Agreed.
While any cell can be discharged at a level far above its continuous current rating the performance is terrible and the cell ages at a faster rate. At above about 25A the 40T struggles and the Samsung 30T is a better choice. It will run more efficiently and cooler, extending cell life.
4 Likes
Quick confirmation please… I’ve bought a blade fuse cradle and blade fuse to wire between the charge port and battery. I’ve seen it wired into the negative lead in other threads. Is this correct? Also, I will be using a 4amp charger so what size fuse is recommended? 10amp?
Thanks all
There is no “correct”; it can be on the positive or negative charge port lead. Generally it’s a good idea to treat + as “common” and just wire the positive together, and switch and fuse all the negative wires.
3 Likes
Tho this can cause issue with ground back feed so I put loop keys on the + use the ground for the common e.g flex BMS uses the - as the ground for the opto isolate if you connect the can ground then remove the loop key on the - all the current try’s to flow thro the can ground and magic smoke appears
What advantage dose the fuse on the neg give iv seen a lot of BMS common the pos then switch the negs but no idea why
1 Like
The reason for this is because due to the nature of physics, it’s easier and cheaper to make quality MOSFETs that switch the negatives on and off with a common positive
@Gamer43 may be able to give a more detailed answer.
Every BMS I’ve ever encountered had common positive and switched negative(s).
2 Likes
I do this as well, but beware in this circuit you can’t have any internal connections between the lights circuit and the ESC (bypass) circuit
Is it ok to mix batteries with different cycle counts? I’m thinking about cases where if one cell goes bad maybe 500 charges in, can it be replaced with an identical brand new cell? Similarly with LiPo packs, can I connect two otherwise identical packs except for cycle count together in series and/or parallel without any issues? Would it become unbalanced as it gets used and cause issues?
Its better to only mix cells of the same cycle count, and with the same capacity and internal resistance (which you have to use a machine to test for). Its usually less of a headache to just use all new cells.
1 Like
It’d always be better to get a new pack, but what issues can occur? Series should be fine, as that cell would just go out of balance and end with a higher voltage than the rest, but it’s not going to be a limiting factor, and it should just get rebalanced when the pack is charged. With parallel, I’m not sure what the interactions are.
If your p-groups are too out of balance from each other, your BMS may not be able to bring them back into balance. Most BMS’s have a maximum difference between the groups that it will balance, and if its outside of that then it will just lock out the charging and you will have to open up your pack to balance manually.
If one of the cells in your p-group is different than the others then it will be under constant stress. The older cells will want to discharge faster, which will put a larger load on the newer cell. Same thing in reverse when it’s charging.
Perhaps @Battery_Mooch can more elegantly explain why mixing old and new cells in a pack is a bad idea.
Too many scenarios to go through. Some only performance related, some safety related.
But since there could be issues, depending on the cells and the setup, we recommend not mixing old/new cells.
2 Likes
Are there any general guidelines on what constitutes an old cell? It seems like a waste of a pack if less than 50 cycles in, a cell goes bad, and I’d need to get a whole new pack.
Generally, no, not at all.
If it’s done a certain specific way then maybe. But only in that exact way. Without a machine that tests IR the only way is to
Have each P-pack contain the exact same amount of Age A and Age B cells.
So if you have 20 cells that are old and used TOGETHER the same amount, and 30 new cells, you could make a 10S5P with EXACTLY two of the old ones and three of the new ones PER P PACK. But this is inflexible and you cannot deviate from that, at all. Not one cell can be different.
In general, no, you can’t mix and match cells.
Also, none of the cells can be used laptop cells. Ever.
4 Likes
The best way to deal with that (if one cell goes bad, it will kill the whole P-group), is to remove the affected p-group and enough others to result in a number of cells that is evenly divisible by the S-count of your pack, so you can do what Brian said and put an equal number of new cells in each p-group.
You don’t have to scrap the whole pack, but you’re gonna have to tear it all the way down to bare cells anyway so you can shuffle them around.
This is why we put so much emphasis on building for reliability, in the form of proper cells, appropriate connection and construction methods, good BMS, the whole nine yards. Because it’s a huge pain when something goes wrong, and the best solution is to not have things go wrong in the first place.
7 Likes
Yeah that seems like the most economical way to do it safely. I was looking at the NESE system, and one of the promoted benefits is the ability to remove cells easily, so I wanted to see how that would actually be useful.
It makes the process faster and less annoying because you don’t have to redo all your spot welds. But you still have to shuffle the cells and make sure each p-group has the same number of new and old cells.
The other potential drawback is the increased size and decreased density of the pack, due to the NESE modules taking up space.
At the end of the process you’ll have a handful of old cells left over, and you can get yourself a power bank like this to stick them into!
1 Like
Yeah, but I was going to do a top mounted brick anyways so space is less of an issue, but it’s made it quite difficult to find prebuilts or used batteries, since they all tend to be flat packs.
Battery is definitely the most difficult part of my build so far. I originally wanted to go with LiPo packs because they’re affordable and familiar, but then I realized my charger (50W max) would take 48 hours to fully charge all four of them, and buying a quad charger is half the price of the batteries. Then I looked at some pack builders, but that’s way out of budget. Now I’m looking at the NESE system, which seems ok, but is neither particularly cheap nor simple. I could print all the stuff and do the soldering, but I don’t get the bulk discounts on cells and still need the NESE hardware with overseas shipping, which really eats into the budget.
1 Like
I feel ya. NKON was the best deal running for cells in the US while they were still shipping here. I’m just glad I got my ebike pack cells ordered before they stopped, that would have ben a lot of extra $$$. (150 cells)
1 Like
The manufacturers say a cell has reached the end of its life when, typically, it’s lost 20% - 40% of its capacity (depending on the cell). But that depends on how hard it’s been run, how it’s stored, and how it’s charged. And a cell doesn’t suddenly become unusable after losing 20% of its capacity. It might be frustrating to use but it’s not useless.
You might get 50 cycles or you might get 1,000 cycles. The “grade” of the cell and its manufacture date (how long it was stored before you bought it) can have an effect on its cycle life too but we typically won’t know the date and storage conditions.
It’s why we stress not mixing cells and matching cells where you can. It just makes things so much easier since there are so many unknowns.
4 Likes
Cost reduction
You don’t need expensive high-side MOSFET drivers or expensive P-channel MOSFETs to switch lowside.
The half-decent BMS will switch highside.
2 Likes