Just Another Pretty Face: How DIY Enthusiasts are Outpacing Big Automakers

Some days you are the bug, and some days you are the windshield. And so it goes. We have a busy 2:23 video this week. Last week we saw Damien Maguire install his Siemens 1PV5135 motor and DMOC 645 controller in the BMW 840ci. This week we visit with Ed Clausen and his BMW320i with the Siemens 1PV5135 motor as well as Paulo Alameida in Lisbon and his Mazda RX-8, or more on point this week, his Siemens 1PV5135 drive train.

And so in addition to our own 1974 VW Thing using this Siemens motor, we have now at least three very prominent builds nearing completion using this roughly 100kw drive train and of course our GEVCU controller.

We're doing a second VW Thing with the UQM Powerphase 100 system from the Coda bankruptcy, and should soon have some video from Collin Kidder on a 1959 Mercedes 190SL convertible with this drive train and the GEVCU.

I think it's important that we get a number of practical installations on the street and rolling before these OEM quality AC drive systems are accepted as "doable" in the custom electric vehicle community, and so I'm quite pleased with the progress here. We originally purchased 65 Siemens motors from the Azure Dynamics bankruptcy but also incurred a heavy investment in an additional 100 of these brand new in box motors from Siemens themselves after they had been hung with them post bankruptcy.

Indeed UQM suffered the same fate but on an apparently larger scale from the Coda bankruptcy. We purchased 10 of those at a much higher price but the company reports some $7.5 million in stock left over from the event and are none too happy about it.

The GEVCU was of course developed for the Azure Dynamics drive system so we could drive it at all. And it proved simmilarly useful in operating the Coda drive train. But to more advantage, it allows UQM to provide us with drive trains at much lower cost - almost entirely because they don't have to support it. They won't warranty it as it works with our GEVCU, but they also don't have any means of providing product support, as they don't really know anything about our GEVCU. You can of course purchase brand new fully warranted products directly from them, at about twice the price, as they then have to provide product support basically for your project.

And so GEVCU is working out precisely as I hoped, not only as a way to operate orphaned equipment, but also to pry open this kind of circular pricing conundrum based on the vendors need to cover rather pricey engineering talent devoted to answering your questions. It's a real issue. And GEVCU somewhat deflects that issue.

Indeed we are pleased to announce that we shall soon be distributors for the Rinehart Motion Systems line of inverters. This is kind of key for a couple of reasons. While we did obtain 165 Siemens motors, we only got about 70 of the DMOC controllers. Once they are gone, we had nothing to really run the remaining Siemens motors. The Scott Drive from New Zealand has remained strangely elusive as far as going into production in any quantity at all. But Rinehart, which already features "tunings" for more motors than any other, did do one for the Siemens 1PV5135 as well.

While slightly pricey, the Rinehart is also one of the physically smallest and lightest 3 phase inverters you'll find in that power class. Indeed, their main line of business as it has developed has been in the racing community. They will be in our 2015 catalog, which I'm still working on.

We also received some preliminary test data from John Hardy of the UK. Mr. Hardy, author of the popular EV how-to book ICE FREE has done some interesting work testing cells in the past. This past EVCCON I awarded him six or eight of the CA40FI cells to play with. He has devised a kind of cunning test where he packages all these cells in a single insulated box so their temperature/environment will be the same, and then ran some cycle tests on them. But he runs one set at standard charge and discharge rates, and the other set at fast charge rates of 3C or in this case 120 amperes.

As you can see from the accompanying graph, it would appear my prediction that fast charge would have little ill effect on these cells appears to be in error. It appears from 500 cycles that there is some readily measurable deterioration of the cells in fast charge. The control pack decreased in capacity to 94% of original while the 3C pack appears to have reduced to 90%. As the high charge cells did not exhibit a significant temperature difference, we can only surmise that this is from a degraded anode crystaline structure and perhaps some increased thickening and deterioration of the solid electrolyte inter-phase layer or SEI on the anode.

Worse, if we project the trendline, the resulting decrease to 80% capacity, our usual definition of pack life, would indicate 1600 cycles for the normal charge which is quite to spec for 100% discharge. But this is reduced to 1000 cycles at 3C.

That's still pretty good cell life largely due to the long cell life enjoyed by LiFePo4 anyway. And it is true that this test uses fast charge for ALL the charge cycles, except for the measurement cycle taken every 50 cycles. It's unlikely in practice that you would charge ONLY using fast charge and fast charge every time. But while it remains a real factor to be considered, I consider the level of damage both affordable and acceptable in these cells for the convenience of occasional fast charge.

On another topic, I observe that the custom EV crowd pretty much accepts as a given that however well they do their vehicles, they cannot approach the polish and function of a "store bought" electric vehicle.

There is certainly something to be said for the uniformity and quality of a manufactured car. But I actually find them a disadvantage particularly with the deeply proprietary attitude of today's manufacturers. It is difficult or impossible to modify the operation of a Volt or a Leaf and indeed probably impractical to do anything to repair it.

Indeed, a not yet evident factor in moving to electric vehicles is that I don't think the manufacturers are making any particular progress in passing repair training effectively to their dealerships. Too often, a fairly straightforward failure requires a visit from the "factory team" to suss out, often engendering days or weeks of delay.

Indeed we were delighted to finally be able to purchase a battery for a factory built SmartED after months of wrangling with the top man in the company. Finally they agreed to sell us the battery for $7880 and even offfered to install it for us as it of course requires a proprietary change to the Vehicle Control Unit to carry the SERIAL NUMBER of the new battery to work at all.

We trailered the car to the Smart dealer in Creve Couer Missouri. WEEKS later they called us to note that the CONNECTOR to the battery was damaged. They have no connector repair replacement capability. All they can do is plug things in. So the motor would have to be replaced and with labor that would be a smooth $7100 additional. We declined that gracious offer and the car is currently in limbo, because it is BEYOND THE CAPABILITY of the Mercedes Benz Smart dealership, which touts the Smart ED, to replace a CONNECTOR????

So as these vehicles age, I see a coming backlash against manufactured electric vehicles that is quite dangerous to the genre. And that again goes to very expensive factory repairs - drastically reducing the resale value of these vehicles. But along the way, my sense is not so much pride of ownership, but rather a new sense of victimhood. I'm a victim of my Tesla Model S and my Smart ED.

As Ed Clausen noted, if he could only use another kWh or so out of the 24kWh IN the Nissan Leaf, his range issue would be done with. But THEY ONLY LET YOU use 20kWh of the pack. Yes, that is what they will ALLOW....on your car.

So more broadly, why can't his BMW 320i be competitive in features to the Nissan Leaf? The BMW will be a car he has COMPLETE control over, thorough knowledge of, and the ability to repair or modify at whim.

As I said in the video, the motor turns the same direction whether we install the motor or Nissan does. We don't seem to have any difficulty installing motors, and now we can have OEM grade and style motors anyway.

I actually think we are out AHEAD of LGChem and the OEM's on batteries in some ways. Our LiFePo4 cells are inherently safer and longer lasting than the chemistries they use. And while they achieve higher energy densities using their more incendiary chemistries, they largely give that back with heavy, huge, and overengineered mounting and heating and cooling and shunting and BMS ridden monstrosities. Over a third of the Better Place pack is extraneous metal. And over half was in addition to the pouch cells themselves. So what is the advantage of lighter pouch cells with more energy dense chemistries. On balance, our lego bricks are easier to deal with and more durable. Net net, we're getting the same range.

No, the central differentiation goes to instrumentation and displays.

Here are the Volt, Leaf and Tesla displays which show all sorts of useful as well as irrelevant information.

You will find it surprising to learn that I devised what I considered a very pleasing and effective display system three years ago.

It is actually relatively easy to engineer such a display, but it is easy largely because they are doing it for one specific vehicle and so much can be assumed. Applying this same concept generally across a broad ARRAY of vehicles using different drive systems, different battery systems, etc poses some unique problems before we can get to a GENERALIZED display approach for electric vehicles. There are actually two many strategies to pick from.

You will find it surprising to learn that I devised what I considered a very pleasing and effective display system three years ago.

The EVu display actually worked quite well. I used an Arduino and made my own shield with an RS systems bluetooth module on it. The module happened to feature a 16-bit analog to digital converter and of course since it was communicating the data via bluetooth isolation wasn't terribly important. It used a LEM HAAS hall effect current sensor and then divided the pack voltage down just using resistors.

Two problems. One, I found the 5v powered hall effect device mysteriously drifty. All who deal with these eventually find the same thing. Second, the bluetooth data was transmitted to a Macintosh laptop and required a program designed just for the Macintosh computer to present this display. I would have had to write teh whole thing over in C# for Windoze for most of you, and then maintained two computer programs indefinitely. And THEN you would have to have a laptop in your car - or at least a CARPUTER to display the data. And so I had once again developed the $2500 unwieldy solution.

We have done some better with the GEVCU device as it has a built in web site that presents most of this information in fairly attractive fashion. By using any browser, we eliminate the dedicated software release.

The problem here of course is the GEVCU has to drive the car. And so the wireless website is a rather low priority that updates rather slowly. Using the klunky web browser interface, it looks slow and awkward.

But we do have plans. Collin has done some preliminary work on an OBDII interface to the GEVCU. Note the display below was designed for the Orion BMS developed by Ewert Systems. This uses an "application" on an Android mobile phone or tablet. This can be done with GEVCU and we eventually hope to get this running.

I myself kind of think the car ought to be independent of Smart phones and Tablets as I don't know what they might look like in 10 years, but my car will likely still be driving. I think the car should be self contained with regards to basic display. We mostly pursue this mobile application front because I want to be able to add OBDII generally to older vintage cars.

Paulo Almeida in this episode very impressively shows an existing Mazda RX-8 instrument cluster working perfectly with his Siemens drive train and GEVCU. Like almost all non-digital displays of the last 10-15 years, the Mazda has a conventional gaged display, but most if not all of the gages are driven by Controller Area Network (CAN) commands from the vehicle control unit (VCU). Paulo has adapted our GEVCU to send the appropriate CAN messages to the instrument cluster making it fully functional.

This is eminently doable. But it has a problem. Paulo had to learn the CAN commands for the Mazda instrument cluster. Fortunately, the Mazda has been kind of a popular car to re-engine anyway, and some Mazdonites had already worked out what most of them were. But the key is they are PARTICULAR to the Mazda. My Escalade has a very similar instrument cluster, but of course uses entirely different messages. So it is not practical to simply have these in GEVCU. Indeed, it requires significant C++ beyond Paulo's range to accomplish and so Collin Kidder wrote him a "special" module just to interface with this legacy, but CAN driven, instrument cluster. There are too many such clusters to generalize this very well.

Or maybe not.

David Seabury, in this week's episode, describes his very low cost Arduino based device that also does instrument clusters. But it has drivers for the earlier instrument clusters with real steam gages driven largely by sensor voltages for temperature, gas gage, pressure, etc. He also deals with the pulse-per-minute common tachometers used in these systems and the idiot lights.

But he takes a very interesting approach. He assumes the DATA is available from CAN and then translates that into analog signals to drive the gages.

We actually DO have analogue PWM outputs on the GEVCU - again you could do it but it would involve some software development.

David has done that for you allowing you to simply configure the CAN message ID and data format, and which gage you want to drive. He even provides a little system to calibrate the high and low of the gages. As we already have half our GEVCU outputs busy with precharge and brake lights and backup lights and cooling fans and so forth, it would appear David's solution, which we intend to carry at about $235, would be perfect for these legacy instrument clusters.

This morning, I've gone elsewhere - poking CAN messages into Brian Gallagher's EVIC display and watching the numbers change. This device CAN be interfaced very elegantly to the GEVCU, but also to the JLD505 device we have in work, or the BMS board that Ed Clausen further describes in this week's show. Or a combination of them. I haven't gotten all that sorted out as to what makes the best sense yet, but we will start with an object module for GEVCU that will drive the EVIC, and select from GEVCU data, BMS data, and JLD505 data as the source. Probably raw LEM CAB300 current sensor data as well.

In any event, I see it as ENTIRELY practical in the future, to drive the existing instrument cluster in your vehicle, AND add a sexy 8-inch eye-candy digital display very similar to the Leaf or the Volt, probably without turtles but otherwise similar.

In any event, that is kind of the central mission of our component side at this point. We want to provide everything you need to do a car, YOUR car, that you own, control, understand, can repair, can modify, and in all ways actually OWN, but I see no reason to "settle" for something less than you could also buy. Obviously this is only going to make sense for people who LIKE to work on cars and do such things. But we are moving quickly toward a future where you can customize and "program" your car to the same degree you now do your personal computer, laptop, or phone. And you won't need anyone's permission to do so.

Unfortunately, the competitive pressures have altered the landscape for America's automakers. They are now following the airlines into an ADVERSARIAL model with their customer base. Madness to my way of thinking but very real. I actually have Mercedes Benz heroically trying to market a SmartED that is actually a GREAT little car and they have so far had comically inept results selling it. I might be able to help. But instead, they would prefer, if at all possible, to SCREW ME out of $7100 for a bad connector. In public and in real time. They cannot resist it. You should have heard the GLEE in the service reps voice because he was CERTAIN I would just pay it. In for a penny, in for a pound. When I told him we'd pick up the car and seek a refund for the battery, he just refused to believe it. In fact, in desperation he notified me that the battery was non-returnable special order. I pointed out to him that thus far, they had never DELIVERED it, weeks later though it has been, and I paid for it by American Express. I can do it online. Doesn't even take a phone call.

The GLEE went GLUM. He had it in the BAG and it just went up in smoke while he's talking it up. How could this be? The man honestly doesn't know. He does it all day, every day, to every Mercedes customer he can get. And as all the stuff is proprietary, he gets them all. How long can that last? He believes forever. I believe it is over and he's already the walking dead.

Is Elon Musk an automotive genius.?? I don't know. Do you mean to compete against Mercedes Benz all I have to do is show up with a good car and declare to people that I will attempt to AVOID SCREWING THEM at every opportunity as my marketing message? And that owners everywhere will know exactly what I'm talking about and flock to it? How hard can it be?

Note that the figures are just in. The average U.S. car has moved from 9.6 years lifetime to 11.4 years lifetime in the past couple of years. Pushback may have already started.

Yes, the vast body of unwashed are destined to be victims until automakers show up to take advantage of the discord with good service and support. Ford is already kind of on a roll these days, for example.

But for those with a bit of tinkerer spirit, it will become ever easier to take any car, convert it to clean quiet electric drive, and wind up with just as good a car, without the victim status. We think we can get the costs down considerably as well, using the OEM's own parts to do so.

So while suffering bronchial viruses, plagued by staffing issues, beset by sourgrape wannabe competitors and imitators, I'm just about to bubble over. The Internet was NEVER this much fun. I'm in my element and can hardly sleep from excitement. I'm nearly overwhelmed by all the development projects and things laying around the shop begging for attention. It is without doubt the most productive period of my life.

So yes, I'm in the fist fight of my life, entirely unarmed, hopelessly outnumbered and surrounded, beset on all sides. I cannot adequately express how joyfully exquisite all that is, or how it brings out the very best in me every time. With ables such as Collin Kidder and Paulo Almeida and John Hardy and David Seabury and many many others, we actually have THEM surrounded. The outcome is more or less certain. We just have a scant 40 miles of piss, blood, puss, broken glass, and razor blades to crawl over on our hands and knees and batta boom batta bing, victory.

Jack Rickard