Thursday, October 8, 2009

Sky Energy's Blue Sky Cell Data



We're nearing completion of the battery box installation on the 2009 Mini Cooper Clubman. This pretty much had us down in the floor wrestling in the mire. Not sure just how much video will be useful. I have posted a new video with the gas tank boxes, but the other two - I can see where we could have shot more, and I can see where a lot of what we did shoot isn't going to survive the editing process. There is just so much of drilling and riveting anyone can be expected to watch on purpose.

As noted earlier, I had decided to wire all the cells to amphenol plugs and build a test set to allow me to balance cells manually. I don't know why I've caved on this. SO MANY people have warned me so emphatically that without a BMS I am TRULY among the lost of the lost and horrible things will happen to the entire Midwest if I don't get with the program. So I've spent hours and hours wiring these things up and cell balance, at least initially, even though we have 112 of them this time, just doesn't seem to be a problem.

After the 107 mile run in the Speedster, I was certain I would need to trim them up with a little charging here and discharging there to bring em in line. Despite the fact that I had run the Thundersky's down to 2.8-2.9 vdc - over the knee of the discharge curve, all the cells were REMARKABLY close. THIS BALANCING ISSUE IS BOGUS. I've put 5000 km on this car and it isn't a problem. I've overcharged, overdishcarged, and done everything to these batteries to TRY to kill them and/or induce some imbalance without actually draining a cell or overcharging a cell myself to prove it could be done. I cannot discern real world data with a huge investment in test equipment to support the theory, much less what appears to have become a religion, about LiFePo4 cell balancing issues. I hear you. I have listened for over a year. But when I look at the meter, there's something wrong between what you are telling me, and what I'm seeing. I don't know what I'm missing, but there's a big disconnect here.

So I feel a bit foolish having installed about 30 unneeded pounds of 16 gage wire so I can access cell voltages without disassembling these largish battery packs. And when I cycle my Soviet Submarine wafer switch through the cells, it SURELY does get boring. They're all the same.

But I feel a little better after today. And I feel a little drained. I basically spent from noon to seven o'clock this evening standing there with a stopwatch and an Agilent 5 1/2 digit bench multimeter. I love the West Mountain Radio load system. It does nothing it purports to do, but the little bit it does, it does pretty good. It's charts are awful and it can't read a voltage if you sound it out audibly yourself. In fairness, putting a 100 amp load on a 3.3 volt cell, tiny resistances in the cables and connections just throw voltage reading out the window. You have to put a meter that ISN'T loading it, directly on the terminals. I explained this a bit in the post about the test box we built.

In any event, about all it DOES do pretty well is maintain a pretty constant current load on a battery that is changing in voltage. And it does totalize AH in reasonably accurate fashion.

We have most of our 100 Ah Blue Sky's in the car. I had done some capacity testing and they were checking out at least as good as advertised and some quite a bit better. Today we received a shipment of the larger 180 Ah Blue Sky's.

I've spent most of the past year working with THundersky's - same manufacturer really but a bit different olivine structure on the cathode. Manufacturers recommendation is to charge to 4.25 vdc, although they do mention in passing that you can extend the life of the cells by limiting this to 4.10 vdc. But the charge curve was so steep, that at about 4 volts, charging a series string, you would lose control and some of the cells would shoot considerably past 4.25 volts while others lagged. So we learned to charge to a lower voltage, kind of right at the knee of the curve where it turns up. As it turns out, you don't get much additional energy into the cells anyway after about 3.75 vdc. So we've been pretty happy there.

The discharge curve comes right back down the hill to 3.3 or thereabouts and then declines very gradually to about 3.0 vdc. After three volts, it again turns wickedly sharp downhill. Below about 2.8 vdc there is nothing there to look at but cell death. In doing the 107 mile run, I ended between 2.8 and 2.9 across the pack. Another mile would have probably been disastrous.

The Blue Sky's are spec'd, in the terse, no information really style I've come to expect from the Chinese. Charge to 3.6 volts. Discharge to 2.0 volts. That's much LOWER on the high side than the Thundersky's, and ALSO much lower on the low side, 2.0 vs 2.5 vdc. So I've been wondering what all this really means.

Today I found out. Incredibly, the Blue Sky's have an even flatter curve than the THundersky's. Worse, after the knee in the curve on either end, it is even more precipitous.

I learned this by standing there and manually measuring the voltage every five minutes while charging them at a constant 75 amps. I then let the batteries rest for 15 minutes and settle to their static voltage. And then I began discharging them at a constant 99 amp rate. The two accompanying graphs tell the tale.





In the case of the charging, the manufacturer suggests that the batteries be stored any time you are not using them at 50-60% charge. These arrived in about the right range. It took 80 minutes to fully charge them at a very constant 75 amps off a bench power supply. That sounds a whole lot like we put in EXACTLY 100 Ah into these 180 Ah cells. As soon as we put power to it, the voltage went to 3.4600 and hung like a rock for about 35 minutes. I kept rechecking my meter connection to the cell terminals, but that was the reading. At 40 minutes it finally climbed to 3.4675 and from there climbed nicely at first and then sharply to 3.59 volts. At that point, it zoomed in the next five minutes to 3.76 vdc. If you must charge to 3.6, I suggest keeping an eye on it. It takes off very fast. Much steeper than the THundersky's I think.

The corollary is they hang in there longer at the lower voltage, before taking off on this very steep climb. That's good to a point. But we have to know where to shut it off. I'm going to err on the side of caution, and realizing you don't forfeit very much this way in capacity per ride, and might gain a lot in cycle life, and call it over and done with at a more conservative 3.5. We want to get everyone into the pool and off this 3.46 plateau, but that's about it. Not much in there for us after 3.5 volts.

On a 112 cell pack, that looks like a charge voltage of 392.

I included something in this graph that I wish everyone would. In talking about batteries and pack voltages, I'm never quite sure WHICH voltage you are talking about, or asking about. We will charge to 3.5 and when the pack reaches this magic 392 volts, our charger will switch from constant current source, to maintaining the cells at as exactly 392 volts as we can muster. As the cells top off, this will require correspondingly less current. The current will taper off pretty quickly in this case - albeit much more sedately at the 10 amps the Brusa can muster at 392 volts. And when the current tapers off to some small value, an amp or so, you're really quite done charging.

When the charger cuts off, within just a few minutes, the cells leak off the surface charge, basically accumulated homeless electrons that can't quite find a Lithium Ion to mate with, and the voltage drops. I have rather loosely adopted the rather loosely bandied about term "nominal voltage" for this, but what I mean is the resting static voltage after the charge process.

So in this graph, after shutting off the power supply, I simply continued to take voltage readings every five minutes as I had been doing. As you can see, the voltage dropped in the first five minutes to 3.3853 vdc. Over the next five, it kept falling to 3.3724. But by 15 minutes it had pretty much stabilized at 3.3668 vdc. So the "nominal" voltage of a fully charged Blue Sky would seem to be about 3.36 vdc.

On the discharge side, you can see we started at that 3.36 vdc, but on applying this 99 amp load, it immediately dove to 3.156. But that's really NOT much sag on a 99 amp output - a little over 0.5C. Two tenths of a volt, across 112 cells would be a noticeable 22.4 volts however.




Again, the cell seemed painted at 3.156 for 20 minutes. It then began a long slow decline down to 3.0 volts. After 3.0 volts, there was nothing good here for us. From 3.0105 it took 5 minutes to drop to 2.9161 vdc. The NEXT five minutes took it down to 2.540 and the next 5 minutes saw it shoot past 2.0 vdc limit to 1.85 vdc. Ten minutes from 2.9161 to 1.850. That's about 16.5 amp hours, but you're welcome to them. It's less than 10% of the pack energy, and I'll just throw out some upholstery or something to get that extra range. Three volts is about the end of the line for me.

So there we have it. 3.50 vdc for a 392 vdc charge voltage. We'll be looking for 3.36 vdc nominal for about a 376 vdc operating voltage. And at 336 volts, we should be getting REAL close to the EVTV Motor Verks to belly up to the bar at the Texaco Fire Chief bar for a sip of some 240 vac stuff before going home.

And I would guess our real pack storage then is something like our NOMINAL operating voltage times our Ah rating for the cells, in this case 100 Ah which is what we put in the car. And that sounds like 37.6 kWh. Another way of looking at it is how much AC you have to put in through the charger to get there, and that would be more like a nice round 40 kWh.

In any event, it's too much for our charger. While the Brusa in the Speedster will charge it pretty handily at 25 amps per hour, it is a lower voltage - about 130 vdc charge voltage. There is no free lunch and the little Brusa can only handle so much power. If you increase the voltage to 392, that's about 3x the voltage. So to produce the same power, you have to decrease how much current you put out. And that works out to a little over 8 amps. That's about 10 hours to charge an 80% DOD pack.
Not good.

These larger packs take some strategizing. We were going to put TWO Brusas in and since they are isolated chargers, we could then do 16 amps between them. Let one drop off a little early, and the other continue to do the finish charge. Easy money. And now we're back at 5 hours.

But it surely did charge nicely at 75 amps. And almost all our charging is done at home. At 75 amps it could charge in just over AN HOUR.

But my 100 amp bench power supply only goes to about 15 volts. Same problem. Scotty, we need more power.

Well, if you recall, for the TExaco Fire Chief charging station, I ran a 100 amp 240 vac circuit underground to the front of the garage outside. And as it so happens, Manzanita has just released a 75 ampere monster charger. And it is a monster. You could never put it in a car. You would be pressed to put it in a bus. And I'm guessing it's not going to quite make 75 amps at 392 volts. But it might charge this car in a couple of hours anyway.




David Kois of EVComponents used to make some replica gas pumps. But his look like a LOT higher quality than the Sky Chief we bought on eBay. He's threatened to pull the forms out and go back into production. He has a 30s style double clock face PollyGas pump that just might be big enough for a Manzanita PFC-75. If we put a REALLY big Anderson connector to the pack in the car, say under the hood. And mounted a two cable DC rig to the Manzanita, we could hook the Manzanita to the 100 amp 240 vac, and we're in business - WHEN we're home.

A couple of issues. I don't have the pump from David. And I HATE Anderson connectors. They can be very stubborn connecting and disconnecting the larger ones, and I have burned some up. We need a heavy duty DC connector for fast charging. I have seen some quick disconnects on welding cable before. We may look into that.....

By the way, EVComponents carries the Manzanita PFC-75. I think I got serial number 3 so they're pretty new, and they are pretty proud of them - even more than the Brusa. But they put out the power.....

Jack Rickard

11 comments:

  1. Hi Jack,
    very interesting write-up of your empirical results from testing TS & SkyEnergy Cells. Perhaps, what people tried to warn you about is not as much about BMS for balancing (equalisation), but rather about BMS making absolutely sure you don't overdischarge individual cells while they are connected in series, aka low voltage cutoff (LVC). So, it is not important what your aggregate number of charge/discharging voltage situation tells you if some of these cells go much deeper bellow it (and this info gets hidden in the aggregate). I understand your approach, just replace the wrong cells along the way and forget it. While in your particualr case with "giant" kWh battery pack and keeping it safe above 20-30% DoD all the time might actually appear to work, for a moment. But the general practical consensus is that it kills batteries prematurely, especially as shown in smaller packs and with deeper discharges, e.g. the electric scooter business was just one giant mess and graveyard prior to this knowledge. In the electric car segment, this issue especially with large packs takes longer to manifestate itself.

    In terms of Anderson connectors, have you tried this fast disconnect thingy (manual release kit/handle), I rather like it.
    http://www.solarseller.com/anderson_sb__sb_cover__sb_cable_clamps__anderson_manual_release_kit.htm

    All best
    M.

    ReplyDelete
  2. Jack, I watched your video about BMS. You have two parallel strings, but I saw only one EVision. How do you have the EVision measure the center offset of each parallel string--or do you just have it hooked to one of the strings?

    Bill

    ReplyDelete
  3. Hi Jack,

    I noticed that when you made the battery boxes, you did not use cushion (e.g. rubber) under the lithium batteries. Are the batteries tough enough for the road vibrations?. Does the Thundersky manufacturers have any data with regards to vibration?

    Thanks & best regards,

    JohnM

    ReplyDelete
  4. Mesuge:

    I know. And I know I think you know. And I don't think you do. I don't know what to do with that further.

    The destruction of the batteries happens of course when the discharge curve knee is hit and one cell goes down before the others. But it's not magic. They have different impedences out of the box. But we don't want to overcharge them either. So we undercharge them slightly and we don't go over the knee on discharge.

    It doesn't get worse. It appears to get better as you use the batteries. No idea why.

    I detest Anderson connectors. We're probably going to go with these camlock welding cable connections.

    Regards

    Jack

    ReplyDelete
  5. Anonymous Bill.

    Excellent question and VERY observant. The two parallel strings are kind of an anomally. I had originally planned on using 160Ah batteries in a single string. I have to say, I've grown to like the two parallel strings. For one thing, it gives me 180 Ah instead of 160.

    The EVision of course monitors the pack at both ends, and conveniently enough, with a parallel string, both ends are tied together. But the HALF pack voltage occurs at TWO different locations, and they are NOT tied together.

    Brian put me a little toggle switch behind the EVision HMI. I have both centertaps routed to it, and can select either with the switch.

    Didn't really think anyone would notice. You're thinking.

    Jack

    ReplyDelete
  6. Anonymous JohnM:

    The manufacturer does not mention vibration at all - or for that matter orientation. Having sawn one of these in two, I can tell you the case is VERY tough, the foils are quite jammed in there, and there is really nothing to vibrate or suffer from vibration. These are VERY sturdy packages, and one of the reasons I strongly favor their use over the "battery modules" made of a couple thousand little cells that is so much in vogue.

    BTW, we had a little 16 cell A123 pack we had been using for 12v around the shop just go totally thermal on us last week. A LOT of heat. This is the model they are using in so many of these modules. You'll know it when you smell it. Yeah, we were pulling 60 amps or so. But they should be rated for 2.8 Amps x 4 and so this was about 6C tops and very briefly.

    Jack Rickard

    ReplyDelete
  7. Jack,
    Very interesting and inspiring videos. Over the past 10 days I think I watched them all including the first weekly show. I have been stalled in a project for a year, parts all too expensive or just not quite right. I like the various options you show. I have been pulled in by the hands on approach to components and the conversion process. Personally I need more engineering calculations and I think BMS is essential. But this is a hobby and everyone gets to do what they think is best and share their experience with like minded people for a multitude of reasons. Thanks for sharing the videos and I hope there are many more to come.
    David Kerzel
    Pompano Beach
    Gold Coast EAA

    ReplyDelete
  8. Jack, your discharge data for the SE180Ah cells indicates there is 181Ah capacity between 3.36V and 2.54V. There is typically quite a bit of capacity between 3.36 and 3.5V so it seems this cell must have significantly larger capacity than 180Ah. The half-dozen 180Ah cells I have charged individually at 5A only show about 10Ah between 3.15 and 2.95V. Much lower than your data indicate. I don't understand.
    tomofreno

    ReplyDelete
  9. Hmmm, I guess one difference is these are voltages under load, so sagged by around 0.2V, whereas the voltages I recorded are "rested" voltages ie, take off charger, allow to relax, record. Also, it looks like this data shows about 16.5Ah between 2.916 and 1.85V, in fairly good agreement with what I have seen. Still don't understand the 181Ah between 3.36 and 2.54V though.
    tomofreno

    ReplyDelete
  10. I have been thinking about chargers and just started a new thread about Harry Landis' charger controllers. Anyway, from what you are reporting here, if the nominal voltage is 3.366v, the Landis approach would be to turn on the charger for 5 minutes if the voltage fell below 3.366v/cell. When the charger turns off, repeat. It sounds like this might provide a simple way to control the charging of your pack.

    ReplyDelete
  11. She may as well have been. It exploded like a weed, a tough one, imperviousto pesticides too.
    free online xrated sex porn stories
    free hard core porn stories
    father and daughter incest sex stories
    black sex stories post
    stories of women and beastiality
    She may as well have been. It exploded like a weed, a tough one, imperviousto pesticides too.

    ReplyDelete