Hi all. I'm new to this forum (see my first post
here...) but I have quite a bit of LiPo/a123 experience in the electric RC world, mostly with big electric helicopters. I am a frequent contributor at RC Groups (over 5,000 posts...), and have probably as much experience with LiPos as anyone over there. I helped Charlie Wang get Thunder Power going several years ago by starting a website (
http://www.tppacks.com) to mainly sell his Thunder Power "gen 1" packs. Back then, we didn't have to worry about balancing. Charlie did a great job of painstakingly matching cells in voltage and capacity. The cells stayed perfectly balanced, for literally hundreds of cycles.
The only problem was that the cells were only good for about 5-6C discharge rates, so most packs were built using four 2000 mAh cells in parallel, and either 2, 3 or 4 cells in series. Being an early proponent of using higher voltage setups in helicopters, mainly as a way to get higher torque, I finally convince Charlie to make some 5s3p packs.
With the 8-10C "gen 2" cells, everything changed. Charlie still did the cell-matching trick, but these cells (from a different Korean manufacturer...) just did not stay balanced. People ended up killing a lot of these packs, mainly due to over discharging. Eventually, almost every pack had problems, and at best, the pack lifes were dramatically shorter. To his credit, Charlie replaced a good number of these packs, no questions asked. It cost him over a million dollars that year. Other pack manufacturers, like Brian at Tanic Packs, started putting "taps" on packs that would provide access to each individual cell junction. With the 15C "gen 3" Thunder Power packs, a new balancing connector was added. Thunder Power also introduced the TP-1010C charger, and matching 10-cell TP-210V balancer. While other charging/balancing solutions appeared as well, The 1010C/210V combo was unique in that a serial connection between to two units allowed the charger access to each cell block in order to indivdually monitor the voltage. That way if one block had a problem, the charging process could be stopped, before a cell exploded, causing a chain reaction. This was significan, as 90% of the problems with LiPo fires/explosions are a result of overcharging cells.
Today, there are lots of very good cells on the market that have higher capacities, and very high "C" ratings (20-30C continuous, 30-40C "burst"...), but they are all still dangerous as all get out. When the a123-based DeWalt packs came out, many of us fully embraced the use of these cells, as they seemed to be the answer we were all looking for, high power, and high safety factor. In actuality, it took many of us awhile to understand a few key characteristics/differences in working with these, as opposed to LiPos, but once we did, they have worked quite well. Other than the higher weight issue, the main "problem" we had was that although these cells have very high "C" ratings (30C continuous and 50C+ burst...), the voltage drop under load is a lot greater than what you would see with a higher end LiPo cell. This is a big problem in helicopters as a voltage drop translates into a drop in speed in the rotor, which is not a good thing.
What we also found, however is that over the duration of the pack, that voltage drop stays pretty constant. With LiPos, even the best LiPos, the voltage drop gets worse towards the end of the capacity. The a123 cells, however, deliver pretty consistant power under load, throughout the duration of the pack. By simply regearing a bit, to plan for the higher voltage drop, we were able to get very consistant power, all the way to the end.
A second characteristic that has proven quite useful is that in a helicopter it is virtually impossible to kill the cells. With LiPos, we have all gotten used to only discharging a pack down to where it stillhas about 15-20% capacity left, in order to increase the cell/pack life. In a helicopter, with the a123s, we can fly the packs until they completey dump, which happens quite quickly with a well-balanced pack, because the "resting" voltage of the cells at the point that a helicopter won't hover anymore, is still well above the point of damageing the cells. The packs/cells seem to be able to handle this "abuse" quite well, for hundreds of cycles, so we just "fly them like they were stolen", to quote one of our more colorful users.
If you have a rainy weekend, with nothing to do, check out this thread:
http://www.rcgroups.com/forums/showthread.php?t=566256, which chronicles our experience with a123-based solutions in helicopters over the last year. It is up to about 100 pages now, I think.
In any case, I have become totally immersed in using a123 cells in my quest to convert as many "glow"-powered users over from the darkside.
For years I have had several different conversion kits for converting the most popular gas-powered model, the 8-pound Raptor (there are about 100,000 of these out there...), into a higher-powered electric equivalent. Although moderately successful, I constantly ran into two concerns from potential converts that were "showstoppers". One is the cost of LiPos (most said it was like carrying around a year's worth of fuel on every flight...), and the second was the danger factor. Once we got the right setups figured out, the a123-based solution has allayed these concerns for many. Now, the Taiwanese Raptor manufacturer has finally introdeced an electric version of the Raptor, but unfortunately, it is not very "a123-friendly", so I sell a lower frame upgrade that makes it easy to use packs, pretty much just like they come out of the DeWalt packs.
As many of you here are aware, the cheapest way to get a123 cells is to buy the DeWalt packs off eBay, from dealers who break apart multi-tool combo setups and sell the individual pieces. It is pretty easy to get these for about $100, shipped, or about $10 a cell. I tried very hard to get a123Systems to sell me cells directly, at a reasonable price, but the best I could do was about $13 a cell, and that was in quantities of about 10,000. Last year they sold Black and Decker (who ownd DeWalt...) 10
million cells for the tool packs. Even though 10,000 sounds like a lot, it really doesn't do much to get their attention. Now, of course, GM is dumping tons of money into a123Systems, for the Volt, so they don't even respond anymore.
When these first came out, charging and/or balancing these were a bit of a problem, as the optimum cutoff voltage for a123 cells is 3.65V, not the 4.2V that LiPos use. What we found, however is that it didn't seem to charge them at the higher voltage. They don't blow up, like a LiPo most certainly will if overcharged, but nonetheless, we usually just picked the closest LiPo setting the chargers had, and balanced the packs separately. Now, virtually all the popular charger/balancer combos have a123-specific modes.In the last few months, I've noticed that if packs are made from fairly matched cells (i.e. -- from the same DeWalt pack, or from packs with similar "resting" voltages...), they will do a fair bit of self-balancing. The key is to fully charge the pack. Even with packs that have different resting voltages, right after a charge, once under even the slightest load for a few seconds, the individual cell voltages will end up pretty close to each other, usually somewhere between 3.48 and 3.53V per cell. The point is that now I don't really worry too much about cell balancing, and have stopped adding balancer plugs to my a123 packs.
What I have found in using a123-based setups EVs is that it is quite possible to kill cells by over discharging them, so some sort of BMS is really needed, if nothing more than a low voltage cutout (LVC) circuit. I know the Crystalyte controllers have this feature, but I think it might be set too low (29V for the 48V controller?). I need to read up about the controller hacks you all are doing to these, to see if this can be adjusted. I'm still trying to figure out how low you can safely take an a123-M1 cell before it a) dies completely, or b) at least has its capacity reduced, but I think a resting voltage under about 2.5V per cell is about right. Under load, I don't know yet. I inadvertantly left one half of a 16s4p pack not connected and then went on a ride that was longer than the other half of the pack could handle. About half the cells were okay, and came back to life from resting voltages around 2.6V, but many had voltages around 1.6-1.8V, and didn't really recover. A few were dead-shorts, or completely dead. The ones that were above 2.6V seem to be fine, and seem to still have capacities around 2200-2300, I've noticed their "resting" voltages right after a charge are a bit lower than the normal 3.5-3.6V, down around 3.4V. I'm keeping my ey on these, but so far, they still seem to working fine.
I've never really understood just exactly what the BMS does in the DeWalt packs. It has connections to each cell junction, but I don't know if that is just to monitor the cell "health", or if actual cell balancing goes on when ithe pack is connected to one of the DeWalt chargers. I do know that the BMS does have a current limiting function, which is somewhere around 15-16A, I think. This precludes the use of this BMS in our helicopter applications as we can easily hit short 80-100A peaks (4000W+...) in some of the more powerful setups. The packs have two negative outputs, one that is switched (enabled by connecting the positive terminal to the closest pin next to it, but apparently it does not have a LVC function, at least not if the unswitched negative connection is used. I was able to kill several cells in a pack in the initial testing I did with using these packs in series (i.e -- 66V...). This happened because two of the four packs I was using came disconnected from the harness I used when the bike hit a prety good pothole. After that, I had my friend Andrey (
http://www.askmanap.com make some mounts for me that solidly hold the packs in place. Each mount is designed to slide into a Topeck MTX channel. I'm using two of these mounts now to hold 8 of these packs :
The nice thing about this setup is that using four of the fairly inexpensive DeWalt chargers, I can be fully recharged, and ready-to-go again, in under two hours. With my other a123-based setups, I use a $400 Zivan NG1 48V charger for charging all the 16s packs, and a 6V 2A Soneil charger for my 2s "booster" packs I use to get to 18s for my Mariner folding bike setup. The NG1 will charge at a max of about 18A. I usually just charge all the 16s sub-packs, from each setup, together, but if I have time, I'll individually recharge the sub-packs. For the two 16s2p packs on the Port Runner, I can charge each 16s block individually, if needed. Each cell in these 16s "chunks" were pretty closely matched, initially, so the occasional individual 16s1p charging helps keep these pretty well balanced.
In the 18s setup I have in the bag, on the Mariner, there are two 6s6p blocks and a 4s6p block. Each of these have special connectors that can be connected to a BalancePro HD charger
http://www.fmadirect.com/detail.htm?item=2044§ion=38 that can individually charge/balance up to 6 cells at up to a 10A rate. Occasionally, I will use three of these chargersto recharge/balance the big 6p subpacks. To get to 18s, I then use two 2s3p subpacks in parallel, for 2s6p. I have two of the 6V 2A Soneil units, but if I'm in a hurry, I'll use a couple of my RC chargers, which all work at 10A.
Sorry for the long, rambling post.
I'll try and keep them shorter, in the future.
-- Gary
