Some Days Being Better than Others at EVTV: Lessons from an Electric Vehicle Enthusiast

Tough week. Somebody tell Selsea Chexton et al that this thing about electric vehicles being maintenance free is a fantasy of theirs, not reality.  Jason Horak and I don't see it that way.

It was bad enough that the transmission in the Escalade imploded.  I find it odd that it did so.  It is indeed a heavy vehicle, much heavier than stock.  And we are putting a bit of power to it.  But frankly, the transmission seemed to be shifting pretty smoothly.  And my driving style is pretty tame frankly.  Picture 0 to 30 in 10 seconds.  And 30 being about it 99% of the time.

Once in awhile, if I coasted downhill for 30 seconds for example, when resuming the throttle, the transmission would kind of "bang" into gear in what I would describe as an unnecessarily hard shift.  Oddly, the experts at LeGrand transmission noted that the gasoline version does exactly the same thing and not to worry about that.

But in the rebuild, they were very confident that what I needed was a HIGHER stall speed, not a lower one.  This is counterintuitive to me.  Diesel trucks all have lower stall speeds.  Lower stall speeds of course lockup the torque converter at lower rpm and this makes the transmission run cooler and last longer.  Diesel engines have more low end torque while gasoline engines tend to develop it at higher rpms - typically 3600 for a V8.  And so to my way of thinking, the electric motor is more akin to a diesel power plant.

But I suffer an ongoing malady.  I probably know more about more different things than anyone you'll likely encounter in a normal environment.  But about 90% of what I know happens to not be EXACTLY technically correct.  I'm keenly aware of this fact, which is the difference between me and most, who actually believe what they know to be true.    And so in any given situation or conversation, I have some thoughts as to how it might work, but I'm never really certain.

But in the course of things, I regularly encounter people who are VERY certain and VERY confident they have the right answer.  That their batting average over the past five years might be 8% of mine doesn't enter into it.  THEY are CONFIDENT and I am NOT.

So I tend to concede the point on the theory that they MUST know what they are talking about, because they are confident and that must derive from actual experience or data they are privileged to and I'm not.  Like Charlie Brown and the football, I learn over and over that those who don't know, generally are unaware they don't know, and in fact think they do - and all too often with my money.

So we are going back to a lower rpm torque converter.  That may fix the problem and return us to our original state.  And maybe the destruction of the transmission will then repeat in a year, or maybe not.  Or not.  I'm not sure.  I know we have to change the tranny twice, and buy two torque converters.

On the matter of the saltwater rheostat, we actually succeeded.  Up to a point.  At a higher voltage (13 volts) and with a LOT of salt, we managed 50 amperes in a fairly controllable fashion.  That implies that with somewhat less salt, and somewhat more voltage - like the voltage of a car, this would be an interesting way to drain a large pack.  But is largely useless in what I want to do, controllably load individual cells or two or three cells at 1000 amps that I can "tune" at least roughly.  We failed there and I guess we'll go to a rebar concept next.

A number of viewers were alarmed at the production of hydrogen.   I just generally toss a lit camel into those situations and they more or less resolve themselves.

The VW Thing is another matter.  And it is related to the DC-DC converter issues we've been having.  The issue is of course inrush current to the essentially infinitely low resistance of the input capacitors that exist in the DC-DC converter, the DMOC645 inverter, and indeed on the output of the Brusa charger.  See our existing diagram.

As I said on camera, one of the clues is that when we turn the thing off, it continues to power the JLD404 with 12v for up to 20-25 seconds.  The DC-DC converter input capacitors must of course drain down, and they are likely being aided by the input capacitors to the DMOC 645.

Mark Wiesheimer and Brian Couchene both advise me that inrush currents at these levels can indeed weld contactor terminals and Brian actually has had this experience and measured the current pulse with an oscilloscope.  His solution is an NTC Thermistor on the input to the DC-DC converter.

I was under the impression, largely from the Solectria manual, that the DMOC had its own contactor and precharge circuit internal.  Mark Wiesheimer assures me not.  We have had it apart actually and it's true I don't recall seeing one.  So in going to the 645 model, apparently they dropped that.

An NTC Thermistor is not going to quite cover that.

We have managed to eliminate the bootstrap switch to jump the car.  We simply take the aux battery as the input to the slap switch AFTER the charge diode.  In this way the slap switch becomes ON/OFF for the contactor K1 and even without the DC-DC converter it will fire it up.  Once the DC-DC converter comes up, it will actually provide the 13.6v through the charge diode to both the battery AND the contactor.

The input to the DC-DC converter is a little more complicated.  Largely because we need to develop something to handle DC-DC converters and hopefully eliminate all the problems we've been having.  Understand that while this is a high voltage circuit, it is pretty low current.  To do 800 watts at 13.5 volts is about a 70 amp max output.  But at 330 volts, that same 800 watts is more like 2.65 amps.

One of the issues we have with DC-DC converters is the case where you step on the accelerator and the pack drops instantly from 330v to 280v for example.  If we have an air conditioner inverter for example or a DC-DC converter with input caps, those caps instantly discharge from 330v to 280v back OUT of the input to the pack.  This can itself be a rather large current.  Diode D1 is a 400v 100A diode.  It allows current IN to the DC-DC converter but not back OUT and so the caps cannot discharge backwards out into the main battery pack.

We also have a theory from Jeffery Jenkins of EVnetics that noise from the controller/inverter switching can exhibit on the cables from the pack and so at the input of the DC-DC converter - sufficient ripple to exceed the current capacity of those same caps.  We have employed a fairly substantial coil here of about 50 uH to filter that ripple.  But what happens when we remove that voltage with K1 or the maintenance switch.  L1 could very well cause its OWN spike on the input exceeding the voltage rating of the caps.  Fortunately, the diodes we use are actually double diodes.  So I'll use the other half to act as a flyback diode across L1.  In the event contactor K1 opens, this should allow L1 to discharge safely through the diode.

We are going to adopt Mr. Couchene's NTC Thermistor as a current limiting device on the input.  But there are a couple of problems.  An NTC Thermistor is a Negative Temperature Coefficient resistor.  Quite the reverse of the Positive Temperature Coefficient resistors we use as heaters.  Instead of going UP in resistance as they heat, NTC Thermistors go DOWN in resistance quickly as they heat. So we can start with an 8 or 10 ohm thermistor, but when it heats up, it will drop to 0.1 ohm or thereabouts as it heats up.  So we will limit inrush current but then go down in resistance allowing full voltage to the DC-DC converter.  This is a very conventional approach with such power supplies.

But NTC thermistors have one little problem.  They need some time to cool to become effective again.  So if we hit our slap switch and then re-engage it immediately, we don't have any resistance to slow the inrush current.

The solution is to put a relay, K3, across the Thermistor.  This is activated by the 13.6v output of the DC-DC converter.  So initially, the Thermistor resists inrush current.   Once the caps are charged, the DC-DC converter will immediately put out 13.5volts and that will engage K3, bypassing the Thermistor.  Once the current is off the thermistor, it will immediately begin to cool.  Subsequent slap switch activations and reengagements will not affect the operation.  And of course the loss through the small resistance of the Thermistor will be avoided as well.

Or that's the theory anyway.  We'll try to get this wired up this week to see if we have better success.  If we do, we will probably package the diode, coil, relay, thermistor, etc into an enclosure and offer them on our online web store.

But I'm never quite sure... And we'll have to run some of these for awhile to see if they cut the death toll on DC-DC converters.... Probably throw in a fuse for giggles as well.

Jack Rickard