ElectricGod
10 MW
Okami said:Ok - so I got a quick reply for you, perhaps you will get the chance to reply soon!
A very clever way to measure the cells! Indeed. I think you have changed my understanding now on how to check cells for high load use..
Ok, some conclusions I got when I tried to discharge my Panasonic CGR-CG pack 4s2P pack (~16v, ~2250mah per cell, when new and 2-3c discharge - datasheet data)
I got to about 3.1 -3.2 volts under load per cell (~25w) and the cells became considerably weaker, if they were pulling the mentioned 25w of load before, then now they pulled only 21-22w of power)
When I disconnected the load.. the charge went up to even 13.43v
I forgot to monitor the total voltage when under load but it was around 12v..
Anyways - I concluded that the performance is quite bad when the cell is below 3.3 under load.. (same conclusions can be found on the net, of course)
And the load voltage is about 3.5v when the cell should be stopped discharging probably
The capacity I managed to sqeeze is out was about 2.650 mah, out of 60-70% full pack when I started monitoring the discharge.
---
So - anyways when I connected to the balance charger, the different between highest and lowest cell was about 0.08v I believe, so almost 0.1v (~3.44 and 3.36)
---
@ElectricGod, though, Im writing all of this to ask you - so what exactly happens when you have different capacity parallel groups? This was the thing why I wanted to test each cell, so that I can arrange them in a manner, the capacity levels are somewhat equal..
How will the pack act when one group has 2ah, another one has 1.8ah and the last one, for example, 2.2ah..
--
I actually wanted to make new topic about mixing high discharge cells with low discharge cells (like laptop batteries).. and what would be the effects of doing so, it looks like there are ppl who have even mixed lipo and li-ion, li-ion and lifepo, and even Pb and lifepo.
So yes, another thing im looking for is, what I should be beware of besides monitoring the cell voltage when connecting two different packs in parallel (same voltage of course but perhaps a little weaker capacity). I do believe that the weaker pack will probably drain faster and will have to be disconnected, while the bigger one will probably have more juise and otherwise.. just supplement the weaker pack, when it has reached its low point.
I run mixed. I have the battery box over the back wheel and then the battery bay under my feet. The battery bay has LIPO cells in it and the box is full of used LION laptop cells. I have in the past run both sets of batteries without a BMS on either one of them. IMHO LIPO will drift out of balance more quickly than will LION so it always made me nervous to not have a BMS on my LIPOs, but never the less I ran for a couple of months with no BMS on them. I was just way more hawkish about checking them every other charge. The LION cells went 2 months between checking if they were still in balance or not and they were fine. By then I had purchased a couple of BMS and didn't worry about it anymore. I still run mixed, bu there is a single BMS for the LIPO packs and another for the LION packs. LIPO is cheap and wont last very long. LION is definitely the better technology and will last for years..even laptop cells. I use a LION BMS on my LIPO pack. LIPO charges like LION except the end voltage is 4.2 rather than 4.1 volts. My charger is set to 82 volts or 4.1 volts per cell at 20S. As a result the LIPOs charge to 4.1 volts for a very small loss in capacity, but longer life span.
LIFE charges very differently from LION and LIPO so as long as I had a LIFE BMS on those batteries, then I wouldn't be afraid to run mixed with them as well.
In my configuration I have LIPO and LION. Pretty much any cell type you get...including alkaline and SLA batteries, they get saggy towards the end of their charge. I expect that as I am nearing the bottom of my charge that acceleration is going to be weaker than normal and that if I do crank the throttle that I am going to sag sufficently to cause the BMS cut-off to happen or the protection voltage level of my controller to stop me..and they do. I also have noticed many times three behaviours. The first is that close to low battery charge that under load, the batteries sag quite a bit. Lets say I am running at 65 volts or 3.25 volts per cell at 20S. I have to hit the throttle slowly and gradually or I will drop below 60 volts very quickly and then bounce back up to 65 volts again. That's a decent amount of sagging! The second thing is I use regen...who wouldn't...so just coasting if battery voltage under load was 65 volts, now it jumps up to like 72-75 volts. Then coming to stop...or no load and no regen, battery voltage settle to like 68 volts. I never look at my resting voltage as true battery voltage. In fact I call this the "recovery voltage". Any partly discharged battery if you let it sit for a little while will slowly increase in voltage as electrons slowly migrate towards the negative pole. It's not magic happening, the battery has exactly the same charge it had before, just the electrons are a little bit more available for a second or 2 and then the battery returns to whatever it's charge state was before. The only battery voltage I care about is the one I read when I am loading the batteries. All the rest are momentary numbers.
All my LION packs are in parallel at a cell level so they can run on a single BMS. I have 6 20S2P battery holders, each with four 5S balance cables. The balance cables are all connected in parallel. So starting from the negative most end of each pack, that's connector 1 to the positive end of the pack where connector 4 is. All the connector 1 cables are connected together. All the twos are connected together and so on. Electrically it's the same thing as welding up a pack with 12 cells in parallel by 20 cells long. The stronger cells and the weaker cells can't discharge at different rates. Being that all 12 cells are connected together in parallel, they act as a single cell...a single cell that is the aggregate of all 12 cells. IE: it doesn't matter that there is a weak cell or not, it can't discharge faster than the rest of the cells it is in parallel with. This is precisely why when I do load testing I only do 2P at most. One or both of those two cells in parallel are weak on a channel that discharges too quickly, but I won't know which cell until I try both of them individually. Most of the cells I harvest that actually charge to 4.1 volts are good for 1800mah at least. It's not dead set to be true all the time, but most of the time it is. Don't ask me why that is, but it seems about 80% reliable. Cells that won't fully charge typically wont get close to 1800mah. A fair percentage of cells charge to 4.1 volts, but are less than 1800mah so I can't depend on fully charging as proof of a good enough cell...just that if a cell doesn't fully charge,that it's already a throw away. Assuming I get the Ah for the entire pack that I am looking for (4000mah) and the cell voltages for each channel are above 3 volts when I get to 4000mah, then it's safe to assume that all the cells used in that load test are at least good enough or 2000mah. I usually load test until the RC battery monitors start alarming. It's quite possible to measure more than 4000mah as a result. It is possible to get a strong cell in parallel with a weak cell and as a result a weak cell slips through if I stop at 4000mah. By going until the monitors alarm, that usually means a channel or 2 has dropped faster than the rest of the channels. Those cells on that channel or channels contain one or more weak cells and need further testing. The rest of the cells in the other channels...assuming I have reached 4000mah are all good enough. It has happened many times that I don't get to 4000mah before alarms start going off. I keep a stack of charged up, but weak cells on hand that are marked as weak. When a channel dips too quickly, I pop out the cells in that channel and pop in charged cells in their place and then let the load test continue. The result is the rest of the 40 cells continue their load test where they left off.
Hi and low discharge cells used simultaneously...
Since all the battery packs are electrically the same being they are in parallel, the same load that goes to a 1C cell goes to a 5C cell and a 100C cell. In my LION pack I have a string of 40 18650PF cells. If I remember correctly, they are good for 5C. They are electrically cell per cell in parallel with the laptop cells so no one cell can discharge more quickly than another. If the 18650PF measures 3.9 volts, the other 11 laptop cells are going to measure 3.9 volts as well...even if I disconnect everything and measure them separately. AS far as mixed capacities is concerned between different packs...I have a 20C 20S2P 20,000mah LIPO pack in parallel with a hodge podge of laptop cells that all meet the "good enough" criteria. Since the load can't choose which battery source to pull current from because the LION and LIPO packs are in parallel, whatever load there is gets spread evenly among all the battery sources there are. If I'm pulling 50 amps...more or less each cell sees 1/14th of that 50 amps or 3.57 amps each. Fortunately, the laptop cells are just fine with that load. and at most I pull 60 amps. However, there is another factor that plays into things. IR is the internal resistance of the cell. Cells with low IR are going to see more current loading than cells with higher IR. That's just basic ohms law. Lets pretend my LIPO pack has a 1 ohm IR for the entire pack and the LION pack has a 2 ohm IR for the entire pack. That effectively means the LIPO pack is going to see 2X the current load that the LION pack is seeing. By the way, my LIPO IR is a tad lower than the LION IR, but not 50% lower. The result is the LIPOs will run down first and then their BMS shuts them off. I definitely notice when this happens, but the IR's for both pack types is close enough that it's just a minute or two that I'm running LION only and battery voltage is so low anyway, that I'm better off walking.
My LIPO pack is on a seperate BMS from the LION pack. The LION pack is good for 24,000mah. The LIPO pack is good for 20,000mah. 4000mah is hardly anything in the grander scheme of things so my packs despite being differing chemistries are pretty close to the same capacity. If I was running without a BMS on each pack type, it would be problematic to run down the packs to 60 volts because the LIPO pack would get to 3 volts per cell and continue to discharge while the LION cells were still 3.2-3.3 volts per cell. This would be bad for the LIPO pack.
Don't let the mega and giga watters on ES talk you out of using laptop cells. They aren't necessarily wrong. Expensive and high discharge rate cells are better. you probably bought them new so you are getting the best capacity they have to offer. This is all true. However, sufficient numbers of 18650 31Q cells to run my scooter would cost me about $1400 while my scrounged laptop cells cost me nothing but the expense of finding the good ones and making battery holders and a few other misc items...or about $250. Depending on your economics and budget, which one is better is a matter of perspective. Me personally, I have access to used laptop battery packs and I can make battery holders easily so my economics say use the free cells.