Oops…wait…the heat sink temp doesn’t matter since we don’t know the thermal resistance of the pad+sink and can’t calculate the FET temp.
We need to see the FET temp directly (with the sink in place) to get an idea how well the sink is working. Otherwise someone could stick a piece of wood on top, show how it barely rises above room temp, and say look how cool the FETs are.
My apologies, I didn’t intend to sound like I thought you were going to try to use that sink on your board. Thank you for that test!
I’m not surprised looking at the chinese BMS’ that they are usually sandwiched or have at least on one side a pretty big heatsink or have multiple parallel fets to lower the losses + increase the active thermal area for better dissipation, as they usually also offer discharge path control as well, so there’s extra thermal generation on the units.
So looks like 8A might be a good powerfull spot but not too toasty?
Do you mind testing again with the shirt 
?
Personally I wouldn’t go over 4A charging current unless you have crazy p-groups. It just wrecks your batteries faster. If you like your pack and you want it to hold up longer, reduce the charging current as much as possible. For my packs I just use as much Amps as I have cells in a p-group, f.e. 12s4p > 4A.
I’ll test it.
Not that people necessarily charge that fast if they don’t need to, but I would rather look at the current tests and questions specifically in the context of this thread, as “How high charging current could I use without overheating the BMS”.
More power to you and your testing, that’s awesome. That was just my two cents on the topic of fast charging.
Shirt test results with 8 Amp current. I’d say completely within operational temperature range. Measured temperatures via USB showed 45C for the LTC and 38C for STM32.
I then modeled and 3D-printed a little enclosure that I could put the BMS into and test what kind of temperatures I get with a little bit of free air inside, but no real space to get open air convection cooling. I’ll be making one more iteration with couple extra holes, so I can plug in the USB-cable too and see what the BMS is measuring for the temperature.
8 Amps.
10 Amps.
Webshop news. I’ll looking into adding Paypal as a payment method this week.
So 8 Amps.
That´s the way!
Thank´s for testing, i´ll get you more work in the future
Got the new case finished (thicker walls for better insulation and a hole for the USB-cable) and did another 8 Amp and 10 Amp tests.
8 Amps, LTC -> 42C, STM32 -> 35C
10 Amps, LTC -> 53C, STM32 -> 43C
EDIT: Walls are now 5mm thick
Somebody asked me about temperatures during balancing, so I did a balancing thermal test today.
Used battery is my production testing 12S pack, which I’ll be doing a complete balancing on with cell #5 being about 30 mV lower than all the other cells, meaning that 11 cells need to be bleeded. This pack is not meant for high current use, but is used as an actual test battery to check that all measured cell voltages are looking okay and within tolerance on the tested units and then do a balancing test for all 12 bleed resistors.
Balancing just started. The active resistors light up nicely in the thermal image.
I had initially set my BMS max temp to 72C, so it ran a little hotter at the beginning and you can see that many of the resistors are brightly glowing because of this. Dynamic temperature ratio was set to 3, meaning that for every 3 degrees below 72C (max temp) another balancing resistor is allowed on. If the balancing is started with a cool/cold board then at the start more resistors will be allowed on and the amount then decreases as the PCB warms up, until an equilibrium is reached somewhere below the BMS max temp setting.
I then set max temp back to the default 57C and allowed the heat to stabilize. And you can see right after switching the max temp lower, a lot less resistors are on, because the allowed amount of resistors ON is lower due to the max temp setting being lowered.
Then as the BMS had time to cool down, more resistors come on again as the thermals lowered. The thermal management is simple, but very functional and most importantly stops any overheating from happening from excessive balancing.
Did you reach the 100mA with these temps?
Excellent! Thanks for continuing to do all this testing.
You wouldn’t expect much additional heat from the MOSFETS being added to the board during actual use, would you? Since balancing is done near the end of the charge cycle when the current levels are a lot lower?
Inside a box might add a few degrees-C?
100 mA of balancing current?
Well the bleed resistor is 47 Ohms and and the Pfet adds a couple Ohm more, so probably around 50 Ohms actual at operating temperature. 4.15V / 50Ohm = 83mA
The reason why multiple resistors are warm, is that the balancing current is being switched over all of them very rapidly. The balancing algorithm prioritizes the highest voltage cells to be balanced first before the lower ones, but as the 11 cells are pretty much at even voltages at this point, the specific cells being balanced are switched constantly over, therefore heating all the resistors pretty evenly.
No I don’t believe that the BMS would be able to heat up the board enclosure in any effective way. It’s mostly as the PCB is just so small and tightly packaged, is that there isn’t much surface area to distribute the heat over, so you get hot spots even with very little actual waste power.
EDIT: also if you’re worried about the BMS heating the surroundings too much, you can turn down the max allowed temperature, so at least the balancing runs cooler, of course it’s also slower overall, unless it’s just like 1 cell that you need to balance.
Sorry, my question was confusing. I was wondering if the BMS itself would heat up much more when actually being used since the MOSFETs could also be conducting current. But balancing would occur near the end of the charge cycle, when the charging current level was low, so IMO the MOSFET heating would be minimal.
The other part of my questions was wondering if the BMS board itself would get hotter in the enclosure when balancing, i.e., higher temps than your tests, and IMO it would but not by much.
Great point about reducing the number of cells allowed to balance at one time…fantastic feature for those who might run the BMS in a hotter setup.
Maybe slightly, but as the resistors are not on the charging Fet end, they won’t contribute that much to their temperature (more the top-of-pack 10-12 cells, as they are closer to the charging Fets). So I would expect more of an even heating over the whole board, not necessarily hotter peak temps per say.
If the charging Fets are running hot, then that already does heat up pcb along with the LTC6803 and the STM32 (where we are measuring board temperature from), so you wouldn’t be able to use many balancing resistors anyway, as the balance resistor limiting will be lowering the allowed amount of balance resistors on.
Totally disagree let’s talk most common size of pack in ESk8. 12s4p. Samsung Q30 rate by Samsung at 1.5A standard charge so that’s 6A in a 4p. Or 4a fast charge 16a fast charge. Any higher is wrecking
@Battery_Mooch him self says below 0.75a there is no real notisable improvement that’s 3a in a 4p as a minimum for maxed batery life.
From experience 4p is the minimum for a dule motor build any lowers and your too compromised in your max load. I run a 3p comuter build and the batery is the limiting factor i have to limit the settings in the VESC. Lot of people out there running 5 or 6p you 4a charger max has no advantage fir them
End of the day batery care is a compromise of use and care. Your absolute max is only 66% of the manufactures standards rating and you get better life out of you battery’s by keeping them at storage voltage when not useing them than charging so slow.
Personally, I never go below 2A/cell charging for any cells I have except the ultra-high capacity (high internal resistance) 21700’s which I charge at 1.5A. I just don’t see any big hit on overall cycle life at those levels…YMMV.
P26A’s and P42A’s I charge at 4A/cell, minimum. This probably does affect cycle life but I don’t care.
My cells do get charged a lot faster, that is while I’m braking. I’m minimizing the standard charge current as much as possible to prevent any additional wear on them. My brake settings are always the highest that manufacturers allow in fast charging terms to their cells, which obv. puts out a lot of stress and reduces cycle life.
That’s why personally I’m recommending only 1a per cell generally. Many people set like -20A bat current on their 10s2p packs for their first buid, so I’m trying to mitigate further damage. That being said, what you said makes a lot of sense.
I’ll be making a video series explaining and demonstrating the use the BMS on my youtube channel.
I’ve 3D-printed a phone holder for my microphone arm, which makes it super convenient to be able to position and show what I want on it.
My phone fits snuggly to the holder right below the mic and I then have great adjustability for the position with the mic arm itself and then fine tuning the direction with the phone holder on the mic.
