Re: "Racing" the Duff Beer Electric Car
Have you considered a constant current bulk charge for the initial charge? You should be able to get them in reasonably good SOC before having to kick in a balancing setup.
You are not logged in. Please login or register.
Have you considered a constant current bulk charge for the initial charge? You should be able to get them in reasonably good SOC before having to kick in a balancing setup.
Sure, that's basically what a battery charger does, CA, then CV. Now I just need a bunch of CA power supplies that put out thousands of watts for really cheap. Actually, I think that Ricky is working on something like that. Unfortunately, all of the high power DC power supplies that are cheap put out a constant voltage. If you connect one of those to a battery, then it makes a big spark and melts the charge connector. It's exciting.
Ah.. Ok. Since you are using about 140vdc you can use industrial DC motor drivers as bulk chargers. They have current limiting circuits. You set one up with the right voltage and with the current limiter dialed down, it could be the ticket for a reasonable cost.
I've been helping Roger in the background - I hadn't even though of using DC motor drivers. That is an excellent idea. They are built to deal with and control high current draw spikes (zero speed startup current), which in a charging application is in-rush current at battery hookup...
Must do more research.
DC motor controllers could handle the CA portion.
Charging needs to happen like this:
1. Constant current at some settable rate until voltage rises to a settable target voltage
2. After voltage target is reached, then constant voltage until current drops to<3% of rated current, but no longer than 30 minutes. You could actually ball-park this to 30 minutes at constant voltage and still be safe and effective.
- This transition needs to be pretty seamless. If you go CV before the battery has reached the voltage threshold, then the charger will try to instantly bring the battery to the target voltage. This will damage the charger and melt stuff.
3. No voltage or current. It has to cut off after 30 minutes.
Now for the fun part:
There are three sets of 147.6v battery packs in the car. Each battery is composed of three 49.2v packs. Each pack is composed of twelve 4.1v cells. Each cell is three 4.1v pouches, but we only manage to the cell level because the pouches are in parallel.
A cell should charge to 4.1v, but must not charge to over 4.2v, or it will damage the cell and possibly start a fire. Some cells will charge faster than others, so this should be managed.
The lowest chargeable unit is a 49.2v pack because that's the only thing that has terminals that will support significant power transfer.
The highest chargeable unit is a set of three 147.6v battery packs in parallel.
At all times, I need to be putting 9000w of output power evenly into a set of three battery packs. That can be broken down into whatever divisions of voltage and current you like, as long as it is divisible by 49.2 volts, the lowest chargeable unit.
All of this needs to be built and tested in the next two months.
I think that covers the charging logistics.
Wow. That is about 20hp in power. Hm. Will have to think about the best way to achieve that. Without breaking the bank.
We will dump some money in this thing if we have to.
Cost-wise, the conventional solution would be something like an Elcon PFC3000 charger for $1100. That would put 3000 watts into a 147.6v battery pack. We'd want an Orion 36-cell BMS for $890 to monitor and balance the cells. Round that to $2000 and we'd need three of those setups, so $6000.
The cheap option that I tried at Barber was EV-Peak A9 12-cell balancing chargers for $160/ea. We'd need 9 of those for $1440. Then we'd need to power them. They require 15v-32v DC, which we did with used 12v HP server power supplies in sets of two in series that can be bought for $15/ea. We'd need 18 of those, so $270. All together that's $1710. Round up to $2000 after I buy spares.
So our tentative charging budget is north of $2K if that helps. We have generators for the AC power. Sam's Club sells a 5500/6875 220v unit for $399 that is pretty good.
Based on my current theory that the charger blowuppiness is some kind of inrush-current/impedance-mismatch problem; I ordered some 500-ohm 50w resistors and built an output cap pre-charge dongle:
I plug in the charger output to this for a couple of seconds, and it smoothly (and with no giant scary lightning-bolt arcing like it usually does) brings the output filter caps in the charger up to within .1v of battery pack voltage. Then I remove the dongle (the output caps discharge some, but it is very slow) and plug the charger directly in to the battery (again, with no big fat arcs or other 'splodiness) and start charging.
I just did this with five brand new A9 chargers in a row and ran them all up to 1000w with no failures. I could not have done that at Barber. I will run them all through a couple of full charge cycles and see how things work.
More batteries arrived today, and a new popup. After getting all of our stuff wet at Barber, we are finally spending some $ on good popups and letting go of our flimsy $99 Wal-Mart specials:
How about something like this??
Those would work, but then I have to do cell management. That charger would require something that would balance 36 cells, which is an expensive battery management system. The cheapest/dirtiest form of cell management that I can find is a balancing board ( http://www.batterysupports.com/44v-48v- … p-268.html ), but those only go to 24 cells, which would leave me electrically breaking up battery packs to charge them, so I'd have to use a 12s battery balancer and three lower voltage chargers for each battery pack.
Craig is working on building a battery balancer that would probably work well with those chargers, but I don't really want to lean on him to hurry it up since I'm not paying him anything. If he gets something built that will handle 4Kw and 36 cells and doesn't cost very much, then those chargers are probably what we'll be using.
I have a 48v 12s eBike charger and 12s balancing board on order so I can test them, but it's still way more expensive than drone chargers, and the last balancing board that we tried exploded. They aren't really designed for taking them on and off of batteries continually.
Based on what you are seeing now do how balanced are the cells staying? You may not need to balance them on every charge cycle if they are reasonably matched. That would reduce how many balancers you need. Also if you don't try to charge them to the top amount you may lesson the need for balancing on every charge.
Unfortunately we don't have enough information to say. We've only turned 44 laps on the new batteries and we don't log cell voltages. I did see several that were over .1v off in a pack after one discharge, which could be bad if not managed. We never charge them all the way up or drain them all the way down. We stay between 3.2 and 4.1v. Also, they are not very well matched. I've pulled them out of four different vehicles, so they are from different lots and are in varying conditions. I also didn't keep track of which batteries came from which vehicle, so within one pack there can be cells from three different vehicles.
The only EV's that I see that don't manage cell voltages are LiFePo-powered, and that chemistry is a whole lot more tolerant of abuse. Nobody that I know of is running big LiPo batteries unmanaged; the failure modes are too catastrophic. That's not to say that we haven't considered it, but I'm not comfortable running these batteries without some type of cell management.
Well, That is about all I can contribute. If you want the contact info for the people that quoted me the battery chargers let me know. You would order those directly from the manufacturer in China. The lead time is about a month according to the documentation.
If I get back to the east coast in time for the CMP fall race I'll bring a generator.
Thanks, I'm guessing that we'll be using something like that whenever Craig gets his balancer sorted out. I'm really looking forward to Craig getting his stuff ready so we can race against another EV.
In our limited practice at Barber, they tended to discharge evenly.
Not really. Within 10mV is "even"; we're getting them way out of balance discharging them so rapidly. The more we do it, the worse they'll get. We're happy to get hundreds of cycles from them, whereas Chevy needs thousands of cycles over a decade.
Ricky tried one at Barber and it blew up...
I found a reasonably understandable engrish version of the instructions. They were very clear on connecting each cell, individually, in order. Or "the smoke will be generated". :-) Cell 13 was connected out of sequence for a millisecond and it fried the balance circuit -- the resistors for sure, the chip most likely.
...inrush current problem ...
And that problem apparently extends beyond vaporizing traces and capacitors. There's research that shows cells within unbalanced packs can actually reverse polarity on those spikes. They were investigating why BMS components were failing.
Using a raw Cisco 4500 PoE feed can work, but there has to be a current limiting regulator. Craig's playing with one I dug up that's pretty basic and cheap to build -- the main CC FET is the only special part, 'tho we may want to use a single chip "current sense amp" instead of doing the calculations for op amp feedback. (and those things can handle in-rush limiting ramp up) If you attach the PS with a limit resistor (I had some 10ohm 100W with me), it'll work perfectly. But it'll take about a week to charge. ;-) If you bypass the resistor after connection ("pre-charge" style), it'll also work... by dumping everything the PS can because the cells will happily soak up over 300A and the alarm light comes on. Of course, that's with zero per cell balancing or monitoring, which will end up boiling electrolyte. (I don't think those LG cells can actually burn -- "vent with flame". I've seen EV West overcharge the hell out of similar packs.)
Here's the frame of the new battery box next to the old battery box. This will let us hold three battery packs instead of two for 50% more power/range. We're looking to turn 30-minute stints at CMP next month. It should also move some weight forward of the rear axle:
I just got the new DC power supplies today. I ran one through a full charge cycle. This works well:
Three battery mounting plates in the battery box. This should give us 50% more power/range than we had at Barber. I could technically fit four battery packs in the box, but that may wait until the CMP season finale. I still have to run conduit and new cables for traction power:
And it runs!
I've replaced every high voltage cable in the car. The under-car cables are now encased in conduit. We now have a good fuse assembly.
We have a new shunt and digital multi-function display
The battery box is big enough for 4 battery packs, but we'll run 3 at CMP, just because that's easier for now. I took it on a run to the beer supply store, and there will be Stout!
Hell, I think this thing is ready to race!
I've been lurking here, because I am both interested in EV technology and horribly ignorant of it. So, nothing productive to add, other than my support and encouragement. Looking forward to racing the EV with our LTD!
After heaving some 120-pound batteries around, the current math is not looking good.
We need about 300 laps to really be racing. Less than that and we're just screwing around. While we do enjoy screwing around, we would like to race at some point.
Under normal driving, an EV gets about 3 miles/kwh. Our car or a new Nissan Leaf do about the same.
Under race conditions we get a little under 2 miles/kwh, which makes sense. Every car gets worse mileage under racing conditions. Translated into laps, call it 1.35 kwh/lap
We run the batteries down to 3.3-3.2 volts/cell, or about 80% utilization.
So we've got 18kwh in the car, and will use 14-15 of it to turn about 12 laps at CMP before we pit to change batteries.
In order to make our lap target, we'll need about 12 sets (a set is three 120-pound batteries) of charged batteries. I don't think that will even fit on our bus. Then we have to load/unload them, swap them in and out of the car, etc. I think that will kill me.
Looking at charging, we have about 7 hours/day during the race when we can charge and get the energy in the car.
9 chargers can put 12 laps back in the batteries in two hours. Each charger uses about 1100 watts. So that's about 1.65 kwh/lap. that also makes our charging around 80% efficient, which is about what I expected. The math is actually a little better because the chargers taper off at the end of a charge cycle, but close enough.
With the 3 sets of batteries that we have now, we'll need 27kw worth of generators running to make this work. That's three 10kw generators running full time. That's a lot of big generators. It's also more power than we can currently put in to the batteries with our current chargers.
With what we have available now, call it 10kw of generator power available for charging, we can generate 44 laps during the day. so now we need 9 sets of batteries. I don't think that I can move that many batteries around without hurting myself. With our current 3 sets, we'll top out at about 80 laps/day.
I roughed out a plan for a battery tray and some small carts that would let us use an engine lift to swap out battery sets quickly. It would also allow us to load and unload them quickly using the wheelchair lift on the bus. This would eliminate humans lifting batteries. I could also rig up some ramps and accelerate the process further.
I can probably fit about 6 trays on the bus.
With 6 sets of batteries, I'll need 18kw of generators to get our lap count to 150/day. That may be doable. I can get an 8K generator from Sam's Club for $650 to add to our current generators.
So for CMP next month, I'll probably only have three sets of batteries and 10kw of generator power, but I think that I will try to get the tray/cart system done and tested.
Both of the cheap Chinese DC power supplies have literally smoked. One throttled itself back to 15V, stopped smoking and seems to be stable, but now takes over three hours to charge a battery. The other still produces 24V, has stopped smoking, but emits a powerful burning smell when loaded. I will be consigning both of the to The Dustbin of History, which is what I call the enormous pile of junk behind my office. Some day I'm going to have to do something about that, but not today.
I reported my results back to the manufacturer and they duplicated my load/time in their lab, and sure enough the power supplies smoked there too. They say that nobody ever tested them near full load at 110V, which is only really used by backwards troglodyte fat stupid Americans, only at 220V, which reasonable enlightened civilized people use, and where they operate correctly. They thanked me for showing them a weak spot in their product, have updated the design to work at the rated capacity with 110V, modified a power supply to test the design, and will be sending me two updated units for testing when they get them produced in a few weeks.
They seem nice enough about the whole thing, but I don't think they'll get me the new units in time for spring CMP, so I ordered some of the more expensive Chinese power supplies that I've been using to charge a bunch of batteries. Based on my earlier battery math, I'll be needing more power later.
Robert wants to mount three power supplies and chargers in a cheap knockoff Pelican case. Everything will be pre-wired and solidly mounted, so we'll just open the case, plug it in to AC and to a battery pack, and fire it up. That should be an improvement over our current strategy of strewing crap everywhere with a bunch of wires running all over everything and a bunch of it sitting on top of the batteries being charged. Then it rains and our cheap pop-up leaks water over the whole mess and we're afraid to go anywhere near it. Hopefully we can get the whole charging issue sorted out at the spring CMP race.
Not that it matters, but all this sounds really expensive, especially compared to an ICE. Since you are typically, and excuse my French, one cheap bastard, this all seems crazy to me. I guess it's the whole journey vs. destination thing?
Pretty much, it's a challenge at this point. We want to turn 300 laps in this POS, except for Charles, who generally has more sense than the rest of us and wants us to scrap this effort and put a cage in my old time trial car (EG Civic with B20 swap) that's been sitting out behind the shop for years. It is kind of neat/weird to drive as well. If we get it buttoned down sufficiently, then we'll rallycross it too, but probably not this weekend.
On the cheap bastard subject, you are correct, but there is a caveat here. The electronic components retain most of their value unless we destroy them, and since I'm buying the electronics in bulk directly from the manufacturer in China via Alibaba, I can actually eBay all of it for at least what I paid for it, maybe more (with the exception of the stuff that I blow up). The A9 chargers sell for $250 and I buy them for $160. The PJ1 power supplies sell for $290 and I buy them for $145. The batteries are even better. I buy them in whole Volt assemblies from LKQ, but broken down into 48V modules they go for more than double what I'm paying for them on eBay, but I have to be DOT hazmat certified in order to sell them. So when I started this mess I enrolled in the DOT course to prepare for that eventuality. It cost about $150, and wasn't difficult, but took forever:
I'm pretty sure I can break even quickly and with little effort when this over, and I might even make enough money to prep the next race car, probably a cheaty-fast class A car so we can race with you guys. It'll probably be this, but without the supercharger:
Not that it matters, but all this sounds really expensive...
Indeed, but it's a lot of small checks. :-) He's going the cheapest route, which is actually not a paved road. In the process, we're blowing up a lot of stuff because none of it was ever designed or tested for the punishment we're going to throw at it. (And we're using server power supplies that likely weren't all that "good" to begin with. It doesn't help that we use them in ways the will blow them up -- in series to get 24V without a safety cutoff, so when one of the pair cuts off, the other one instantly fries it.) I have several very good power supplies (Sony built Cisco power-over-ethernet supplies) that can dish out many kW -- enough to literally vaporize an XT-90 plug, because it has without skipping a beat (the red "overload" LED comes on) -- but the stupid Chinese chargers only accept up to 34V despite needing to output 50! (it's an 8S charger they hacked up to be a 12S charger for agricultural drone. No joke.)
There are proper, commercial solutions out there to do exactly what we need. But those are very large checks to write, and we need many of them. If this were a normal EV, yes, we'd go buy a $3000 charger and $2000 BMS designed for the pack that bolts in the car and has the industry standard connectors. A proper charger setup is 150V@120A; that's slightly more than half of a standard household feed (220AC@200A) -- and 80% more than the RV generator can provide. When you need *12* of those, you start looking at $8 HP server power supplies and $200 rc car technology. Besides, it's LeMons! Do you really think any of us are sane?!?
Let's not forget this is all going down at CMP, where just a desk fan can pop a breaker! I have looked into renting an industrial 30-50kW generator, but we're nowhere near that at this stage.