I think what you are looking for is the Wh used by going from full to a certain voltage at a given current?
yep that
sorry if i explained badly
but is there a graph to represent it
That is a different graph for every cell. And it changes depending on whether the cell is at rest or by how hard the cell is discharged (when voltage sag changes things).
Are you configuring something that displays a percentage or state-of-charge and need more accurate numbers?
1 Like
not really
no prob if you cant
i have a 12s8p p42a and i noticed today i used around 300wh of capacity and it dropped by around 0.3v (4.15-3.85) just wanted to know if that is reasonable and how much more capacity i will get between 3.9-3.6v
I can. I just had no idea what cell you were using.
My P42A test report has a graph with capacity delivered vs voltage for some different discharge current levels: Bench Retest Results: Molicel P42A 4000mAh 21700...beats 40T, one of the best batteries available | E-Cigarette Forum
I don’t know if that is reasonable and what you will get between 3.9V and 3.6V per cell depends on how hard the cells are being used.
By far the best way to do this, due to all the variables, is to measure the Wh (energy) you actually get. Or the delivered capacity if that is the only number available.
1 Like
We actually have a (pretty old) thread about this that I would love to revive/get your input on, if you find some time to check it out!
2 Likes
Some good stuff there. Graphs like the ones that are linked to are a good way to derive the SOC vs Voltage numbers.
It can be done using the resting voltage (wait at least two hours) or it can be done at various load levels. Which is the problem…there is no one “correct” set of numbers. It all depends on what you are using the numbers for and when you are using them.
Coulomb counters can help out but typically need to be configured for the particular cell. But they’re often not great with pulsed current applications.
There are other chips that can estimate SOC based on voltage but the pack needs to be occasionally cycled and let come to rest at different SOC levels for the chip to calibrate itself to the particular pack.
1 Like
ah nice
all the numbers i have are from metr
roughly per parallel group its 10amps max
thanks for the link and will look more into it
1 Like
Am I drawing a correct conclusion here that if I charge to only 3.9 volts I would not be loosing much capacity? @10amp load.
I saw some information a while back that charging to 4.1 instead of 4.2 could double the cycle life. So that’s what I have done for years, I knew there was very little capacity between 4.2 and 4.1 but this graph seems to show there is very little capacity until 3.95v.
(Assuming you have a smart BMS / charger that could balance to 3.9v)
I guess a lower starting voltage will sag to a lower level under load, so whats the best compromise for charge voltage if your main concern is cycle life and your only pushing 10amps?
1 Like
TBH, I don’t know if discharging down to 3.9V and starting is the same as charging up to 3.9V and starting. I think they would be the same but I have never done a direct comparison.
There is no best compromise. 
Just whatever one best fits your personal references and priorities. I can only recommend what everyone should try…experiment and see what kind of run time loss you get for charging to a couple of different voltages. See how low you can go before the capacity loss results in an unacceptable shortening, for you, of the run time for the pack.
I would start at 4.0V to 4.05V or so as I have used that before with clients and they were willing to accept the shorter run time for that voltage.
2 Likes
I got a question for yall. I was putting together some pgroups out of used cells…I matched everything up pretty perfectly mostly because all the cells were at the same voltage accept for a few. Here is my question on one of the groups, this is a 9p group…8 of the cells were at 3.5 and the last one was at 3.9…the group averaged out to be around 3.6…did I screw something up? Or will that one simply average out with the rest over a period of time.
1 Like
From what I know about cell recycling, looking at the voltage of each cell alone doesn’t tell you much. The best way is to do a full charge/discharge test and match up cells with similar measured capacities, as well as their internal resistance. I dont know how to do all of that, so I just stick to using new cells.
Interestingly, I actually just spent my night trying to fix my buddy’s ebike pack that we made with recycled cells long before I knew what I was doing with battery building haha.
4 Likes
Ok so actually these cells are all pretty new, havmt been used yet, but I had to deconstruct the pack for reasons. One of the groups was lower the the rest, and I mixed up the cells accidentally
I think for parallel groups the cells will eventually even out in charge, may take a few hours or a day to actually bring the higher cell down, but should level out.
2 Likes
Yes, the high cell(s) will drain into the lower cell(s), until they’re all the same.
That said, 0.4v of difference is a bit high to just slap them together and let them do their thing. The high cell could have dumped a LOT of current with that much difference between it an the rest, which is not good for either the high cells or the low cells.
In the future, don’t do that.
2 Likes
Yeah this is what I am thinking, a bit worried that the cell might have sustained some damage. But I guess we will see. Pack is almost ready for a ride. Worst case scenario I will just replace the whole group
That’s by far not the worst case scenario 
5 Likes
Ok yeah, I also think it could have been worse! Its probably fiiiiiiiiiiiiiine
2 Likes
What is the general opinion about the samsung 40T cells vs 30q’s? Thinking of building a new pack with them, because figured i could build a lighter pack with more discharge and slightly less capacity (wasn’t using the full range anyways), and also do a much better and safer job with it. I also looked at the molicel p42a’s but the price is very steep (more than double of the 40T’s), and i don’t really need to extra discharge.
This isn’t comparable, because they are different form factors.
Any 21700 is going to blow away 18650s — but they will be heavier, take up more space, and likely be more expensive.
Compare within the same form factor.
If you want a cheat sheet, just get Molicel P26A (18650) or Molicel P42A (21700). Those options are pretty hard to beat, unless you can’t get those in your country. edit: (this was written in 2021 and seems still good for 2022)
1 Like
I’m comparing 40T’s with 30q’s becuase I had 30q’s in my old pack. The old pack was 12s5p, and i’m thinking of building 12s3p with 40T’s. This means 12Ah with 90A continous discharge instead of 15Ah with 75A continous discharge. Also the 36 40T cells would be lighter by about 1kg than the 60 30q’s. If I was to go with molicel P42A’s, i would still need to go for 12s3p to keep the capacity, but hen the price goes from about 250usd to 650usd.
3 Likes