This week we try to catch up a bit on the eCobra project, which I would like to finish sometime in THIS life. This car has been a struggle.
The bad news is it has just featured some really difficult geometry to fit in a 41 kWh battery pack in a small two-seat sports car. I'm very pleased with the power plant. The Netgain Warp 11HV and the Netgain Controls Industrial controller are very strong in this 3000 lb monster. We haven't done any performance testing yet, but I can already tell you the performance will be very strong.
This week we tie up some cooling system issues and some instrumentation. And most of all wire in our Elcon 5000w charger.
The Elcon has emerged as the default charger by virtue of its power and price. It is NOT fully configurable and that poses some problems and the guys selling them have just enough information to be dangerous about the charger and are actually doing some damage out there to cars. Their cocommittant desire to sell battery management systems with them being part of the problem.
We need the Elcon's 5000 watts because we have a 41 kWh battery pack. At full power from 240 vac, this would still take about 8 1/2 hours for a full charge. So we can't really use a smaller charger. And this unit is physically BIG. By lifting the body almost off the car, we managed to squeeze it into the rear trunk area right behind the roll bar.
We also wired it into our J1772 connector and the little AC31 board David Kerzel of ModularEVPower provided. Given the mirroring of the proximity switch pin and the copilot signal pin between the plug and the socket, I can NEVER get them right. Sure enough, I had them swapped in our install and it would not charge. Swapping the wires on the little circuit board solved the problem. The car now charges quite well on the Clipper Creek EVSE we had installed for such testing. Our eCobra should charge smartly from any available SAEJ1772-2010 EVSE.
I prefer fully configurable chargers such as the Brusa, which you can easily change everything about the charge process including the stages, currents, voltages, rest times, etc. But a Brusa is close to $4000 now and this project would have required TWO of them. Eight grand for a charging system seems a little bit much since my first car I drove at age 16 cost me $60 cash money hard come by.
The Elcon at $1695 from Evolve Electrics just seems a better buy. But you have to specify voltage and charge curve algorithm. We've had a lot of questions about these charge curves. In this episode, I talk a bit about why we use curve 502 and how you can add a little bit of "configurability" by having 10 charge curve voltages stored in memory.
Here's an image of the 500 series algorithms they ask you to choose from. We picked 502.
The document basically describes a very simple charge curve, which works well for these cells. Basically, you pump all the current you can into the pack, until it reaches a certain voltage. You then HOLD that voltage by ADJUSTING the current, until the amount of current needed to maintain that voltage diminishes to some set value.
The spec on most of the Thundersky and CALB and Sinopoly cells actually does call out this value - 0.05C. This is 5% of the rated capacity in amps. So a 180 AH pack would cutoff at 9 amperes.
This is not one of the choices. But AH/30 seems to be. Since we are at 180 AH, that would be 180/30 or 6 amperes.
With 69 two-cell pairs, we are looking at a charge VOLTAGE of 251.85 volts. AGAIN and I repeat for the 40th time, this charge voltage has almost nothing to do with the fully charged voltage of the cell, which in all cases will be something less than 3.4 volts if you are indeed using LiFePo4 cells. It is part of a PROCEDURE to get you to that fully charged state. We use 3.65v rather arbitrarily. The original spec on these cells was 4.2v. Then it was lowered to 4.0v. I think it is now 3.8v. It doesn't matter, we developed our OWN procedure that seems to work better. I feel ever more validated as the MANUFACTURERs' procedure is clearly dropping a couple of tenths of volts per quarter as we go along.
In any event, the terminating CURRENT is an important part of this procedure. You do not WANT to keep putting energy into the cell past this point. With 502, we actually ARE overcharging the cell. But since we used 3.65v to UNDERCHARGE the cell, it all works out.
The proof is of course in the pudding. Twelve hours after charging, we are seeing a pack voltage of 230.4v. This works out to a cell average of 3.34 volts - exactly where I like to see it. 3.35 would be good too. But with this many cells, I actually feel better at 3.34.
One thing stressed in the video is to actually observe and measure the end of charge activity before entrusting your pack to ANY charger. This is kind of important - under the rubric that in any product shit happens. You don't want it to happen TO your expensive batteries. So its worth the time to go through the process with the charger a couple of times to make sure it is doing as you THOUGHT you had it configured to do.
In this case, instead of charging to 251.85 volts we actually hit a peak briefly at 253.4 volts. But it settled down quickly to right at 253 volts and maintained that very nicely. The current tapered quickly since we are charging at a low percentage of capacity. We were seeing about 22 amps IN to the pack from this charger and it appeared to be a very nice 95% efficient comparing the current into the cells with the current from the wall AC. I was quite surprised by that.
The current tapered nicely and the voltage was pretty steady, minor wandering in the 0.2-0.4 v range. As the current decreased from 22 amps down to six amps, the charger terminated abruptly at an indicated 6.2 amps. This is actually very good.
Bottom line is that bang for buck, this is an excellent charger operationally. It is a bit of a pain to order and configure. But once installed it works well and is quite powerful for the price.
We also discuss in this week's video the announcement pending from EVnetics of a monster controller they are calling SHIVA. This controller will crank 3000 amps peak and 2500 amps continuous using EIGHT 600 ampere IGBTS. It is pricey at $7500 and they are only going to make 25 of them initially. But it promises to become the immediate darling of the drag race community. It probably has little application in a general electric vehicle UNLESS you happen to be doing a Cadillac Elescalade with twin 11 inch motors. In that event, if you paralleled the connection to the motors and backed it up with some 400 Ah cells, you could probably do 1500 amps into each motor briefly and tear up transmissions all the way between here and Florida.
I think this going to be huge for EVnetics. That the vast majority of their sales will be Soliton Jrs is not the point. Everyone will know the upgrade path to the BIGGEST dc PWM chopper on the block. And I actually think they'll sell out of the 25 run quicker than they think. The Tim Allen/Tooltime riff on MORE POWER worked for that show and that comedian because he hit the nail on the head. Just because I don't USE all the power under the hood, doesn't mean I don't want it there. At $7500 it's just beyond the reach of most builds. But I would goess the drag racing community will waste no time neatening up their wiring with this little controller. As always, the package SEB does is just gorgeous.
The IGBT's have an interesting feature. They have internal temperature sensing. What this means is that the coder guy, Martin, can actually throttle this thing back very quickly and very accurately based on temperature. ANd what THAT means is the better you are at getting cooling glycol to this beast, the more power you can get out of it. The spec limitations are 4800 amps actually but they've got it cut back to 3000 amps. I don't know the voltage drop, but if it's a volt at 3000 amps you also have a 3000 watt heater. NoGiven the forward voltage drop on six IGBTs, that's still a lot of heat. But there's room there if you beg.... and don't mind blowing a $7500 controller every other race. it's all software...
We did talk a little bit about the Tesla party. What I didn't mention, but am looking into is the Tesla/Panasonic battery. March 2012 will be the first run of a new 18650 cell by Panasonic that is really quite a thing. It uses a couple of innovations. They call it their Nickel New Platform or NNP cell. It is actually a Lithium Nickel Cobalt Aluminum Oxide (LiNiCoAlO2) cathode with a 3.6v voltage and 3.4Ah in a single 18650 cell. This works out to 46 grams and 12.2 watt hours or 265 wH per kilogram. By contrast, our cells are about 106-109 wH per kilogram. That's how you get a 300 mile range.
Interestingly, they have even bigger plans for March 2013. At that point they will do a production run of the same cell, but they intend to replace the carbon anode with a silicon alloy anode. This will boost the Ah rating to a full 4.0 Ah in this cell.
As to cycle life, there's some bad news but then some good news. To the standard quoted 80% of the initial capacity, this little cell will only do 500-600 cycles. But if you can deal with 70% capacity, it will go to 2000 cycles. And so given the 265 wH per kilogram, build in a little extra capacity. It will be a MUCH improved package over the Tesla Roadster battery module. And it explains the Tesla 300 mile range claim.
I found it interesting to note that since acquiring Sanyo, Panasonic may be the world's largest Lithium battery supplier. But it is also interesting to note that they are building factories to produce all this in CHINA. In fact, due to the strength of the Yen, they've dramatically cut back plans on a new factory in Japan.
Bottom line, the cell march goes on. Our 80 mile cars will soon be in the 250 mile range (really). But it will cost you. And of course, all roads lead to China.