Yeah, there are pros and cons to each method. I make all kinds of crazy packs for work (up to 10S10P Li-ion setups). I’m just looking to further my knowledge 
1 Like
Haha, we are all here to learn 
@willpark16 They are packs of new drill batteries. Ryobi 40V packs. I got them for less than 50% of regular price from an overstock seller on craigslist. I bought him out 
Do not get 25Rs if you can avoid it. They are rated for 20A but are no better than Sanyo GA cells which are rated to 10A with respect to power output.
I recommend Samsung 30Q cells. They can be found somewhat cheap in eu
2 Likes
@PXSS I saw it here: https://endless-sphere.com/forums/viewtopic.php?f=35&t=72026
It’s at the bottom of the first page and a bit of the second page. @whitepony 's thread
1 Like
Do not get 25Rs if you can avoid it. They are rated for 20A but are no better than Sanyo GA cells which are rated to 10A with respect to power output.
I’m not following.
25R vs GA Samsung 25R discharges 20A at 77C while maintaining 92% of the rated 2500mah life (2300 total) Sanyo NCR18650GA discharges 20A at 113C while maintaining only 78% rated 3300mah capacity (2600 total).
The 25R looks quite a bit better than the Sanyo?
25R vs 30Q 25R discharges 20A at 76C and has about 2300mah capacity 30Q discharges 20A at 83C and has about 2550mah capacity
25R pulse discharges 50A at 80C for 14.5 min 30Q pulse discharges 50A at 84C for 16 min
So the 30Q has slightly better capacity, but runs slightly hotter and costs significantly more in the US
25R - $3.50-4.00/cell - 2300mah @ 20A - 77C - $1.74/Ah 30Q - $5.60/cell - 2550mah @ 20A - 68C - $2.19/Ah
10S4P cost 40 cell pack 25R - $160 @ $4.00 368wh 40 cell pack 30Q - $224 @ $5.60 408wh
I think either the 30Q or 25R is a solid battery and it’s probably worth getting which ever one is cheaper.
2 Likes
Dont forget our aplication is constant power and not constant current. Voltage drop plays a very big role in this scenario.
Voltage drop for a given current per cell. 25R vs GA: 0.2A
5A
10A
The voltage sag is slightly worse in every scenario in the 25Rs. That means that for the same power setting, you will have to pull more amps on the 25R than on the GA up to 10A. This also means that the 25R is more inneficient. Most people do not run continuous power at current draws larger than 40A. It’s usually just bursts which the GA cells can do just fine while staying under 60C.
If you’re drawing 20A average per cell while riding then I would not recommend the GAs, if you’re drawing 10A on average then the GAs have 33% more energy than the 25Rs.
25R vs 30Q 0.2A
5A
10A
15A
As you can see, the voltage sag on these cells is significantly lower than on the 25Rs. This means that these cells will run at a lower current for a given power setting, not stall during hard acceleration due to power drop, and still have a significant amount more of capacity over 25Rs at high average current draws.
I can easily source these under $4 per cell, both in the US and EU
E: Another good set of comparisons. This is the discharge energy at a specific current. 25R
30Q
GA
6 Likes
https://eu.nkon.nl/rechargeable/18650-size/samsung-inr-18650-30q-3000mah.html https://eu.nkon.nl/rechargeable/18650-size/samsung-inr-18650-30q-3000mah-button-top.html
These 30Q seem to be the same on Nkon, but the price is very different. Is that only due to the fact that the second has button-top? Is it worth paying more for that?
Thanks for the link! Really interesting conversation that went on. Learned somethkng about not welding directly on the center of a cell maybe? I need to go verify this on the webs.
1 Like
You dont need button top
Its threads like this that get me thinking. I have created a 10s4p pack as per diagram below.
(ignore the balance connector marks I know they are backwards)
I am not planning on having a bms and just using a brick charger with cutoff and then periodically checking the cells are balanced. I soldered the pack making sure that each connection was solid. (build of pack here)
Now my question is what happens if the connection to a cell in a parallel group goes bad? How would I know before it is too late? I am assuming that the voltage across the parallel group would still be consistant with the other cells and the capacity would be reduced. Would that mean that I would just be overcharging the battery each time and destroying them without knowing? How would a BMS stop me doing this? (maybe I just suck it up and fit one and divert the BMS for discharge)
I am not worried about a series failure as its hugely unlikely and rolling to a stop isnt the end of the world!
Norcs
Interesting on the 30Q voltage drop. Is that due to the fact that it has more total capacity though? I would expect a cell with lower voltage drop to run cooler?
The 30Q still gets a lot hotter than the 25R, looks like lygte’s tests had to stop at several high amp tests due to temperature?
I’m not sure that a small voltage drop will have any impact in practice? It would need more amperage, but a very small amount?
Due to duty cycle, actual battery draw is usually pretty light so maybe looking at it from a high amperage point of view is a bad approach and I should be looking at it from a capacity POV?
Overall the feeling I’m getting is the 30Q has a bit more capacity but struggles with high currents, and the 25R has lower capacity but handles high currents like a champ.
Thanks - now I have more questions than answers haha. Good discussion though. NKON has the 25R at 3.69 vs 3.99 for the 30Q. That price is about right for the differences we see in performance. Still having a hard time seeing a clear winner - will have to do some more research!
You can’t just replace voltage with current.
I think the only reason is the lower internal resistance of the 25r.
The 30Q is a really interesting, underrated battery and we´re group buying a few from nkon.nl for EU guys for 3,55€ at the moment 
You can’t just replace voltage with current.
I think at 100% motor load you are correct, but at less than 100% load the motor will just draw more amps if the voltage is less. Since we are talking 0.1v per cell I’m not sure there would be any noticeable difference while riding since most riding is done under 100%
Nice group buy! Very attractive battery at that price
That is not possible. The ESC basically just shorts the battery to the motor in a controlled way. The current that flows is not pushed but is the max current that can flow due to the internal resistance of the motor at that voltage. If your voltage is less you can not make up for that by “pushing” more amps. I have a 10S3P pack with 25Rs and I get voltage sag from 37V down below 33V. 0.1V multiplied by the number of cells becomes 3-4V in the end.
Ask @evoheyax about his experience with a 25R pack. I bet he won’t agree with you.
The 25Rs do not handle high currents like a champ. At 50% (1.25Ah) used capacity. If you draw 15A, you will sag from 3.7V to 3.2V. In a 10S4P pack that is a loss of 5volts, what should be outputting 2220W is now only doing 1920W. That is a power loss of 13.5%. If you’re trying to climb a hill, it will stall.
On the other hand, 30Qs drop to 3.45V at 15A draw with 1.25Ah consumed. In a 10S4P pack, this means 2070W. That is an improvement of 8% over the 1920W the 25Rs output. Things get much worse for the 25Rs the further down you move on the curve.
Yes the 25Rs might run cooler. They still have less power output at any setting than the equivalent 30Q pack.
Not only that, the sag means that you can only run it hard for about 50% capacity without hitting your lower voltage limit. Using 30Qs expands this to 80ish%
@Maxid How are you doing the full quote?
All cells will sag quite a bit when under load, so 37 to 33 is pretty normal. The 25R just sags 0.1v more than the 30Q under the same load, so 37 to 33 on 25R vs 37 to 34 (10x0.1) on 30Q.
I think what you are saying is the motor always takes the same amount of amperage, but the amperage is switched on and off by the duty cycle, so you’d have to have a higher/longer duty cycle at lower voltage to get the same wattage output, but the amperage wouldn’t change - correct?
But isn’t amperage = voltage / resistance?
40v / 1ohm = 40 amps? 50v / 1ohm = 50 amps?
So what I was saying before is backwards, lower voltage would also lower your total amperage?
Sorry - thinking out loud here, just trying to understand.
Power is constant. P = VI V = IR V is a variable I is a variable R is mostly a constant.
talking about the different ways to wire up 10s4p, either in a 10s chain and those in parallel or 4p all wired in series, the obvious benefit is with the 4 all connected in parallel they will balance each other and you now have 12 cells effectively to really worry about as apposed to 40…but…when I say worry, this method is safer in that the cells will balance each other but also it has a liability in that you dont know the individual cell voltages and neither would a bms and if there’s a dud in the pack of four you wouldnt know and you’ll get reduced output from the pack of 4 as the one sinks all the rest. a bms would likely need to do the 4 cells wired together as if all the cells are individual as in the 10s wired then in parallel would be too many individual cells for it to monitor so with a bms it’s not an option anyway but the point is when you put cells in parallel you dont know how they are individually doing anymore and you could be suffering from reduced output as apposed to when using 10s chains then wired in parallel.
also someone said when you were to put 10s chains then in parallel it will be a lot more current going through than the current going through with the 4 single cells being paralleled. true but it would have many more cells for that current to go to and ultimately it would be the same inrush of balancing current per cell. How much variance in voltage between the cells being parallleled is too much I’d like to know. in theory even if the cells being paralleled had a slight variation in current and paralleled it would be a huge current inrush. when cells are charged appropriately with a regulated supply they are limited to a constant current to avoid the high current that would be generated from a higher voltage connected to a cell
parallel connections pass very little current and its just the series chain that it runs in so the 4p then put in series would very likely also have reduced resistance with it’s reduced series chain wiring,…four times less