The battery builders club

this is exactly what seems like is happening with my battery, except with imbalanced cells.

my charger keeps going in CC until one cell reaches the cell overvoltage cutoff, then the phantom voltage leaves and the highest cell drops down to about 4.11V. it repeats this many times, adding 0.01V to each cell at best.

the only way I’ve found to deal with this in the LLT BMS app is to set the overvoltage release to about 4.05V, but keep the pack overvoltage release and overvoltage cutoff the same and set cell overvoltage to 4.25V-4.3V.

this way it charges slightly past 4.2V but then cuts off letting the phantom voltage release and the balancer kicks in bringing the cells closer. so far this has helped me cut my 0.12V delta of imbalance down to 0.06V in ~1.5 days.

My charger won’t go into CV until the pack is >50V at least. I have opened it up though and there are 3 potentiometers. if I knew which controlled the current I could tone it down for better, quicker, easier, safer, better balancing.

I may also just be completely misunderstanding phantom voltage and/or balancing, someone please tell me if I am.

This is a good reason to do your voltage adjustments on the charger end, rather than the BMS. If you lower the charger’s voltage, it will go into CV mode earlier.

Easily checked with a multimeter. A clamp-type meter that will measure DC is best (Uni-T UT210E for example), but any meter that can measure current and voltage will do.
Just hook up the charger to a suitable load (A partially charged battery will work) with the multimeter set up to measure current, then twiddle the pots till you find the one that controls the current. Label it, and put the other two back the way they were.
Then, without the charger attached to the battery and the meter measuring the output voltage, twiddle the other two pots until you find the one that controls the voltage. Label it, and put the other one back the way it was.

Et voila, you now know which pots control voltage and current.


If you want to charge to a lower voltage, including a CV stage to top off the pack, then you need to set the charger to a lower voltage. Not the BMS.

The BMS will never allow the charger to go into CV mode if the BMS is set to cut off at below the charger’s voltage + the phantom voltage rise.

Yes, if the charger voltage + phantom rise is higher than the BMS cutoff then you can get lots of cycling on/off. This might not hurt anything, if it doesn’t happen very quickly, but it sure isn’t good.

Every time you stop current flow you get a voltage spike. At these relatively low current levels these spikes are small but they’re never going to increase the life of your BMS and could potentially shorten its life.


Lower current certainly helps reduce phantom voltage rise but I think you should be able to help things by lowering your charging voltage or getting all settings up and away from the charger’s voltage (which might not be a good idea, depending on the charger voltage).

You have a very complex set of thresholds set up now, all trying to handle the symptoms you are seeing, but IMO never addressing the cause of all the symptoms…which hasn’t been discovered yet.

Try a lower charging voltage, even if you will never use it. This is just for testing to see if the charge finishes properly, going through a full CV stage down to the termination (charge stop) current setting level…if there is such a setting.

Keep your protection settings high and start balancing about 0.2V/cell below your charge voltage, e.g., 3.8V/cell balance start for a 4.0V/cell charge voltage.


2018/19 everyone used Samsung 30Q

now Molicel P26A is the new norm it seems
i see that they have lots more A, but does it matter that much?

Also they’re 2.6Ah vs 3Ah which i don’t like… (if i’d have the space i’d go for 21700 cells)

In actual runtime the p26a ties or beats the 30q within our range. Rated capacity is generally not a factor we can look at for the high drain of esk8.


can’t really see what you mean

Moochs e-Rating is about the same and the as expected 30Q last more Ah than the P26A
only thing i see is that P26A has less initial sag and is about 0.1V higher (does this make much of a difference?)

in a 12S pack 0.1V adds up to 1.2V sag
But that just means i don’t have the same top speed?

sry, quite some time since i last looked into batteries

When draining at 10-15A, the P26A remains in a usable voltage longer than the 30Q. It also doesnt unbalance as badly.


what’s ‘usable voltage’ for us?

The p26a delivers as much energy total as a 30q at 10 amps, more at 15 amps, and can be pushed to higher currents. It is also much cheaper from what I’ve seen. The 30q may be better below 10 A, but it’s still much more expensive and most of the time we’re looking to use more power than that.

How I understand it

30Q has more initial sag and drops to 3.5V within 300mAh
P26A has less and drops to 3.6V at 300mAh

at around 1500mAh they even out at 3.2V

2300mAh both 3V

P26A more or less dies within the next 100mAh
30Q still has about 300mAh

are we powering off at a higher voltage or what am i missing?

using following charts

*looking at the 20A data

On the escore table, the p26a beats the 30qs at 15 amps. 30Qs have an unbearable amount of sag at 15 amps compared to the tolerable amount of the p26a.

In terms of usable voltage, when your battery is lower, the 30Qs are gonna sag down to the soft cutoff of your esc much earlier than the p26a. You’ll have more usable power on the lower end.

Not to mention how bloody unreliable 30Qs can be. Dead cells, really quick degradation, inconsistencies make them a pain.


I think the main thing is that energy, which is what really matters in the end, is basically the integral of this curve. Being higher for most of the curve outweighs the little bit at the end.


so this basically means we don’t really care about everything below 3V anyway (for battery safety?)

ok, found the meaning of the escores (Wh down to 3.2V)

didn’t know they have quality issues

One question though… how do i need to calculate range if using the mAh won’t do it?

I use total Wh divided by the Wh/Km number to find out approximate Km able to be traveled on a charge.

So this is how the electronics and batteries are built over at Boards of Sweden!

Cells and all the electronics are glued/potted to the enclosure with natural curing silicone.

Everything uses connectors so it’s a simple task replacing parts.

The Xenith and Bmses heatsink is permanently glued to the enclosure so you replace everything but the heatsink if something where to fail.

Enclosure has a layer of glass fibre on the inside and deck has a thick coat of epoxy on the bottom.

Enclosure is super rigid at each segment to protect the cells and handle impacts. While less material and a flexible epoxy is there so the enclosure can flex with the board.

3D printed pieces with a grid structure on the bottom siliconed to the enclosure and velcro plus hot glue for the wiring harness

Butyl tape to make keep the water out, and acts as a strain relief for the enclosure screws.

Always open for input & improvements!


but like…
Wh = V * Ah
So for a 12S3P (12 * 4.2V) * (3 * ?Ah)

wh is measured at nominal voltage.


1 Like

I use Wh because it takes voltage into account. it’s better than saying “I have a 10Ah battery” and that meaning different things when considering the Voltage. 1Ah on 12s isn’t the same as 1Ah on 10s. Wh however is more representative of capacity, so I use that.

what affects capacity values are things like internal resistance and the load applied. if we want to pull tonnes of amps from our cells, the capacity will reduce faster as a result. I don’t know of a way to accurately communicate capacity without caveats like that.

1 Like