people. So a happy weekend anyway.

      My approach is to do something positive, and then work toward a win/win/win for everybody assuming the very best of intentions all around. The rest of the world can lead, follow, or get out of the way. Some I win. Some I lose. And some get rained out.

  4. On the Powerwall announcement they announced that they are doubling the power output capabilities, not the capacity output capabilities, for the same price. It sounds like the charger/DC-to-DC converter they have in the package was working at about 0.5C and they’ve upped that to 1C.

  5. Slightly off topic, but for anyone who has been following the high rate rig saga, I’ve stopped the test after a rig failure (not a cell failure) after 988 cycles. No evidence of cell drift: I balanced the cells to about 50 millivolts and after an overnight rest the cells in pack A were 2.79, 2.74 and 2.75 volts. Three cells don’t make a definitive test, but let’s say that I have not seen any evidence to support the contention that high rate charging exacerbates cell drift

      1. Cell testing is really valuable – too bad I don’t have the patience to do it well. Luckily John has the right mix of ability and patience for the task. If we need 1000A to feed the P85 beast, we’d better feel comfortable with how far we can push these cells. The more we know, the more flexible our battery pack designs can be.

        Hard-core battery testing is brilliant! Unfortunately, when I try to share my enthusiasm for battery testing with others, their eyes glaze over. What a bunch of neanderthals. (Did I offend anyone by not capitalizing neanderthal?)

        1. Many thanks for the kind words. Deming said “without data you are just another guy with an opinion”. There are a lot of people out there (many of them academics) with opinions on the behaviour of these cells.

          Got some interesting data on LiCo and LiMn, plus about 20 odd cycles on a pack of NCAs, with NMC in the pipeline

  6. John Westlund

    A P85D drivetrain with enough batteries to maximize its power in a converted 4WD vehicle that weighed 3,000 lbs or less would be ridiculous. Try to imagine a salvaged P85D drivetrain of this sort with the 85 kWh pack in a stripped-down Subaru Justy, Honda Civic Wagon AWD, first generation Mitsubishi Eclipse GXS, Subaru Brat, or any other lightweight 4WD car/truck that can be found(with the necessary modifications to make use of these components without destroying itself).

    0-60 mph in under 2 seconds and quarter mile drag races of under 10 second would be within grasp, for not a whole lot of money spent(less than a typical new car), once someone figures out how to get these salvaged parts to work.

    I wish I was in a position to help with debugging this. This is very interesting… and only a matter of time before someone builds something stupid!

  7. If Tesla is on board with the idea of selling access to other cars, then there’s no reason to “hack” the Supercharger in the first place.

    Hacking the Supercharger network is a very foolish idea. There’s no legal ground to stand on like there is for hacking drivetrains.

    1. There is a group of people hacking door locks. They usually have two different door bells. One meaning please open, the other meaning hey, if you dont mind, I am opening the door right now. Dont bother to rise from the chaiselong.

      Those people are invited to parties they would never invite you or me. They get all the information about new locks and sometimes they get a free sample without keys.

      If we want to join the party we have to hack the supercharger. We can always say thanks for all the fish and would you mind a donation for the electricity we helped us? We are not burglars after all.

      With Wrecklas we have to find out, is it the supercharger refusing to offer a charge or is it the Wreckla refusing to take a charge. It could be either or both. The Tesla is connected to the internet. If the Tesla knows it must not charge then there is no need for the supercharger to lookup a database. But if they do some kind of hidden accounting, telling when I charge how much and where I go afterwards, that is information some people would not like to be publicly known.

      If Tesla knows it, everybody knows it. I dont think Tesla likes other car makers to know how much each and every Tesla consumes and how far they can drive with it.

      Cheers
      Peter and Karin

      1. Peter, I suspect that Tesla DOES collect some statistical data on how often a vehicle uses a super charger station. Here’s my reasoning: Elon stated, I believe in his shareholder meeting, that the supercharger network was for “free long distance” and that some people were utilizing it excessively supposedly for non long distance.. and that those drivers might get a letter asking them to not do that…. I might have gotten the wording a bit wrong but that’s the idea. I don’t know if they are logging simply that a vehicle charged at a supercharger on this date, or are recording detailed info like KWH used, time spent, etc. For all we know they could be using GPS data to determine when the vehicle is parked at the supercharger, but I would hazzard to guess that the supercharger logs the vehicle and reports back to the mother ship.

        1. Thankyou Brian.

          I do not drive a Tesla but I have seen in reports about driving a Tesla Roadster, they do know each and every power socket here in germany. So when you want to get more than 16 amperes on a single phase and from a socket they do not yet know, you have to phone them. It is not a problem. They do enable it almost immediately. If I remember correctly then the test car had to be enabled to “fast” charge at a Tesla owners home. Only those two cars were enabled to “fast” charge at that very socket. GPS was used to identify the socket.

          That is not the model S and not Supercharging but it might help us finding out how they are thinking.

          A Supercharger has more than a single charger. GPS is not good enough to identify one of them. So it is the charger … if only to find a broken charger just in case.

          Being a space guy we can be sure he will log data at both ends.

          So borrowing a VIN number, be sure to have your Tesla parked at that very same Supercharger to keep GPS reasonably close.

          1. Peter, while reading down I was itching to say the very same thing about location/VIN.
            These cars are tracked and any serious mismatch represents a breach of contact.
            Maybe it will be handled quietly… Who can buy supercharger sockets with cash and no receipt ?

            The Mennekes sockets in EU-land are not the extra deep(?) versions as used by Tesla.

  8. Mr.Greenjeans

    Congrats on figuring out the Tesla sekert sauce. How about teaming up with Art Morrison Enterprises to mount the drive unit in one of their independent rear suspension frames, and fashion a roller-skate chassis with a small battery pack mounted under the floorboards and between the frame rails so you could mount various bodies (63 F150 Unibody) to it. Range extending packs could be installed for and aft.

    1. Yes – but not forgetting that we need 1000 amps to feed it: that is 10C for a 100 Amp hour pack which at 375 volts is north of 35 kw hours – so it can’t be too small a pack unless you are going to fiddle about with model aircraft cells that can handle 100C. If I have done my sums correctly 375 volts of CA100 fi cells would weigh over 800lbs

  9. Congrats EVTV HT!! Jack, funny you should mention about tech and automation. Last night I just discovered an automated tap beer dispensing machine at a sushi restaurant in Japan. It tips the glass and even has a head dispenser. Here’s one like I found on YouTube:

    http://youtu.be/MiXoqry5zIo

  10. I haven’t heard much about boats for a while and was hoping we would see some video on the build. Did you ever get the insurance claim settled for the Speedster boat? We were all waiting for it to show up at EVCCON last year and then after the show when it did show up, was quite a disappointment!

  11. The garage dinged my Ampera slightly during maintenance and while they are fixing it they have loaned me a puddle-jumper powered by a piston engine with a stick shift. I had forgotten what a thoroughly nasty thing a piston engine is for propelling a car. An absurdly narrow rev range and if you forget to push the clutch down as you stop the stupid thing judders to a halt and has to be restarted. And they don’t have the anti-roll back that you get on AC motors/controllers so a hill start is a complex co-ordination of handbrake, clutch and throttle. Modern cars are brilliantly executed examples of a lousy operating principle

  12. Congratulations, Rich Morris, on your first drive in the Studebaker! It looks like a great team effort and you all should be very proud of your accomplishments. Interesting that you commented on the “creaks and rattles”. It’s remarkable how much of the chassis and body noise is masked by the ICE. I was astounded with my first conversion at the noise shock absorbers make, and tires really sing on the roads around here. Hope to see you and your Stude’ at EVCCON.

  13. Well done everyone!

    I am in the process of gathering information to do an Mazda RX-8 EV conversion in a years time and was thinking with the move to ‘standard off the shelf OEM manufactured parts’ are we possibly limiting out battery choice. With fast chargers apparently not charging battery packs operating below 240V (if I remember correctly), and battery packs using say the CALB 180’s not able to hit that voltage without large space requirements and weight penalties, tends to force us into using battery packs out of Volts/Leafs/Teslas etc – which in their prepacked forms may not fit well, or at all, in a doner vehicles form factor. Dis-assembling them to fit could be either impractical, challenging, or maybe not even possible for a someone doing a conversion at home.

    Perhaps the reborn A123 company might bring a viable option to this potential issue, I dont know the answer to it, so am throwing the above out there as to how Jack and the team think we can get around this issue (Assuming it is actually an issue and not something I have missed)

    Cheers

    Mark

    1. Yes, you’ve kind of missed something here Mark. In going to much higher voltage, the necessity of 180Ah cells diminishes rather quickly. 100Ah cells would represent an enormous battery pack.

      At 335 volts, we generally use CA60FI or CAM72FI cells, and they are nice and compact and tidy.

      At 120v and 180 Ah you get about 21600 Wh of capacity. To get that same capacity at 320volts requires 67.2 AH of cell. The remaining issue is current. But these high voltage AC systems tend to be limited to about 400 amps max on even pretty high powered equipment.

      Jack RIckard

      1. Thanks for enlightening me Jack, I must admit I didn’t consider going down in battery size capacity to bring the pack voltage up. I think in part I was going on the assumption that the larger capacity batteries were more energy dense than the lower capacity ones, and hence ‘better’.

        I have another technical related question which doesnt really fit into this blog topic, do you mind if I message you privately, or just post it here regardless?

        Cheers

        Mark

    2. You’ll probably find, as most of us with successful EV conversion have found, that public charging is really not all that necessary.
      If your intent is to do a coast-to-coast with your REVX8, you’ll likely be disappointed…even if you manage to get CHADEMO working.
      If you’re one of the 2% of Americans who routinely drive more than 90 miles/day, again, you may be disappointed with current technologies.
      If, however, you have no C2C road trip plans and only drive 10mi to work, charging at home will be more than enough.
      Good luck with your conversion!!

      1. Hi Bill

        Even though I will be charging at home I like to have the option of using the limited charching infrastructure here in Australia if something unexpected occurs. However, that is not the reason why I was talking about needing high battery voltages. Jack was talking about using OEM parts from Leaf/Volts/Tesla’s etc, and they are set up to use high voltage battery packs. I am not even sure if say a Leaf controller/inverter will even power a motor if it is fed say 100V to 144V from a typical battery pack that has been ‘standard’ for conversions over the last few years.

        Anyway I continue to research on using a Mazda RX8 as a doner car for sometime next year. To my eyes, besides Tesla models, it would be one of the most attractive modern EV conversions that could be made – I could be biased however lol.

        Cheers

        Mark

  14. Interesting info from battery field:

    Solar Impulse 2 (www.solarimpulse.com)

    NO SI2 FLIGHT BEFORE AUGUST
    Following the record-breaking oceanic flight of 5 days and 5 nights (117 hours and 52 minutes) in a solar-powered airplane, Solar Impulse will undergo maintenance repairs on the batteries due to damages brought about by overheating.

    Despite having completed the longest and most difficult leg of the Round the World Solar Flight, #Si2 has suffered battery damages due to overheating.

    During the first ascent on day one of the flight from Nagoya to Hawaii, the battery temperature increased too much due to over insulation. And while the Mission Team was monitoring this very closely during the mission leg, there was no way to decrease the temperature for the remaining duration of the flight as each daily cycle requires an ascend to 28’000 feet and descend for energy management issues.

    The damage to certain parts of the batteries is irreversible and will require repairs and replacements that will take several weeks to work through. In parallel, the Solar Impulse engineering team is looking at various options for better management of the cooling and heating process for very long flights. Solar Impulse does not see the possibility for any flights before 2-3 weeks at the earliest.

  15. As you may have heard on the show this week, I have sort of taken it upon myself to rev-up the Facebook page a little. I have created a photo album for 2014 so if you have some good shots then post them as a comment and I will move them into the album (or you can email to me nick@emotionev.co.nz).
    I will of course be creating an album for 2015 and soliciting all your images while at EVCCON!!

  16. Loaned Ampera to a friend for a few days; Civic is my only set of wheels. Over 500 miles on the conversion and it just goes when you turn the key. The HPEVS AC50 is great in this car if you don’t mind having to change gear occasionally on fast roads – not wild performance but keeps up with the traffic.

    I am coming to the viewpoint that dispensing with the gearbox is less about power than about usable rev range. The Civic is fine anywhere from 0 to 45 m.p.h. in 2nd gear. At 120 volts, the torque drops radically thereafter. With twice the voltage and twice the rev range second gear ratio would be good to 90.

    Mark – a friend over here converted an RX8 – it would be my choice if I were doing another conversion mostly because there are cheap donors with blown engines. There is a power steering gotcha (my friend burned his out and had to get it repaired)

  17. DC BUSS CAPACITORS

    Damian has mentioned some on these before, and I would like to learn some more about them, hoping to stir up some discussion about it here.

    1. What is their purpose and why are they are located in the inverter. Is more in parallel, or higher capacitance (Farads) better?
    Say we don’t care so much about range, but still need high energy for a lighter drag racing application. Lets just say I want to run 2 Tesla P drive trains, one front, and one rear, using a pack of CALB 60s In a car about 2500 lbs. That would be still a usable range, but every now and then, I might just need a bout 10 seconds of maybe 30c-40c from these batteries. So, I was wondering if adding a bank of caps could be an option.

    2. Does voltage rating need to be close to operating voltage, or is higher better?

    3. Is Aluminum Electrolytic the cap of choice nowadays?

    1. Doug Ingraham

      Jack P:

      1. The caps need to be as close to the power switches as possible. This is because part of their purpose is to absorb a voltage spike that would destroy the power switch when it turns off. The other part of their purpose is to average the current on the battery side of the inverter. If you don’t have them the battery sees the full motor current when the power switch is on and zero current when the power switch is off. Since they are a somewhat expensive and bulky item you want them to be as small as possible. More is not better, you want the right size. One way to make them smaller would be to increase the switching frequency. This of course has other side effects namely increased switching losses.
      2. The voltage rating needs to be greater than that of the battery at the highest point during charge.
      3. There are several types used. I believe some sort of film cap is the currently preferred type.

      Supercaps as the power source for drag racers is still about two orders of magnitude away from being practical. If you want to add something to allow higher currents for a few seconds the most practical thing to do from a weight and bulk and cost standpoint is add more LiFe batteries. In a 2500 lb car you would get 250 wh per mile. If you have 100 of the 60AH cells in series this gives you a pack with a nominal 340 volts and a 19.2kwh capacity. A drop dead range of around 77 miles. A 10C 10 second peak output power of around 204kw (273 HP). As you posit this is not even close to enough battery for a single Tesla drive train. Your best bet would be to use the batteries John Metric sells. They can do 100C. I am building a drag racing pack from these and expect to have it at EVCCON. A 50 lb battery with a 9 mile range replaces my 60 mile range 400 lb LiFe battery and it puts out a lot more power.

      I hope that answered your question.

        1. Earlier in my life I did some experiments charging Batteries with asymmetric AC or biased AC. For small batteries (NiCd mignons) little resistors did the trick. For a 12V car battery I tried electrolytic capacitors. Worked nicely for a couple of days until one of those bastards complained. Unrolled all the paper in my room – oh, what a mess.

          250 volt capacitors (dc rating) are not meant to work as resístors at 24 volts ac.

          Modern capacitors have a vent but when they blow they release a lot of smoke and can spoil a computer case and all that is inside.

          My old capacitor did not have a vent. It did not blow it exploded. I guess if you really need a capacitor it grows gray hairs quickly.

          I dont mean get rid of them. Make them big enough, capacity and voltage. If they are meant for dc take care they dont see negative voltage.

          Cheers
          Peter and Karin

  18. Popular Science has an article about a new “Breakthrough” electric car, with a 20 minute recharge, 300 mile range, and a 90 MPH top speed:

    https://books.google.com.ec/books?id=pgAAAAAAMBAJ&printsec=frontcover&source=gbs_ge_summary_r&cad=0#v=onepage&q&f=false

    Oh yeah, one more thing: the article is from February, 1971, about 45 years ago. I keep telling friends of mine that we are still in for a bit more of a ride before crossing that chasm. Tesla P90D Ludicrous speed aside it is still a slow journey, but we all know how it plays out.

  19. Peter, intersting that you mention capacitor longevity. The reason I started wondering and asking about the DC Buss caps, was that on the S-10 electric email forum, it was mentioned the other day that the GM Power modules’ caps don’t hold up well. They have been a common problem through the years. The inverter has (4) 2200uF 450v caps, and I was wondering if I can replace them with something of an upgrade, or maybe just keep the spec and find high quality. Elon Musk had dreams of Super Capacitors at one time too.

    I also found something on Wikipedia of interest. It is the Capacitor Plague.

    1. Jack, thankyou for pointing me to the Capacitor Plague.

      Yes, it has hit me several times and not so long ago. Seen from the backside the fan blades of the power supply seem to have collected more
      dust than usual. If you ignore this your power supply might die killing mainboard and disks.

      Inside the power supply you will find the capacitors of one type actually of one rail have vented and maybe some of the other rails as well.

      Some people have successfully replaced those capacitors as long as the power supply unit itself was not damaged yet.

      Karin and me are in the process of moving into our computer. There is a problem of pc cases collecting dust and not getting rid of heat. We want to spread the insides of some PCs all over the room and use something like a kitchen fog remover, a big fan to transport air through pipes to the place where it is needed to both get good ventilation for the room and cooling the computer mainboards. That is going to be the i-MiEV remote dashboard.

      Side effect, 12V LED lighting is feed from the computer. No more 220V lighting with changing sockets and regulation every year.

      Cheers
      Peter and Karin

    2. Jack P, To elaborate further:
      As Doug pointed out, the caps need to be in the inverter, and newer designs use film caps. Using film caps instead of electrolytic caps has many benefits except for cost. The electrolytic caps usually need to be oversized to handle the level of ripple current caused by the PWM switching. By over sized, I mean the design calls for X capacitance, but if you used x capacitance in an electrolytic style cap, the level of ripple current would be far above what the capacitor is rated to handle. Larger capacitor banks have larger ripple current capability, thus you have to size the caps with more capacitance than is required in order to get the ripple current within the caps ratings. Now with film caps, they usually have MUCH higher ripple current capability, and thus you don’t have to over size them.
      As for your issue with the S-10 electric, and the caps being a common failure….. In any electronic switch mode power supply or inverter design, the electrolytic capacitors are usually the first to fail. They are rated to operate for X hours at Y temperature. A common rating would be 2,000 hours at 105C. 2,000 hours isn’t long, but if you can operate it at much lower temperatures, then the operating hours increases dramatically, it’s not linear by any means. The trick is that a capacitors temperature is affected by the ambient temperature inside the inverter, as well as the ripple current to the capacitor. The ripple current heats the cap up from the inside, thus the higher the ripple current, the hotter the cap get’s. If you were a clever engineer and designing a piece of equipment, based on the temperature and capacitors used, you could design it such that it would never fail within the warranty period but fail in under a year after the warranty expired.
      I have replaced the capacitors in numerous pieces of consumer electronics, Just last month I replaced the caps in the power supply section of my laser printer. I’ve also done my DSL router and numerous computer power supplies. I usually replace them with higher hour rated caps and also lower ESR (internal resistance) rated for higher ripple current. This provides the replacement caps with both longer hours and causes them to run cooler because the have lower ESR.
      To the S-10 electric folks, I would recommend doing the same, don’t replace the caps with OEM parts, because the will last just as long as the originals. Replace them with higher grade caps.
      From a consumer aspect, the best thing you can do to keep your electronics lasting longer is to keep them clean internally and well vented to keep the caps temperature as low as possible.

      1. In regard to super caps, they have a few main problems as I see it. First they have poor frequency response, so you still have to use film or electrolytic caps inside your inverter. Second, their SOC is a function of their voltage, thus at 50% SOC, they are at less than 50% voltage. So in order to use them as an energy storage device, you have to have an application that can operate with HUGE swings in supply voltage. A boost converter could be used to boost and provide a stable DC bus voltage with a variable super cap voltage. Then with a boost converter you have another point of losses, point of power limitation (power limit of the boost converter), and more space and weight consumed. They also have a self discharge rate that most people seem to ignore…. There’s a reason you don’t see super caps on any large scale production EV’s.

    1. Yes, I was aware that certain model years of the prius had a 2:1 boost converter. Keep in mind though that this is a more or less fixed ratio conversion and not a variable input voltage, which makes the power supply more complicated… I still can’t fathom it being a good idea though… If you need 650V, why not just use twice the number of batteries at half the size? Extra credit to anyone who knows why the DC bus on AC inverters is always 300-400, or 600-750V but rarely anywhere in between?

        1. I’ve been confused about Jacks conversion from AC to DC for quite a while now. I learned in school that you divide the peak AC voltage by 1.414 to get the equivalent DC voltage. What gives?

          1. It represents the relationship between peak voltage and the “root mean square” voltage. When discussing the alternating current on our grid, here in the U.S we talk about single phase electricity as 120 vac and two phases 180 degrees out of phase, as 220vac. These values are actually the root mean square of the peak voltage. So more properly 220vac rms. The PEAK of the sinusoidal waveform is actually about 311 volts. And if you put this through a full wave bridge rectifier into a set of input capacitors, they would quickly charge to that value, but no higher. This kind of represents the AC input circuit into a switching power supply.

            Conversely, if you want ot calculate the root mean square from a sinusoidal waveform, I suppose you could divide it by 1.414, but I was simply taught to multiply again by 0.707.

            480 x 1.414 = 678.42 volts peak.

            The concept is that for any given power switching electronics, if you use commonly available AC in the US. 220vac gives you 311vdc as an input while 480v 3-phase gives you 678vdc. There IS actually quite a lot in between as 3-phase is also available in lower voltages, but I’m guessing this is the answer to Brian’s question, which is why do you see AC inverters of 300-400v DC input and alos 600-750 vdc input, but rarely anything between those two zones.

            Indeed, we routinely use Vicor’s excellent 220vac switching supplies as DC-DC converters. We simply rig up the input to high voltage DC and get 15vdc output which can be trimmed to whatever you like, 13.5vdc most often as we do it. The full wave rectifier can’t really tell 300vdc from 220 vac rms, except the current only goes through half the full wave bridge. The Escalade has two of these Vicor power supplies onboard. Mark Wiesheimer first suggested this. We load up the DC output card slots with 5vdc 40A cards and strap them together for 80 amps at 15vdc.

          2. Jack pretty much hit it on the head. Forget about residential voltages because few residences utilize variable sped AC drives.. Now look at industry, they are either (in the US) 240Vac 3 phase, or much more commonly 480Vac 3 phase. Inverters are manufactured by the brazillion and sold throughout the world. Each of them needs IGBT’s, and thus the IGBT’s that the IGBT manufacturrers make are those that they can sell brazillions of. Thus the IGBT’s that are available for purchase were designed around an inverter design for either 240Vac or 480Vac input. Some countries utilize 380Vac 3 phase. Most inverters that are used in those countries are actually 480Vac inverters that are dual rated to simply work on a lower DC bus voltage. It’s common to see the voltage rating of 380/480V on an inverter. Same thing with electrolytic capacitors, common ratings are 450Vdc. To get higher voltages (for a 480Vac system) they just put 2 in series.

  20. Thanks Jack, I learned some important stuff here. I do have one more question. Any full wave bridge rectifier I’ve seen outputs about 1/2 of it’s input AC voltage. Peak to peak AC input from residential is a bit over 300v. A bridge rectifier will output around 150 volts, no?

    1. Doug Ingraham

      Josh, Not sure where you got that but in a residential service in the US we have what is called 240 volt split phase. There are three wires coming into the residence. One is the neutral which is a center tap on the transformer and the other two hot leads are one phase of the AC transmitted power. In your box the neutral will be tied to earth ground. If you feed the 240 volts AC into a full wave bridge you get a peak DC voltage on the output of the bridge of approximately 339 volts. If you take the neutral and run that into one side of the full wave bridge and either hot into the other AC input of the bridge you get 120 volts AC input and you will see DC peaks of approximately 170 volts.

      1. Doug, Josh might be referring to a 2 diode configuration applied to a center tapped transformer. Josh might be familiar with this approach on the secondary of transformers inside of consumer electronics. It was common to use a center tapped transformer, place 1 diode on each hot leg, and then tie the outputs of each diode together. Then the DC output would be taken from between that point and the center tap of the transformer. This method was commonly used in consumer electronics when linear power supplies were common. This method required a center tapped transformer, but eliminated 2 of the diodes as a cost savings. We’re a bit off topic here…

  21. Good discussion Brian and it is jogging my memory from my electronics classes 40 years ago. it’s a shame my Fiero project was started just before the AC motors and controllers became popular and began to be priced in an affordable range. It would have been nice to have been able to hook my Better Place modules in series with an AC motor as it seems that at 120 volts running the DC motor, there is quite a bit of sag, with just one module. I made pretty good progress this weekend getting the second module hooked up, so it will be interesting to see if putting it in parallel will help with the voltage sag when accelerating.

      1. Mike,

        Maybe there should be a contest for Co-Host of the week and if you win, you could travel to the Cape for a week to work at EVTV and then host the weekly show…but then again, I still think Jack’s daughter would make a great Co-Host!

        I saw your scuba diving video, where you holed up at now?

        1. Co-hosting EVTV could be a top priority item on my bucket list! I’m in Melbourne, Australia right now, about a month into my 7 month around-the-world trip. I went to the Tesla showroom in Sydney last week, one of the Melbourne stores yesterday and the other one is on the schedule for today. I’m trying to hit as many as I can on my travels.

          1. Well you shoud definitely stop by and sit in as “co-host” then. Actually I think you’ve already been one of us. I recall a shot of an electric vehicle conference, maybe a formula E report, and then a BMW i3 test drive. So I suppose you are already an “associate producer” I believe is the industry term. You might as well handle the front desk if you have an opportunity to get out this way.

            Jack Rickard

  22. CollinK

    I just watch the Mercedes it looked like it was purring like a kitten. It sound Real nice.

    By the way I told you so 🙂 I have never any doubt that this would work and keep on working.

    Best regards Allan

  23. Jack,

    I’m wondering where you heard the Elon Musk quote about a degree being optional, I looked on the web and the only MIT interview I found was done at MIT AeroAstro but there was no mention of degrees. Maybe I’m missing something?

      1. Yes Jack naturally, but there are over 100k videos of Elon Musk on YouTube ranging from 2min to 2hrs so any criteria to narrow that down (or better yet a link) would be most appreciated.

        1. Well it might be appreciated, but I watch a LOT of video online. ANd as far as I can tell, neither Google or Youtube search has my social security number.

          I found it a very interesting comment, but not interesting enough to Bookmark it.

          Jack Rickard

          1. I also found it to be an interesting comment, though your remembering of it sounded a little direct even for Elon. I believe this is the video you were talking about http://www.youtube.com/watch?v=FE4iFYqi4QU

            Below is the actual quote:

            “There’s no need even to have a college degree, at all, or even high school. I mean, if somebody graduated from a great university that may be an indication that they will be capable of great things but it’s not necessarily the case.”

            Very similar, but you can’t just put your own paraphrase in “quotes” unless you are actually quoting someone word for word.

          2. Obrien28:

            I’m not certain just what you’re on about, but take it elsehwere. I can already tell it’s a bunch of BS and you really haven’t even tuned up yet.

            First, I can do EXACTLY as I like and I don’t need any instruction from an unknown wannabe about what I can say or can’t say.

            Second, I did more or less paraphrase it, but quite accurately, and no, I was quite clear in that it was an address at MIT, which is what made it a striking response. The video you’ve pulled up has nothing to do with it, other than he seems to be consistent and wasn’t just yanking the MIT guys chain. More or less, so there, he said it AGAIN.

            It is mildly interesting, but just barely. You’re obsessing. Might I ask why? It was certainly not my, nor actually Elon Musk’s parrticularly inventive concept. The notion is quite widely held in Silicon Valley these days, given the unusual percentage of players without the piece of paper, and some of them rather pointedly – Jobs, Gates, Zuckerman, et al.

            Musk and Straubel both did graduate Stanford and regularly visit there and support the institution.

            I take it all as somewhat damning of higher education, and indeed our educational system itself. I understand Musk has started his own kind of school for his kids, and if it follows the Tesla/SpaceEx/Solar City model, it might behoove any of you with kids to see if you can get yours in.

            But beyond blathering accusations and demanding citation and footnotes, what the fuck is your point?

            Jack Rickard

  24. Pingback: jordan 11

  25. I agree with Musk on the doers vs the watchers… I used to own a Network Operation Center. I was doing all the setups, updates and security and wanted some help since I wrote, designed and were hosting a shopping software platform I sold to 1 of the top 3 search engines in 1999. So I hired a couple of system administrators… With college “debris’s” and MS certifications… Gave them each a server to set up and lock down… less than 24 hours later they were both hacked into… Needless to say, they became immediately unemployed. I have an associate’s degree in Electrical Engineering and 0 computer schooling or classes, I learned everything by reading anything I could get my hands on, playing with any computer I could work on and writing software for any language, starting in 1974. So having all the diplomas, don’t mean squat… I have met a lot of college grads (even some with PHD’s) that have no clue, common sense or won’t try and look outside the box.

    As far as hacking it, Tesla customers will get pissed off waiting in line for another car being charged, plus what if some home brewed system blows up or damages or destroys a charging station or other cars? These days people will sue you for just looking at them sideways… And if it gets wild spread, Tesla could just install cameras to recognize if the car is a Telsa model or not, or match license plates to registered vehicles. Not to mention the R&D money spent on designing the system… Being a programmer, I have spent a lot of time and effort keeping my software and Intellectual Property from being stolen. Some of it has taken years to write and test. Reverse engineering is theft… it wasn’t your idea to begin with… Just like all the patent trolls that are running around… Pay to use the technology to the company or person that actually wrote it.

    1. Old school and very very self referenced. Reverse engineering is NEVER theft. But particularly not when you own the device with the software in it. You bought it. You own it. And your “protection” crap you spent years developing IS theft. It denies me access to what I’ve already bought and paid for. That my “use” is not what you intend isn’t even pertinent to the conversation.

Leave a Comment