As many of you may have picked up, I’m not really by standard definition a “car guy”. I don’t really like cars. I don’t even like to drive. When we go out, my wife normally drives. Strange thing I wind up doing given that context.
But I am a battery guy. I’ve always been fascinated by telephones and batteries. That you can carry an electrical device around without a cord, never ceases to amaze me. At age four I thought flashlights were magic and I just never could put together how they could magically make light, and then how they would run down and COULDN’T make light. My later education in electronics didn’t help much. Batteries were always kind of glossed over, largely because the instructors were electrical engineering types, batteries were kind of chemical, and beyond the basic textbook explanation, we kind of moved on to acknowledge that they could indeed produce voltage and current – for awhile.
Batteries have changed a lot in my lifetime. I recall entering a store called the Hobby Horse at age six and asking the guy for a dry cell and a minature socket. He brought out a little blow up clown with sand in the bottom that you punched and it swung back up. He didn’t know what a miniature socket was nor for that matter a “dry cell”. I was confused that this adult was showing me a plastic toy when I wanted a battery and a bulb holder for a very basic experiment I found in a book. My disconnect with the populace of planet earth was pretty well completely formed by age 10 and I became convinced that I was supposed to be delivered to another planet and there was some heroic mixup in the postal system of the universe that had created this mess with me here instead. I’ve never really changed that position. An uncorrectable if tragic mistake that I will just have to learn to live with for now.
That you could move a car with batteries is indeed one of the most exciting things I’ve learned along the way. A 1979 experiment with a Ford Pinto was less than satisfactory. We had no PWM controllers. You switched batteries around in various combinations and an 11 mile range was the best I could do with a Baldor aircraft starter and some lead acid cells. I also learned a lesson about unobtainium as I ordered and paid for $5000 worth of Nickel Cadmium aircraft batteries and waited over a year before receiving my money back. As I was making about $19,000 per year and it WAS 1980, you can comprehend my lack of sympathy over battery prices today. IT would be the equivalent of a $50,000 purchase. After ayear, they DID send my money back. They simply had been unable to produce their brochure.
The Pinto kind of haunted the garage for a few years and was finally towed off to the junkyard.
Four years ago I ordered some very bad white Seiden LiFePo4 cells from China. 12 weeks later they showed up in my driveway. I had almost forgotten them. After a few months of playing around with them, it occurred to me that they would in fact drive a car somewhat better than the lead acid cells and I asked Brain to obtain a suitable lightweight vehicle. He picked a Speedster replica. I didn’t even care what KIND of car it was.
But I fell in love with the Speedster. And I ordered seome Thundersky 90Ah cells to power it. We started work in September and rolled first on Christmas day. To say I had an EV grin was an understatement. I had calculated a range of 46 miles – we did 100 on the first charge. It went 94 miles per hour the first week. I had expected a very capable golf cart. I got a Porsche Speedster instead. Game on.
The central issue to me at the time was “If people knew about this shit, they’d be doing it RIGHT NOW”. Some early EV pioneers had also ordered some Thundersky batteries in a group buy. They immediately tried to “equalize them” after the fashion of their beloved lead acid cells and ruined them all. I was following their progress quietly but intently. They’ve been heroically rewriting that historical episode ever since and I’ve watched that too. This was the beginning of the BMS crowd, which I struggle with to this day. They STILL do not understand these batteries.
[jwplayer file=”news083112-iPhone.m4v” hd.file=”news083112-1280.mov” image=”http://media3.ev-tv.me/news083112.jpg” streamer=”rtmp://s3einxnpkaij93.cloudfront.net/cfx/st/” provider=”rtmp” width=”950″ height=”584″ html5_file=”http://media3.ev-tv.me/news083112-1280.mov”]
As they come from China, the instructions with them were a little sparse. We were supposed to charge them to 4.2v per cell and Winston himself said that charging them to lesser voltages would “damage” his cells. The literature was totally rife with lead acid myths and carryovers. There was nothing to go by. We looked for Battery Management Systems and I actually bought several – typically at $2000-$2500 each. One from China had so much equipment and cables that the idea of instaling it in a Speedster was preposterous on the face of it. I would have needed Houston Control to mount all that stuff.
And so I set out to learn all I could about large format Lithium Ion batteries. But as noted, there was almost no information available. I started hounding the Chinese by e-mail, and bit by bit did gain some communication with them in Chinglish. A guy in The Thundersky Yahoo group and I even devised a current shunt that worked with an LED and a Darlington transistor and I got it to shunt about 7 AMPERES across the cells. I also learned about thermal runaway of semiconductors and worse, that top balancing cells CAUSED cell death at the end of a drive – if you drove it far enough.
And so I started aquiring better volt meters and ammeters and power supplies and loads and experimenting with cells. I was crushed when I would lose one as they were expensive. Finally I decided to GET OVER THAT and just destroy them on purpose to see when and wny they failed. Enough accidents. Let’s just do it on purpose.
I got to the point with Thundersky’s that if I undercharged them to about 3.6 volts, they behaved very well and I really didn’t have any less capacity than the spec sheet – often more in fact. Then I bought a Thundersky charger for 24 cells, which was what I had in the GEM. When I got it, it was really two little chargers bolted together. And it had NO controls on it at all. In fact, I opened the box and looked all through it. There was not a SINGLE trim pot in the entire device. It was hard wired for an output voltage of 87.4 volts as it turns out. Very interesting. Thunderksy, who still had 4.2v on their spec sheet at the time, was selling a charger for a fixed 24 cell system that charged the cells to 3.65 volts. I was off by 0.05 volts from what they were DOING not what they were SAYING.
The experience along the last four years has made me TOTALLY cynical about manufacturer specifications, battery experts, and the entire body of literature available on these cells. I distrust my own measurements and experiments, constantly questioning that I could be wrong, because of the variance between what I see and measure and what I read and hear. And over time, even more distrustful of others. I actually got in an argument with a guy on Endless Fear Sphorum and he sent me some graphs that were FAKED ENTIRELY. He just made these graphs up with software with no measurement at all. They had no outliers, no anomalies,they were gorgeous – albeit impossible. I questioned him on how he did the measurements and the story came unglued. I did the purported measurements, and showed my messy graphs that showed exactly the opposite, with the methodology I used and noted that he was either a liar or had faked the graphs entirely. He simply disappeared from the discussion.
So am I a little cynical about “experts” and self proclaimed “authorities” on the topic. Duh yayuh. Just a tad.
I also learned some interesting things about the Chinese. They will fill any blank on any form you want to present in order to close the sale. But consistently they UNDER promise and OVER deliver. Americans not only OVER promise and UNDER deliver, but there is a recurring admonishment that the Chinese products are shoddy in construction, unreliable, and probably dangerous – all coming from American companies in the space of course. The cynicism grows, and my nationalism fades, leaving two enormous fears. The first fear is that we will all wind up working for the Chinese. They believe this is their century and I can’t make a case. The second fear of course is that they won’t want us at all and will hire Brazillians instead.
In any event, the Sky Energy company received an enormous investment from the China Aviation Missile Academy and was reborn as the China Aviation Lithium Battery Cmopany – CALB. We had kind of developed a bias in favor of the Sky Energy cells and they continued to manufacture them. We used them in most of our builds, although we also used a number of Winston battery company cells as they provided some cells and advertised with us for awhile. Over a year ago Keegan Han, the U.S. marketing rep for CALB, showed us some photos of their new “grey cells” that they were introducing. It was always a month or so away and this went on for over a year. Finally in June they became available. We got a few and tested them and not only ordered some, but began to discuss selling them.
Keegan provided us with an internal document describing the gains of the new grey cells. One of the claims was dramatically improved cold weather performance. So we set up a test of the SE180AH cell compared to the new CA180FI cell. The results were persuasive. The CAFI cell was dramatically better at -20C or about 0F. We also graphed the charge and discharge curve and noted that it was even flatter than the SE cells. We actually like flat charge/discharge curves as this moves a greater percentage of the capacity of the cell onto the safe “flat” part of the curve. This means we can cut off charging later, and cutoff discharging later and still be safe.
One of the claims was a 70% improvement in output power. All lithium cells face a design tradeoff. If you put more active material on the aluminum current collector, you get more energy in the same weight battery. But the migration path of electrons from the current collector to intercalation sites in the LiFePo4 material becomes longer and more tortuous, and at the same time, the migration path for lithium ions out of the electrolyte into the same material to join with the electron and so intercalate into the material becomes longer as well. This reduces the amount of current you can deliver at any one instant. And so the ENERGY density of a cell, that is its ability to store power, and its POWER density, it’s ability to deliver power at impressive current levels, are almost mutually exclusive. We want more energy density for range. ANd we want more power to actually send 1000 amps for 15 seconds to our Soliton controller for example. This then is the central design tradeoff in Lithium ionic cells.
The CA180FI cells mostly come int at 196-198AH – typical of the Chinese propensity to under promise and over deliver. The CA180FI cells are expensive for 180 AH cells. They are pretty INEXPENSIVE for 200Ah cells. By contrast, we’ve NEVER had an American made A123 20Ah cell actually achieve 20Ah of capacity. Just doesn’t happen.
But there was also a claim of 70% improved power output. Now I struggle with this. What does it mean? 70% of what. I have to assume that for the SAME decrease in output voltage, it would deliver 70% more current. Or conversely, might it mean if it delivers the same current, would their be a 70% decrease in the voltage drop? Or what?
A few weeks ago we carried some video of Damian McGuire heroically challenging a CA180FI cell we sent him with a SPANNER. A SPANNER is like a wrench, only white hot. He shorted a single cell with the spanner and recorded currents up to 1976Amps as I recall at least momentarily. And it appeared to hold 2.90volts at some impressive current levels. But the whole thing was sufficiently quick and uncontrolled that I couldnt’ make much of it.
I don’t have any real way to test 2000 amps beyond sticking a coat hanger in a bucket of water, hooking up a shunt to it, and praying that I don’t blow all of us to Kingdom Come.
So on the assumption that the 40AH cell, the 60Ah cell, the 100Ah cell and the 180Ah cell are all different sizes of the same chemistry and construction, we ordered some 40Ah cells from Keegan and were disappointed to learn he was out of them. In fact, supplies of all the CA cells were uncertain as they were selling somewhat better than expected. Finally, they arrived last week.
The 40Ah cell is somewhat more naaageable. 5C would be 200 amps and 10C consequently 400 amperes. I have loads that when used together can do 400 or 500 amps for brief periods in a controlled fashion. These are constant current loads.
Testing a single cell is problematical. When you only have 3.3v to start with, and it decreases to 2.5v, in order to do 400-500 amps you have to have very low resistances. Cables and connections and so forth become significant and you find you just can’t make the current, not because the battery won’t put it out, but because you can’t load it sufficiently.
So we like to put 3 cells in series and get up to 10v or so to start. Our constant current loads can then maintain constant current and we can actually read our meters without going back to the video later to try to figure out what happened.
So we put three CA40FI cells in series and hooked it all up with a very accurate voltmeter and ammeter. We’re within a hundredth of a volt and perhaps 0.3 amps on 100 amps.
I dont’ really need to get 40Ah SE cells to compare. First I cannot. But we had done a dynomometer test on Speedster Redux with a Soliton1 where we DID achieve 1000 amps, any way you want to measure it, by flooring it for 15 seconds in third gear under load. We had 188volts in our pack when we started, and 147volts at the point of highest current and lowest voltage. The SE cells did it pretty well, but our pack voltage dropped from 188volts to 147 volts a decrease of 41 volts and 23.4% of our original voltage.
This is kind of important actually. Had we maintained 188 volts, at 1000 amps, we would have had 188kW driving that car. In electrical horsepower, that is 252 horsepower. Unfortunately we didn’t get 252 horsepower. The voltage dropped to 147 volts and at 1000 amps that is 147kW, a 41kW drop and more like 197 horsepower. In other words, we lost 55 horsepower to voltage sag.
1000 amps for a 180Ah cell is a current rate of 5.55C. In other words, 5.55 x the 180 rating in current. The higher the voltage at that current rate, very simply the more power and the faster the car will go.
In this weeks test, we attempted to duplicate this with the 40Ah cells. 5.5 x 40 is a much more manageable and measureable 220 amperes. So we set up the load to do 220 amps and switched it on for 15 seconds.
The results were nothing short of astounding. After 15 seconds at 220 amps, we had dropped from 9.935 volts to 9.263 volts – a drop of some 6.76%. This is a bit over a quarter of the drop the from the SE180AH cells at 23.4%.
What would this mean in Speedster Redux? Well, it means our 188 volts would have dropped to 175.3 volts and we would have produced 175.3kW instead of 147 kW, and produced 235 horsepower-e instead of 197 horsepower-e. With the same size and price of battery cell. That’s 38 horsepower for free.
We repeated the test at a 30 second drain and the drop only increased to 7.66%. Again, compared to 23.4% at 15 seconds for the SE cells.
So we decided to go for broke and up the measurement to 12C or 480 amps. At 15 seconds, the drop increased to 15.94%. But that is OVER TWICE the current level and still over 7% LESS voltage sag than the SE cells. So we did a final test at 30 seconds and 12C. ANd we were STILL less voltage sag at 17% – more than 6% below the SE cells STILL.
Accelerating at full current for 30 seconds is actually preposterous. You will never do this in a car. You could go zero to 100 in a little over 20 seconds normally. If you can’t you are doing direct drive and simply lack proper gearing for your vehicle.
Where to go with this? It pretty much obsoletes the A123 cells here at EVTV. True, they will do 20C. But the packaging requirements have proven daunting.
We have a new speedster coming our way from B&B Manufacturing in Granby Missouri. Recall they built the eCobra, which last week we welcomed back for some minor work before delivery. This Speedser will feature a carbon fiber body and a molytube chassis with rack and pinion steering and an IRS rear end. IT is supposed to come in at under 1000 pounds before adding wheels.
We’ll use an AC-50 motor from High performance Electric Vehicles on this car – a three phase AC motor with a Curtis 1238-7601 controller. That controller can be fitted for 126volts but not over 130. And so sag is an issue. But it can also do 650 amps. If we had a 70Ah CALB cell, we have demonstrated it can do 12C or 840Amps. 38 cells at 3.3 each would give us 125.4 volts. At 840 amps, we could expect a voltage sag of 16% or 105 volts. Of course, it won’t be that bad because our controller is limited to 650 amps. Let’s call it 12% and 110 volts. That’s 71.5 kW and 96 horsepower. That is not only FULL power, but much fuller power than we have ever been able to produce from this combination.
That’s a very small pack at 8750 wH, but we should be down around 1550-1600 lbs total for perhaps 160 wH per mile. And so we could still hit a 55 mile range at 100% DOD or 43.75 miles at 80% DOD. That’s still four miles past the 39.4 daily average in America.
And the price becomes attractive. At a little over $100 per cell, we have a battery pack at $3800. That’s getting down toward Pb AGM cell prices.
ANd so the small pack short range vehicle at a lead acid price we were shooting for with the A123 cells becomes viable – without the packaging nightmare. And it’s available now.
So we’re a little excited about batteries. How about weight? The 70Ah cell should come in at just over five pounds each for a battery pack between 190 and 200 lbs. The curtis and motor come in at about 140 together. The wheels will probably be 160 lbs. And that’s 1500 lbs for the car. Add a charger and DC-DC converter and cables and boxes and we will probably be 1550-1600 lbs. That’s 450 lbs LESS than Speedster Duh, which was a 550 amp Curtis 1238-7501 controller. With the better voltage, the 100 additional amperes, and the 450 lb weight loss, I’ve got to believe this car will get out of its own way. I was actually enormously pleased with Speedgter Duh’s performance. But this one should scream. Rack and pinion front end and independent rear instead of swing axle should improve the handling.
AND Brian Anderson advises us he has a 6 x 9 inch channel 40 inches long under each door. It so happens, the 70Ah cells are about 5.5 x 8.5 inches and less than 2 inches thick. I could put 20 down both sides (actually I only need 38) and have NO forward cells and NO rear cells. A heroically low center of gravity and a MUCH improved polar moment on this car. It should handle heroically better, steer better, accelerate better, and literally dance down the road at 1600 lbs. Range anxiety? Keep your range. I don’t need it. I need that Speedster, and it may just convert me to BE a car guy. If that won’t, nothing ever will.
72 thoughts on “BATTERY JOY AND THE CAR GUY”
Capacity Balancing the Battery Pack:
All kinds of effort is spent balancing battery packs, top, bottom, middle, upside down, and right side up. I propose that capacity balancing is the true way to go. All this balancing is due to different capacity of the cells. There are tab type lifepo4 cells available in 1100,2200, & 3300 mah increments. I say mount these on top of or wired in parallel to your lower capacity cells that brings them up to the capacity of your higher capacity cells. Then you can individually top off the cells to a common set point with a Cellpro Powerlab and check the capacity variance. As the complete pack is drained during normal use any individual cell groups that need more or less capacity would show up as higher or lower voltage individual “parallel cell modules” at the lower end of the charge. What I think you end up with is a pack that is very difficult to get out of balance whether it’s balanced at the top or the bottom. Jack originally started down this path with the concept of sorting out the cells into premium groups of increased ah capacity. I propose that fine tuning these “parallel cell modules” to equal capacity is of more importance as a balancing technique.
The negatives far outweight any advantages to this concept. But beyond that, they change nothing with regards to balance. Balance, indeed has nothing to do directly with capacity. We use it to ensure the cells ARRIVE at the same point in regards to Depth of Discharge at the end of the discharge cycle. Others use it to make sure they all arrive at the 100% SOC at the same point. That they are different capacities can complicate this, but it is actually NOT the same thing. You could have two perfectly identical cells of EXACTLY 100.00001 Ah with one at 10% SOC and the other at 80% state of charge and both in the same string.
Jack, I am not saying that bottom or top balancing can be done away with. I’m saying that making all the cells have the same capacity makes the balance stay more aligned throughout the charge / discharge curve than if the cells have different capacities. Other than the small parallel cell failing I don’t see a downside. It’s not like a BMS but more like parallel battery strings which you have run in the cars without a problem. If you foresee a problem with this let me know what it is.
Jack, if you have time, there are still a few things that aren’t clear to me about charging lithium cells, which you might be able to clarify. You’ve said that the “charge at .3C to 3.6V then taper to .05C” formula that CALB sends is only one possible formula for charging a cell. I took that to mean that it is the formula for a .3C charge. Based on the tests in your video this week, I calculate the internal resistance of the new cells to be about .002 ohms, which is impressively remarkable for a 40Ah cell. When charging at 3C as you were (i.e.,10 times the .3C rate), does the knee still stay in the same 3.45V-3.6V range, or does the knee also move up closer to 3.65V-3.8V?
Also, you’ve said that the natural resting voltage of the cell is 3.4V. So, would another valid formula be to just charge a cell to 3.4V at whatever current, then hold the voltage there and taper the current to near zero amps? Would that fill the cell just as completely as charging to 3.6V and tapering to .05C? Or is there something magical that happens when you charge the cell over 3.4V then taper?
Generally ALL charge rates use the same procedure – charge to a fixed voltage and hold it there until current diminishes to a fixed value – usually C/20. With the new CA cells, fast charge at 3C gets you further up towards 100% charge before the voltage reaches the CC/CV point because the curve is flatter.
But nothing else really changes. The probably generally is making enough current from an affordable charger.
Enjoyed the show again this week. I have an AC50 for the Civic and CALB CA100s on the high seas from China so this weeks show left me thinking that was a good choice. I believe, BTW, that the Civic we are converting is mechanically similar to the contemporaneous North American model but the body style is different. Also our base vehicle is the Turbo diesel which I suspect was not common in the US,
He would probably find it comic, but I greatly enjoyed Richard’s accent: reminds me of a Christmas spent with a delightful family in Louisville in 1976.
I would not have previously recommended a 100Ah cell with a full vehicle of that size. But with the CA cells, yes, I think they will do just fine. Occasional forays into 6.5C and normal operation at 2-3C will be very easy for these battery cells. I would not have done that with SE cells frankly.
I am excited to hear in this episode that you will be carrying 40-6 braided straps for the CA60FI cells. Will you also be carrying 50-6 braided straps? The 50mm straps would be ideal for end-to-end interconnects on the CA60FI cells.
We will design and have manufaxctured straps for all cells we carry. Right now we are talking with them about a 70Ah cell order and they are trying to push us toward the 60’s they have on hand. We’ll have to see how it comes out.
Are you able to divulge the cost of the “super speedster”? Is this a one-off vehicle, or is it in production?
You’re right about some of us getting all hot and bothered about the CA cells performance. What an amazing difference! Considering your advice about “less stress = longer battery life”, do you think that the new 100Ah cells would be a good match for a Soliton 1, as long as repeated drag racing isn’t the plan?
Great show this week!
You know I’m not sable to divulge it because we haven’t started the project yet and I have little information on what it will cost to complete. But with either 60 or 70 Ah cells, much less than previous Speedsters.
We’re talking with B&B about carrying the roller at about $25,000. But it’s all graven in jello at hte moment.
Enjoyed the show and the weekly article that you write in your blog. In fact I enjoy your weekly article as much as the video. You are such an excellent writer and storyteller. Are you writing a book in the near future?
About the Speedster Redux SE voltage sag: the cables and bolts in series also introduce a voltage drop which increases when current increases. Shouldn’t that be partly the reason why there is that much voltage sag discrepancy between Redux battery pack and the CA that you just tested?
We have plenty of cables and connections on the test bench. No I don’t think the car connections add measurably to the voltage drop. Were this the case, it would show up quickly as heat and probably blow something explosively at 1000 amps.
Interesting idea, but no.
Pearls before swine:
Excellent show. What is it about bottom balancing that is difficult to understand? After seeing a few disasters, the idea that once the pack is charged “TURN OFF THE JUICE” seems to be obvious. As you have said and demonstrated there is simply no point in attempting to get every last bit of energy. Keep up the good work eventually it will sink in.
At the risk of typing myself stupid … again … I have a question that I haven’t seen posed before.
If a battery A sags less at the same current as battery B with both having the same AH capacity. Would this mean you’d get more range from battery A – especially if driving with a lead foot?
Never mix disparate cells. Your pack is only as good as the weakest in the string.
Don, I can sell you a pill. Only $2,500. It will make you live forever and it comes with a lifetimes guarantee.
Once again, John chose well. A diesel. Low rpm, strong clutch, gearbox built for high torque and the brakes/steering are electric so don’t need a vacuum pump.
Jacks latest Speedster. Very exciting indeed. Would this do 160WHr/mile?
While I’m talking about consumption; 1W/s = 1 joule. So a ready reckoner for the metric lads in the EU will make this car
Jacks 10% rule: 1600lbs,160WHr/Mile convert to metric = 160*3600 / 1602 = 359Kj/m.
1600lb =725KG /2 = 362.
Or 362Watt seconds a metre. Only 1% out.
So, simply divide Kg/2 for Kj/m. 😉
The best part about this calculation is the readout will look very technical which will be the envy of your friends and the girls will want your babies.
Andy, how are the parallel connections of these cells outside of each other any different than the internal parallel connections. If we aren’t willing to try something different than has been done previously we learn nothing. Without trial and the risk of failure there is no success.
You can parallel differing cells on a per parallel basis but where have you mentioned this earlier?
Failure is no success so if you drive an experiment to fail you have re-emulated the obvious. It’s all been done before. Why so many times simply flummoxes me.
I put forward a query about a chance to upset a bottom balance by continuous cell charging at very low currents above their resting mean voltage. It was just a thought that simply does not matter because I do not aim to ever “go there”.
The answer is you should get more range. You will get more energy out of the pack. The only time where battery amps is the same as motor amps is when the motor controller is at 100% duty cycle. The motor controller is generally not at 100% duty cycle because it is limiting the motor or battery current. You only reach 100% duty cycle at the point where the motor is no longer trying to pull more amps than the current limit is set for. This generally happens somewhere between 3000 and 4000 RPM. You can tell when it happens if you have an ammeter on the battery. The battery current will go up as you accelerate and when it peaks and heads back down the motor controller is at 100% duty cycle. The only case I can think of where you wont get more range is if you operate a significant part of the time with the controller at 100% duty cycle. I don’t see how you could really do this in a car because you either get arrested or you overheat and breakdown. In any case the range increase will be modest at perhaps a few percent.
“I could put 20 down both sides (actually I only need 38) and have NO forward cells and NO rear cells. ”
It’s my brother that has the PHD in physics, but isn’t it usually best to keep the weight not only low, but as close to center as possible? Re Ice skater spinning with arms outstretched vs tucked in.
Yes indeed – low polar moment of inertia, the aficionados tell us, makes a better handling car. Can’t tell the difference myself though.
Those arms will still be a LOT closer to center than is the case with fore/aft packs, and lower too. Jack doesn’t say but I imagine the location of those channels is a result of structural considerations. Hanging the packs from the middle of a belly pan might require more structural reinforcement than it would be worth.
Great Blog & Video this week. I can’t wait to see how these CA cells perform in the EVThing. Since the PLC and HMI I am using can independently log pack voltage and current accurately down to about 100ms increments, I hope to be able to offer you guys some good real world data to chew on about how these thing perform over actual extended drives.
This week for me is all about getting the batteries mounted…. If I can pull that off I have a good chance at getting this “Thing” running by EVCCON. (However, I might still be wiring and add parts in the parking lot…)
Using information from your April 16, 2012 blog I built a budget bottom balancer for under $100. I thought some of your readers might be interested in it as an adjunct to the Revolectrix Powerlab 6/8. Detailed build and programming instructions can be found in my EV Motorcycle build blog at: http://st1100ev.blogspot.com/2012/09/budget-bottom-balancing.html
If you would like to include it in a future episode of EVTV, I can take better pictures or a video… although my video skills are severely lacking.
I like the bottom balancer Stan. I’m using an Arduino and a Mosfet rather than the JLD404 and relay – but the same principle. It cycles impossibly rapidly initially. I leave mine connected for a couple of hours after it starts cycling. I find the hardest thing to source cheaply is a good resistor/dump load capable of handling the power. 40 Amps at 3 volts is a rather uncommon scenario and my last lot of “100 watt” resistors got warm enough to char paper at about 50 watts
I suppose that if you use the same procedure & equipment to bottom balance a pack of cells that they will all be at the same SOC at the bottom but, does balancing at a lower current draw have any drawbacks other than taking too long? Does creeping down slowly to the balance voltage make it any more closely balanced, or does drawing a reasonable current more accurately simulate the cells being drawn down under load in the car. Assuming one stops pulling pack current at 90% DOD (in the car) it probably doesn’t matter, but which method do you think results in a “better” balanced pack, or does it make no difference either way.
I’m not sure it matters either way. The higher current (40A/0.66C) method bounced back and fourth (on/off) faster and more often than the lower current method. I switched to the lower current (20A/0.3C) because of too much heat from the resistors and wires. In the end both methods seemed to do the job and both were well under 1C. After I complete the pack, I’m going to go back thru one more time to knock ’em all back down to the same voltage… it should go rather quickly with most of the energy already removed.
I used the 12 V systems of two motorcycles to initially drain the cells by connecting four into a series and turning all the lights in the bikes. The final step of individually balancing the cells I did manually by connecting a single cell to the ME1003 motor I used until the voltage dropped to about 2.65, switched to the next cell and repeated this over and over again as long as I had patience. Last time I checked the situation on an empty-ish pack the cells were within 4% of each other in voltage. When I get my 25th cell I may improve upon on this to improve the balance a bit more. This was just to let you know you might be able to use the motor for the balancing, perhaps initial draining too, which I didn’t do, and get some brush seating done for free while doing it.
3x 2.5KW household kettles should be ~40 amps. 😉
John, I use a 10 meter 15Amp 2 core and earth extension cord with all three cores paralleled. Different combinations can give you any current you want. I used a similar arrangement to load 2 batteries in series at 90Amps
Stan, I find it amazing that I chose the exact same contactor form amazon, and the same .1 Ohm 200W resisistor from digi-key. I have not set it up yet, but I was concerned about the resistor getting too hot without a heat sink. When you read the fine print its only rated at 50W without a heat sink. Is this a problem in practice?
If you are the David in the April 16 blog, then I got those part numbers from your post. The resistor gets fairly hot, but not “burn down the house” hot… after all, 3.2V * 20A is only 64W. I have a small desk fan blowing across the setup and haven’t had any problems. A small CPU heatsink and fan stuck to the bottom of the resistor would suck away enough heat to make it a non-issue.
Stan, I am not sure I understand why you only see 20A. I would think it would be closer to 32A. 3.2V/.1 Ohm. This would give a dissipation of 102W.
3.27v / 22.7A = 0.144R, so I’m assuming that between the contactor, the 3 alligator clips, and the wire, I’ve gained 0.044 Ohms. My cheap volt-ohm meter isn’t accurate at that low resistance, so I can’t really measure it… so I’m just taking the word of the JLD404 that it is drawing about 22.7A at 3.27v with that resistor.
When I connected the JLD404 to a cell that was being tested with my PowerLab6, both showed 30.0A, so I trust that the JLD404 is at least in the ballpark for measuring amperes.
I have built a bottom ballancer using the JLD404 and a JLD5740 digital Volt meter. It works fine but ended up being a cluge. You can see it in action on YouTube. Please tgo to:
This will also be on display in Jacks garage during EVCCON 1012.
Al, I’m impressed with your bottom balancer and charger setup. Your junk drawer must be much better stocked than mine, what did you use for relays?
That would be an excellent rig for life cycle tests. Since I have the Revolectrix PowerLab6, my needs were much more modest.
I bought one of those Cellpro battery chargers and have fried it twice. Can I get a schematic of the balancer you made? I have the meters and the Power supply… I also have a lot of automotive bulbs and believe it or not I have a hot tub board I removed to replace my old one.
After a cell is bottom balanced, do you recharge it to 3.1 for storage until each cell in the whole pack is bottom balanced and each one of them is charged to 3.1Volts, then charge them all up at he same time? Or can they be stored at 2.7V until the whole pack is done, Then charged to 3.5Volts all at the same time.
It remains my opinion that main reason for choosing LiFePo4 cells over any other Lithium cell is the safety factor. That said, has anyone performed the same physically destructive tests of the CA series cells to see if they as (relatively) safe as the SE cells? Nails, bullets, et al. I’d hate to see these cells sell like hotcakes only to find out they go up like roman candles on impact.
The 3 relays have 12 volt coils and 30 amp contacts. They came off of a dead controll board for a hot tub. These were the motor controll relays and the motors were rated at 1/2 to 3/4 Hp.
I never thought about Life cycle testing as a use for my bottom ballancer
Here is a video all of you might like.
L.A. to Vegas in One Charge? Tesla Model S Road Trip!
Thanks Jack for all your great work and videos.
Thanks for the Tesla video. Really nice car!
The new speedster project sound exciting. Have you considered using the upcoming 144V/500A system that HPEV is working on (due out in Nov, hopefully)? Paired with their new AC-51 motor, it should give better torque across the power band. The folks at HPEV say the controller can handle up to 170V, so you could run a 160V pack. At 500A, it comes out to 72kW, about the same as the -7601 controller/motor. Downside, your battery pack won’t fit into the door channels unless you want to drop down to 40Ah batteries, which is a pretty big hit to the pack power. If you wanted to stash another 10 70Ah cells somewhere, you could get a 11270Wh pack for an additional 50 lbs.
And if you’re patient, HPEV is also working on an 11″ motor (AC-75, I think), that takes the same 144V controller. That would make the speedster really scream. No clue when that would be available.
Spoke with Bill Ritchie on Friday. The 144 is officially put out to December now, and he did not seem at all hopeful that that date would be any more likely than the 10 previous ones they had heard from Curtis. The AC75 and 76 would be far too heavy for Speedster Lite at about 180 lbs. Much of what you don’t like about the AC50 is inherent in my liking it. It is light, and small, and in ratio to its power and rpm a very nicely and even somewhat elegantly proportioned design.
Far more to my liking is their concept, about to be announced, of productizing their siamese AC-35. Though somewhat less powerful, the AC35 is actually quite torquie, and when you put two of them together, and hook that up to two Curtis controllers, the results could be very impressive.
I might add that while it is very preliminary and very anecdotal at this point, the first week of actually driving the Elescalade is a bit of an eye opener. First, it takes a LOT of power to move that much mass. Second, once it is moving you can actually coast a very very long way on the momentum. But thirdly, while the current levels total are somewhat high, divided by two motors not so much so. And the BlowMe superchargers are apparently doing their job with temperature. I actually have the Cadillac ECT probe in the front motor, and I keep having to pull it out and put the cigarette lighter to it to convince myself that is working normally and working the temperature gage on the dash. These motors are running VERY much cooler than I would have expected.
I suspect that although we are using a lot of electrical energy to get the beast in motion, and it does in fact have a heroic frontal area requiring 350 to 500 amps at freeway speeds, it is all spread along nearly 35 inches of shaft and two commutators. Because we just are not straining these motors. (Fingers crossed).
So sometimes more IS more. And sometimes it is not. And in exploring higher powers and higher weights, what is obvious may not be true, and what is true may not be obvious. I am much more excited about the siamese 35 than a 180lb AC75. And I’ve actually kind of lost interest in the 170v controller. I’m told that Brian Seymour has one in his Scion XB and that it really IS the tits. But I really do like the AC50 and 7601. I haven’t needed more torque at the high end actually, and found the power curve almost perfectly suited to Speedster Duh as it was. If I can shave 450 lbs over the package, and up the current to 650 amps from the 550 we had in Speedster Duh, I have to believe Speedster Lite will be a very responsive vehicle and at a price that frankly we just have not explored at EVTV to date. This car should come in at retail prices for everything at somewhere between $35,000 and $39,000. That’s less than half the cost of the eCobra for example and just over half the weight. So if you buy your electric cars by the pound, a very interesting project.
It’s also quite a contrast from a 7500 lb Cadillac to a 1600 lb Speedster. All things being a matter of taste of course.
A great show once again
I bet with the work you’ve done on batteries, you are now the foremost expert on Lithium Ion Batteries for EV applications
I think you shoud get an intern to follow you around all day, take some notes from you and collate, collect and publish formally your findings
You can trouble the battery manufacturers for some Lab Samples, QA batch samples and Random production run samples
This can be a side endeavor on your part aside from producing the show, you can complete the circle and life cycle of battery chemistry / construction design and improvement
You are very much correct in being able to hit that sweet spot in being able to string 38 or 40 70AH Lithium Ion Batteries.
Imagine going back with all these Sag and Swell data back to the Manufacturers of PWM controllers and they can custom design compensation circuits to squeeze every drop of performance from your pack (CA series) for your new speedster platform
Then you will have the beginnings of a prototype car everyone can enjoy with maximum performance at 1600 lbs
But imagine that still using the new stamped classic Mustang body
Wow that would be a show stopper and a head turner
Aside from controller and battery chemistry improvements, I think with your wealth of field testing information, you can be a consultant in improving DC motor designs and automatic transmission designs.
Please give it some thought in doing consultancy work for these manufacturers so that the EV community can benefit further and further as you help propel the technology.
I have some ideas of my own with regards to coupling 3 warp 9 motors to a custom designed automatic transmission and I bet you may have some better ideas of your own too.
All the best to your show and to EVCCON 2012
Thank you and I’m very pleased you enjoy the show. It is interesting that this sparks the idea of 3 Warp 9’s in a custom designed automatic transmission.
And that’s rather the point. Yes, I suppose I could be a consultant to the battery industry, or the controller industry. Or the transmission industry. I also make pretty good whiskey, and pretty good iced tea.
The point is, that our economy has become somewhat formulaic in hiring consultants, and developing intellectual property. And the formula has to do with secrecy. You purchase expertise and develop IP and treat it as a trade secret.
The problem with that is that we all stand on the shoulders of giants. And that lets us see further. If you develop one thing, it can spark someone else to develop something else, perhaps entirely different. Perhaps three 9 inch motors and a transmission. Or a different way of packaging pouch cells.
Or something else. And that something else can spark FURTHER developments that in turn trigger ideas in others.
I have no problem with profits. And I think a free market is the most nutritious medium for this bubbling petri dish of innovation and development. But at the same time, the meager and miserly view that all information we develop we should directly profit from seems to me small and miserly.
We live in a universe of plenty actually. And the more you give away the more it tries to ruin you with plenty – heaped on and flowing over. I take no thought as to what I might wear or what I might eat or what I might drink. I have worn until worn out, and eaten until I can hardly get into one of these cars, and drank until I had to get up to pee five times a night.
The most effective thing I can do to further this cause, I believe, is to publish a video each week of what we have done and what we have learned, in the hopes that you can not only learn it as well, but specifically that it will trigger further innovation and further development and more ideas and in turn those will have the same effect on others.
Some will take that and hide it and try to profit from it in the same old fashion. And some will pay it forward many times over. I am for the latter group. And never did have any concern for the former. They are left to care for themselves and that’s what they are very good at being concerned with.
So the battery companies are more than welcome to what we have found and I hope they make a brazilion dollars from it. And likewise anyone else.
I hope to expand this and report what YOU have done so this in turn triggers others to do likewise. And to large degree this then defines our mission. The idea of me being a paid consultant to industry that exists to make things proprietary and closed is intriguing, but only in its absurdity.
My very presence in their chambers would cause unnecessary and grave discomfort to them, which is not really what I am about. They are welcome to watch the show, just as you do. And I won’t even reveal that they do so. It might surprise you indeed to learn who does in fact watch EVTV.
“…the meager and miserly view that all information we develop we should directly profit from seems to me small and miserly…”
Yeah, yes, amen.
Innovation has a social element and the ability to participate brings responsibilities as well as earning potential. How many overcharge fires would have been prevented if Nissan, Ford, GM et all had put their battery research on line? And, by-the-by, how much better would all their products have been? Sitting on potentially safety critical information for commercial gain seems to me to be both culpable and short sighted.
A sure-fire way to get rid of someone annoying is to hire them as a consultant with an NDA. You’ll never hear from him again.
Here in Singapore, a guy like you would very much be a perfect fit for Exploit Technologies Pte Ltd at the Agency for Science, Technology and Research where I have a lot of friends
Your better ideas would turn into sought after catalyst to propel EV well on its way to become the standard and dislodge Internal Combustion Engine Cars
ETPL is into developing Intellectual Property and turning them into products and putting up start up companies, so Industry field proven field tested information coming from you will be sought after.
Please write a treatise or a paper and publish your work for all mankind
If you don’t have the time, have your interns do the documentations
The guys in the suits at GM, Nissan, Toyota are driven with a different agenda than you are
You are all about getting into the heart of things, thinking out of the box and innovating.
Although I would love to tidy up your wiring (ha ha ha ha)
Case in point is the Caddilac Escalade, how many out there said it’s impossible but you went ahead and succeeded, your idea of having a low stall torque converter from a Diesel Engine is an insight to how the DC motors would actually be maximized in terms of torque.
Given the right iterations and enough time with possibly a data logger you can fine tune your transmission to perfection
All that’s missing in your shop as I see it is your very own Dyno, a generalized test bench / jig fixture / dollies to mount universally different motors and transmissons
The way I see it, material science is key in getting performance breakthroughs, but along the way a paradigm shift is what is needed to re-think strategies
All along you have been saying, take out the ICE motor from those hybrids and just make those things into 100% electrics
Why not do things how things were accomplished in the past
Get enough publicity and following, tie up with a major newspaper and set forth in a race / endurance rally across the US and do some documentary filmining and then getting gov’t to sponsor a prize like the “X” prize
It would bring the best out of every EV individual true to the american spirit of “can do”
The time is ripe, think of the movie “Those Magnificent Men in their Flying Machines”
We are all open source here.
We’d be brothers where it comes to Jacks wiring.
In propped fashion, the work is coming ahead of the video posts or blog over at our site. So I’m going to share our initial observations about the 50 CA100FI cells we got from Jack two weeks ago.
First: always have good, clean connections. My dad was literally burned by a terminal bolt during a 35a test. Leaned over while organizing the next batch and got singed. The graph was a big red flag but we let it go for over :20 before the incident. Afterwords, duh was the primary sentiment.
Next: these cells all came in at .3 mOhms on the cellpro bl6.
I’m contrast to my headway pack, these all had a very regular and consistent discharge curve. These cells have conquered the inconsistency problems Dr. Denis Dorfler was inspired by to develop his BMS.
These cells are all labelled by measured capacity. All are between 110 and 116 amp hours on 100ah specd cells. About 2 out of three are 112.
Most had 59-62 ah in them down to 2.8v. But a handful were less and randomly so. I would advise actively bottom balancing before powering the car. There is no way to spot the less charged cells without individually counting them down.
Our batch was fully sequential minus two cells that were skipped over and are probably in one of your garages. On top they were all labelled with either of two letters. The cells that were less charged were from one letter set, the cells of the other letter set were all within 5 ah of each other. The letter designations were sequential and all of the lower number cells were the former letter. I suspect some sort of technician stamp and that the first tech was getting distracted. The latter set were so spot on that it was beyond boring.
The pl6 cannot handle 1000w from these cells. We were limited to about 520w out of a 6s string.
See you at Evccon with the Porsche!!
That’s careful observation and good info Nabil – particularly the reminder about terminal resistance. I’ve got 100 CAs en route from China so it is very relevant to me
I got my GBS 100AH batteries about a year ago and when they came in I tested them for capacity. I charged them back up to approximately half charge by putting them on a Vicor 5V 40A module that had been trimmed down to 3.6V. I manually pulled them off when they got to around 3.32 volts. These have been resting at this near half charge point for 6 to 9 months. So now it is time to put them in the car and they need to be bottom balanced. It occured to me that there is no reason to do a careful bottom balance by hand. The goal is to have the cells all dump at the same time and to accomplish this all that needs to be done is pick the voltage you want them to get to at the same time. In my case I chose 2.5v because that is the factory listed minimum voltage. I used my iCharger 3010b and discharged each at 25 amps down to 2.5V. Because of voltage drop in the contacts and the fact that the iCharger doesn’t use the balancing wires to sense voltage during discharge of a single cell the actual cutoff voltage is 2.7v. With this approach there is essentially no pain doing the bottom balance and all cells should reach 2.7v at the same time assuming the current is 25 amps or higher.
I took the opportunity to record the initial resting voltage and then the AH taken out of each cell during this test and have corrected the data plotted for percentage of charge rather than AH removed which gave a more consistent plot. It is a rather interesting curve. I tried to record the temperature of each cell but I don’t believe the temp data is good enough to apply as compensation once I come up with a formula. Anyway this is a link to my google docs spreadsheet of the results with graph.
You should be able to use the resting cell voltage to approximate the state of charge within a couple of percent. At least it seems to work with the GBS batteries. It would be interesting to know how close these numbers come out on other manufacturers cells. As Jack has pointed out you do need to be able to read the voltage to 0.0001 when in the middle because the curve is so flat in this range.
Hope to see you all at EVCCon with my Mazda in less than 3 weeks.
Doug, I agree with your statement about not needing to meticulously rebalance them only if you can be certain that each cell received the exact same amount of current for exactly the same amount of time. Measuring that by mid-curve voltage alone is naive. If you cannot account for the amp hours put into the cells independent if voltage, you need to rebalance them.
State of charge is not tied to voltage so much as it is tied to where those lithium ions are – are they in the cathode or annode? If a 100ah cell had 100 total ions in play, you need to know how many of those ions are on the cathode and annode, not what their voltage is when you decide to stop moving them from one side to the other.
Let’s anthropomorphize a bit; the lithium ions are lazy teenagers. The cathode is home. With external persuasion, the leave and hang out at the carbon couch at the neighbors. Now, eventually they get tired and have a desire to go home. However, the correct circumstances need to be present for them to do so – a closed circuit. Once there is a circuit, the scoot home according to their intent. But no matter how intentful they are, if they aren’t allowed, they don’t. Now, let’s go back to that couch. It holds lots of ions, but only so many. Once it starts getting full, the ions’ intent to go back home increases. But again, their intent does not directly tell you how big the couch is or how many ions are sharing that couch. You need to count them each as they arrive and leave.
The thing about they’ve cells is that only at either end of the charge can you clearly infer just how many of those ions are chillin on the couch from measuring that intent, or voltage. So I say, you need to get to that bottom end of charge to know just how full that couch is.
I don’t see anything in my post that should have given you the impression that mid cell voltage could be used to predict bottom balance. Then I know what I was trying to say and probably left out something important. I agree you can’t use the cells out of the box even if the voltages are the same to a millionth of a volt. I did what you did on a few of my cells and measured the AH out of the box and found a 4AH difference. That is enough to ruin cells at either end. My bottom balance approach is to simply discharge all cells independently of each other at the same current to the same voltage. I don’t care what the cells float to after that. I will know in a few more days if this is adequate or if further fiddling is necessary. All we really care about is that the cells all dump at the same time. If I could do the discharge at 100 amps I would because this is more like what will happen when someone is driving the pack to death. It certainly won’t hurt anything to use a low current but I don’t think it is necessary.
The generation of my chart indicates that you can tell the state of charge from the voltage if you let the lazy teenagers rest long enough. It is affected by temperature and I am not compensating for that. I don’t think my temperature measurements were accurate enough or meticulous enough for that.
If you read the voltage of the cell after allowing it to rest for several days you can tell the SOC. To get anything meaningful you also need to compensate for the actual capacity of the cell and the temperature. You might also need to compensate for atmospheric pressure as well. That chart was generated knowing the initial resting voltage and then measuring the amount of amp hours removed and then compensating for the capacity of that cell. And over the 48 cells checked so far (6 to go) it is pretty consistent. The plot is a lot more messy if you remove the capacity of the cell compensation (size of the couch). It would probably look faked if I had good temperatures for the cells and applied the offset to the resting voltage.
I plan to do some testing with the 26650 A123 cells I have and see if the voltages correlate there as well. If they do I will fill in the sparse areas of the chart by testing, not extrapolation.
Doug, I trust we are in the same page.
My albeit limited experience with headways is that they do not have a consistent discharge voltage curve. I could see two distinct trends; one with a more plateaued yet smoothly curvy knee and one that held the plateau till about 40% left, then turned and kept a very linear trajectory down, then had a less distinct end knee than the other profile. Have you seen such behavior? The CA100FI’s held a very consistent and textbook curve.
I also have a PL6 , and I can discharge 6s string at the full 40A of the PL6, but I´m using a very big 24V 400Ah lead battery bank , If you are using a 12V battery the PL6 also as a limit of 40A out to the lead battery., on the software you are able to see if the power suplly / lead battery is in any way limiting your power.
With a 6s string the power out of the PL6 is about 780W to the 24V lead battery
Jack, Thanks for the bench test results of the new CALBs. Looking at the “battery sheet” on lithitumstorage.com for the new CALB FI cells they have a max rated continuous discharge of 2C. I think the SE series was 3 or 4C. I’ll be re-building my EV with a 320 V pack possibly with 100 of the CA60FIAHA. I’ll need a bit more than 2C at times for climbing hills in California. Should I disregard their spec sheet as being very conservative and assume they can handle 3C+?
Certainly. We just demonstrated 12C for 30 seconds and I could see these cells easily doing 3-4C continously. The specification sheets are indeed entertaining, and we read them carefully and cover to cover.
This is not a complaint in any way, just an observation on working with the CA100 cells I have gotten. The packing slip I received with the cells states their battery # and capacity in ahs. In testing with my powerlab the ah figures while might be an slight indication of ahs,the figures might be taken with a grain of salt. I have seen cells marked as 112ah show more ahs than those marked as 114ah. Just passing the information along, all measurements are taken charging the cells to 3.5 volts and then discharging to 2.9 volts.
For capacity measurement, the metrics they use are 3.65v and 2.50v. If you use different ones, you will get different results methinks.
With all due respect I think your lack of “car guy” has shown up in your El Escalade. Try an avoidance maneuver at 40 MPH (not on a public road) and I think you will find that your vehicle is dangerous. As us “car guys” know, loading up your pickup with too much weight at Home Depot makes for a very scary braking and swerving dynamic. It is not a matter of shocks, airbags or anti swaybars. It is simply too high a center of gravity. I have heard you state you like “elegant design”(yes, I have watched EVTV since the first episode). Loading up that vehicle with a gazillion pounds of batteries above the beltline is not a good idea. Yes, it may go down the road smoothly, but a sudden over-correction will have you looking at your rear bumper in nanosecond.
You appear to have a lot of knowledge of a vehicle you’ve never driven Neil. But your comments are pretty much on target. The vehicle is overweight. But yes, I’ve tried some panic stops and the brakes work pretty well and some sudden lane changes and it does wallow a bit but no end for end swapping. It is certainly heavy.
But to my way of thinking it was when I started. MOre so now. There is no doubt 1700 pounds of batteries is an issue. But they are pretty evenly distributed obviously and rather low in the bed. I guess the spec would be for 1500 lbs so we are over. But it does not seem to feel alarming in any way.
So yes and no. It is portly and it does wallow. BUt it does not exhibit the characteristics you so confidently express.
Jack, I just watched your sept 7 episode, nice job, Did I miss something? Where is Brian? is Brian busy with EVCCON?
Regarding lithium and its nuclear activity and its state as an ion or not an ion …are we dealing with a plasma here? I found the following EXTREMELY interesting and thought you might like to review. I don’t want to drag anyone too far off topic but I think this is pertinent: ( I know you like Plasma Balls and such)
Plasma Physics’ Answers to the New Cosmological Questions by Dr. Donald E. Scott – Full Video
I am using 90 amp-hr ThunderSky batteries in a Miata conversion. 40 cells came from a Jeep conversion two years ago and 8 from Jacks garage. Before installing i made sure all were at 3,.22 or 3.23 volts. We also added a mini-BMS that shunts current above 3,7 volts and alarms above 3.8 or below 2.5. Some cells seem to have lost capacity as i can get the low voltage alarm with just 5 miles of hard driving at 65. Or I can baby them for about 40 miles. Due to a comedy of errors we were without a charger for 5 weeks and the BMS seemed to slowly bleed down weaker cells. I then had to play a game of whack a mole with a USB charger to bring those low cells back to voltage. Hopefully i can top balance using the BMS and get them all to the top voltage. I wonder if a year of hard use in the Jeep has changed the capacty of the cells as it is challenging to get them balanced.
Are you sure your BMS isn’t just throwing up an alarm at 2.5 V because of normal voltage sag under load? You could try removing all BMS devices, bottom balance all cells by bleeding them to 2.7 until they don’t bounce back anymore, recharge to 3.6 per cell while counting amp hours and see how far you’ll get before those amp hours run out. You can put back the BMS later if you want to, but you might find that your vehicle works better without.
Note! Do not top balance the cells at 3.6. Just charge the pack as a whole.
Normally you shouldn’t notice any change in capacity (unless using accurate measurements) for several years at least.
You are (and by extension your equipment) doing something wrong.
I suspect your BMS is unbalancing your cells and no actual capacity is really lost. Remove BMS. Bottom balance. Charge to 3.5 or 3.6v volts using CC CV.
Drive and count AH.
This does sound like a comedy of errors. PARTICULARLY with a mixed pack like this, an initial bottom balance is imperative to get them on the same page. Driving them any distance at all without doing this will almost certainly lead to death of some cells. Take them all down to 2.75 volts as precisely as you get them, then charge the string together to 2.65v. I don’t know how many times I have to iterate do NOT top balance. You can use the BMS to monitor if you must. A year of hard use in the Jeep probably didn’t hurt. But putting a mixed pack together without bottom balancing probably did – and you most likely do have some bad cells at this point. They should become obvious when you try to bottom balance. If they don’t, you’re left with the possibility of a cell that sags to much under load. Basically, you have to set the brake and block it. And set a meter on a cell then and apply accelerator with the brakes locked to reach a specific current draw – say 100 amps and note the cell voltage. Move the meter to the next cell and repeat. Do that for the pack and if you have any weak sisters, the voltage will be much lower when you hit the 100 A load level for that cell compared to the rest.
If you get them all bottom balanced and they are all within 0.10 or even 0.15 on the load test, you are probably ok. Charge normally to 3.65v x N cells.
In driving the car, having the cells hit 2.50v under high loads is not an indication of a problem, something I did convey to the builder of that BMS but he obstinately refuses to listen. Thunderskies DO sag a lot under load. If you are working with a single string of those 90Ah cells, then indeed they will sag at even 400 or 500 amps quite remarkably. If you have two strings paralleled, not so badly.
Yes, it’s a lot of work. But when you take on the task of putting cells from different sources together to make a pack, that’s what you’re talking about from the beginning. THere aren’t really any easy ways around it and no magic BMS is going to make it easy or allow you to “get lucky.”
Thank you all for the comments. The low voltage alarm is coming on during hard acceleration of 500 amps. While there is a 4 second delay one can easily hit it on an on ramp. Sounds lile i have more work ahead of me.
Small remark to Cougar video: the number of impulses per revolution should be divided by two (you had four stroke engine), so for 4-cylinder you should assume 2 impulses per revolution not 4 and for 6-cylinder it should be 3.
I just recently found your website and videos, as I am in the fantasy/research phase of my own conversion. I came across one battery manufacturer’s (flux power) claim that their “electrolyte starved” LiFePO4’s could be oriented in any direction. I thought back to all the conversions I’ve seen (many!) and realize I’ve never seen cells installed in a car in anything other than upright (electrodes on top). However, I imagine it might be smart to keep the cells on the floor, between the wheels, and out of crumple zones laying flat. I would love to try and replicate the Tesla Model S battery pack design in my own conversion as much as possible. Obvious advantage of a low center of gravity, etc. Do the CALBs and all other prismatic LiFePO4s have that same feature or is it bad for them?
Thank you for sharing your work!