Dyslexics - UNTIE!: The Push for Affordable, High-Quality EV Components

We're getting a bit sloppy around here. Part of it is reality intervention in our normal reality distortion field. Last week everything we touched turned to a mess. And I'm struggling here. The Ford Edge battery pack we just reconfigured we rather didn't bottom balance. This is kind of interesting in that we have 28 90AH cells now in series with 74 100Ah cells - not good from the get go conceptually. But it was a lot of work to attempt a bottom balance as well. We drove the vehicle a bit too far, very short range anyway. And we have lost a couple of cells irretrievably.

That's not TOO much of a disaster. Recall we had to cut our pack by a couple of cells to get down to the 342 volts that the Soliton would allow anyway. So we have replacements right in the box. But it kind of drives home the point of bottom balancing. So we are going to have to perform this.

With the cells all fully enclosed in battery boxes, it is kind of interesting in making the discovery before there was further damage. We don't have a cell level BMS of course for infamous reasons. So this serves as an interesting illustration of failure discovery. It's just not that hard.

Even with little experience with this new pack configuration, we had charged it several times to set the charge level to 366 volts on the Manzanita Charger. With 102 cells this works out to about 3.6v per cell. High by CALB standards, recall these are very old Thundersky cells gleaned from the upgrade of the original Speedster from Thundersky 100Ah cells to CALB SE180AH cells. So we charge them to 3.6v x the number of cells or 367.2 volts. Actually we set the Manzanita for 366 volts. Close enough for government work and erring on the undercharge side.

After charging, these cells typically return to an open circuit voltage of 3.34v per cell - a pack voltage of 341 volts. And so if you plug in at night, by the next morning you typically see 341v. One morning I saw 331 volts instead. This is all the indication in the world. You don't just ignore it. Things happen for a reason. If you normally get 341 volts, you should ALWAYS get 341 volts. I recharged the car on the off chance that it just had not fully charged for some reason. A few hours later, we were back to 331. Not good.

Drove the car up to the shop and began disassembly. We found 3 cells in the main pack at less than a volt. Two of them were pretty much 0 v. We put a charger on them to see if we could get them back. No dice.

The failure mode here comes from cell reversal - the other cells driving current through a fully discharged cell. I've never found a way to bring one back. If the cell is simply low, you can put a low level 5Amp charge on it and the voltage will slowly climb. If it is a fatality, it will climb up to about a volt and a half, and then it reverses and begins to fall again - even while putting current. if the reversal is clear, disconnect the charge. It is a goner.

So we replaced two cells and nursed one back to life.

So why the failure. Well first we are mixing and matching cells and in this case of this not very valued vehicle, mixing and matching even cell capacities. Don't try this at home. In a series string you obviously want all cells to be as nearly the SAME capacity as possible. But in the case of the FOrd Edge, we're kind of junking around anyway. But in this case, for obvious reasons bottom balancing becomes critical. You do want your 90Ah cells to reach full discharge at the same time as your 100Ah cells. And so bottom balancing is the only way to really do that. We rather failed to attempt it. Yes, it's a lot of work. BUt it just has to be. So we are working at it at the moment.

The Cadillac Escalade is another story. We blew up our whole house tankless water heater. The element simply failed for unknown reasons leaving us without heat. We have a solution I'm quite pleased with. More on THAT next week. But we developed a pronounced stutter at 2950 rpm in the cold. If it was above 65 degrees, the thing ran smooth as glass. At 45 degrees F, anytime we hit exactly 2950 rpm, it bucked a bit. I played with our Mass Air Flow sensor simulator and even swapped it out several times with an improved model Ed Clausen had built me. Oddly mine worked a iittle better, but still had the stutter.

As it seemed to be temperature dependent, I decided it must be the IAT - intake air temperature probe. We had retained this from the original MAF sensor so the ECU would have an ambient temperature. The only other temperature input it has is the Engine Coolant Temperature. This is mounted on the front motor and gives us a temperature indication on the instrument panel when we reach a hot motor situation. I had kind of ruled this out for a couple of reasons. First, in these temperatures, we are not really getting a hot temperature on the motor. But it does warm from operation, and if it was the ECT input, the stutter at 2950 rpm would improve during continued operation. It does not. So I replaced the IAT with a simple 20 turn trim pot of 5 ohms. And adjusted for a temperature of 80F.

No change to the problem.

Recall that we do have an rpm sensor that uses the original engine reluctor ring. This ring is normally mounted on the drive shaft INTERNAL to the engine. A small plastic sensor plugs into a socket in the engine block. What is amazing about this is that this sensor has to be EXTREMELY close to the reluctor ring to pick up RPM. ANd of course the signal is a little bizarre. 58 teeth with a single gap in the teeth give us 58 pulses and blank pulse for RPM.

We had mounted this ring to the FRONT of the assembly on a pulley used to drive the air conditioning compressor and power steering/brake pump. So we had the sensor mounted on a thin aluminum tab. We could bend this tab a bit to align it but it was very tricky.

After a couple of glasses of whiskey, I decided this must be the culprit. It's kind of hard to reach with all the front end on the truck now. But I reached up and tweaked it a bit. The truck then quit idling or even running really. It was just a jerky mess trying to run it at all.

And so we embarked on a mission to see how many times we could raise and lower the Escalade at 7777 lbs up and down on the lift. Adjusting the tab, and then lowering the truck and trying to start it and idle smoothly. Then raise the truck and adjust it again. This went on for some time.

We finally got it smoothed out. And on test drive - we have a stutter at 2950 rpm. Worse, after driving it a bit, the sensor slipped again. SO we are going to have to re-engineer a mounting system for the sensor - with very limited access this time.

Meanwhile, we received two kind of special motors from Netgain. Technically, these were custom motors I ordered last November. They've been awhile in coming. They have all the new features of the latest run of Warp 11 motors. BUt they have some additional tweaks. Most obviously an EVTV green powdercoat paint job. But more importantly, a 1.375 32 spline shaft.

This is the normal output shaft of the Transwarp 11. But in this case, they made us a Transwarp with the 1.375 splined shaft going all the way through. And they made us a second motor with a 1.375 splined output shaft, and 1.125 keyed aux shaft for accessories.

The concept always was to use Wayne Alexander's billet aluminum adapter to mate two motors siamese fashion. This is SIMILAR to what we have in the Escalade. Not precisely the same. But we waited six and a half months for a motor Jim Husted had promised in six weeks. It actually marries two motors somewhat more closely and with an end designed for a GM flex plate. He even removed the internal fans to reduce the overall length.

But using Wayne's adapter, we could marry these two motors together. We bought a standard Dana Spicer yoke off of eBay and sawed off the Yoke part to a length just shorter than the adapter. In this way, we could mate the two shafts with about a 4.5 inch coupling of 32 spline material. This would make a very stroung coupling that would be very removable as well.

Richard and I embarked on an assembly procedure that was actually quite comical. We did everything we could think of to get these two motors to got together over the course of about two hours. I finally put a caliper on the motor shafts. One shaft was very nearly 3 inches long and the other about 2.6 inches. That's 5.6 inches. Unfortunately, the adapter is 4.85 inches long. We had spent a couple of hours trying to get about 0.75 inches of 1.375 splined shaft to occupy the same space at the same time. Hard work even for a couple of hefties.

As the entire project was based on this premise from the beginning, I don't really understand why the shafts weren't cut to proper length at Warfield. But according to George, it was no problem. I just had to cut off the shaft ends and it would be fine.

So we did.

And at that we installed our Garret Turbochargers repurposed to be auxiliary cooling fans. Netgain actually provides replacement "shrouds" for their Warp 9 motors and I had happened to notice that the outlets of these Garrett's that have been redone with electric motors to serve as an electric supercharger, would fit right over the round port on those shrouds. So we made similar shrouds for the 11 inch mmotors on the Escalade out of aluminum. Netgain liked the idea well enough that George's son Hunter had engineered similar shrouds for their Netgain 11 series. They come in two pieces, and we ordered 10 sets as we offer them now with the Garret blowers we call the BlowMe. We pulled a couple of sets and attempted to mount the blowers.

Unfortunately, Hunter's design winds up with the blowers mounted at a very odd angle mostly DOWN below the motors. I cannot really picture a motor mount scenario for a Netgain 11 where this would work at all. We called this to George's attention and again there wasn't much of an explanation beyond Hunter had consulted with a shop that does conversions and that's how they thought it should work.

So we had to hack up a couple of sets of these. They weren't simply reversible. In order to get the Blow Me's and the motor terminals to be more or less "up" on this motor assembly.

I suspect demand for such a device to be a little thin. It's a bit longer than most of the cars you guys are building and weighs over 400 lbs. But for some racing or larger vehicle applications, it offers a lot of power. If it doesn't sell, I've kind of got plans for a 2014 Ford F150 project. But the important part is that it is in stock. No six month delay. No six week delay. We can ship it tomorrow morning. It would drive a pretty good size boat for that matter.

We also continued work on the test bench. We have our battery pack assembled, a 12v DC-DC converter, and both a disconnect switch and an emergency slap switch. Using Collin Kidder's software and some coaching from Ed Clausen, we managed to get the DMOC645 and Siemens 1PV5135 motor to spin for the first time using the RechargeCar MACCHINA as a controller. This was kind of a big deal for me as we had no way to control this setup. It kind of proves you CAN develop a CANbus controller to drive the DMOC645 and hence the motor. At this point, much further development is called for, but the basic premise is kind of a done thing.

We did not really intend to be in product development. But this may well be worthwhile. This motor will do 297 Nm or 221 ft-lbs at 300v and 400 amps - 120kW which is just about what the DMOC645 will do. It's a liquid cooled package in AC induction format and gives you regenerative braking and the advantages of lower maintenance - no brushes to deal with. At $5000, we think it's a breakthrough for DIY EV builds. That's about the price of a Soliton1 and Netgain Warp9. Actually, we set that price to BE about the price of a Soliton1 and Warp9.

If we sell enough of them, I'm pretty confident we can get an unlimited number of these motors from Siemens. There should be additional DMOC's come to us through salvage from decommissioned trucks. And we're working hard on some other realistically priced controllers from Scott Drive and Rinehart. The key is to to establish tje fact that there IS indeed a DIY market for these devices and that they can be SUPPORTED in kind of a crowd source fashion so that the developers don't have to walk each person through the basics over the phone. We hope to accomplish this with documentation and user forums where we can share information.

If you want OEM quality components, this is the path. We have talked with developers with promising products for four years now. The mantra is always the same. We only sell to OEMs. Why do you only sell to OEMs? (I happen to know why they DON'T want to and now A123, Enerdel, and Siemens know this as well). But it is very simple. If they think they can sell 100 of them and have to provide 40 hours of product support, that's cool. If they only sell one of them and have to provide 40 hours of product support, that isn't cool. Hopefully, along the way they can persuade an OEM not to do something stupid with their product and wind up in the New York Times with a disaster. They do not have this control with you.

The reality of marketing in developing technical areas is you cannot predict the winners and losers. We now know, and the previous shareholders painfully so, that General Motors can indeed go bankrupt, and lacking an $80 billion handout from us, leave ALL their suppliers holding a lot of bad debt and unused parts. We also know that a 14 year old can dig a golf cart out of a junkyard and restore it to operation. If he's lucky, he winds up as the Chief Technical Officer at Tesla Motors - JB Straubel. I would strongly advise all vendors to note that selling to the 14 year old, might long term be in their best interest. And a contract with General Motors, or Fisker, or Azure Dynamics, Coda, Think, Aptera, et al is not necessarily a guarantee of success. I learned this on the Internet. The three happy yokels who founded Cisco Systems would not likely find a good reception from Siemens. Nor the Yahoo or Google founders. Steve Jobs of course would have convinced them he WAS a major OEM, standing dead broke and with no furniture.

Our strategy is pretty simple. IF we can form a nexus, a kind of buyers coop, that focuses the buying power of the DIY community into one identifiable entity they can deal with, with some indication of at least modest sales, and we can get YOU to support each other in ways that limit their product support costs, we can gain access to these components ongoing. We have to have a nexus they can sell to in promising quantities, and a way to limit their product support costs. We do now have a relationship with both Borg Warner on the eGearDrive and Siemens USA on these motors to provide product ongoing. I'm working on Rinehart and Scott Drive to add these inverters to the line. And with GEVCU and stallions like Ed Clausen and Collin Kidder, i hope to get us set for a cornucopia of very reasonably priced salvage parts from Nissan Leaf's Tesla Motors, Toyota Prius, and Chevy Volts that will rather inevitably become available in the future. With just a little bit of cooperation, we can form an entity where we can deal with all of that and vastly improve the tools available to DIY conversion builds of the future.

Jack Rickard