FlexiBMS Lite - Flexible Configuration BMS w/ CAN-bus

I’m pretty sure this is an old thermal picture of an earlier HW version (looks like “0.2” on the silkmask) at 10 Amps, but I can get a fresher one next week.

The still air and relatively insulated internal space in the enclosure, limits the amount of heat the BMS can get rid of and as this is the method that is used to balance the pack, then it directly limits the speed at which it can be done at.

What wires specifically?

3 Likes

I see a lot of cancelled and waiting payments on the payment processor’s control panel.

Couple customers have reported problems with paying for their orders, so I’ll be checking with Paytrail what the possible causes for the failed payments could be.

EDIT: I’ll start a FinnkoTronics webstore [SERIOUS] thread on the forum today, so all the store related discussions can be had there in the future and leave this thread to the project logging and discussion.

5 Likes

Nice one Simon!
Just got 3 for some customer´s boards.
Will use them in future Board too i think :slight_smile:
Keep doing good stuff!

1 Like

I paid last week and my payment is still in pending I think this is normal for international visa transactions. Fingers Crossed.

1 Like

So I took some new thermal images of the Lite with different charging currents.

It was hooked to a DC load and I was supplying different charging currents via programmable PSU. The temperatures were allowed to settle for 10 minutes before taking any images. Room ambient temperature was 22C.

5A:

8A:

10A:

11A:

12A:

Used charging cables were quite thin at 20AWG, which is starting to be on the limit for what I would recommend for 10A, but still usable.
Thicker ones could be used to lower the resistive losses in wiring itself and to conduct heat away from the BMS itself for extra cooling. I’d say with thicker wires (16AWG), 12Amps is doable, but I can’t say how the thermals will be if the unit is stuck into a tight enclosure.

Wires (20AWG) measuring at 65-70C on the PVC insulation @ 12Amps

I’d expect balancing to be limited in this case, especially during the high current CC-stage. It would likely improve towards the end with the CV-stage, as the current starts to ramp down and there is less heating happening on the charging Mosfets and current shunt.

15 Likes

Awesome, thank you for doing those tests!
Could I request a high-current test (10A or or the max you’re comfortable with) done with the unit in shrink wrap or under a cloth? Something to simulate a tight fit in an enclosure?

5 Likes

Yea, I can do some more thermal tests, currently doing some heatsink tests out of curiosity.

Probably 3d-print a little cover for it to limit the dissipation.

4 Likes

if it helps i did some conformal coating, and can say using thin acrylic in the hot area works well, using silicone everywhere else works…

on one board it managed to insulate a little to well and a resistor melted off :smiley:

2 Likes

Yea, that shunt resistor is thermal density wise the hottest component on the board. Better let it breathe.

No need to go through that trouble just for my test request. Just toss a cloth over the board…my favorite quick&dirty method. :grin:

Great to hear about the heat sink tests. I’m very interested in seeing how much of a difference it makes!

4 Likes

I would also be quite interested in the test Mooch suggested. Planning a 12s build that will use the flexi and a 10a charger, and would love some reassurance that it will be ok :grin:

3 Likes

I´d expect 8A on 12S should be perfect.
It will get hotter in a tight enclosure and with higher room temperatures in summer.
So a bit headroom is always a nice thing!

1 Like

So I did another thermal test like @Battery_Mooch suggested and wrapped the unit in a shirt. Test current was 10 Amps, and I let the thermals stabilize for 1 hour.

I also hooked up the USB and was logging the measured temperatures. Note that the measured temperatures on the PCB are from the internal temperature sensors in the LTC6803 and STM32, so they aren’t right on top of the hot end of the PCB, but rather the LTC6803 is pretty much in the middle and the STM32 is closer to the USB-connector end, which explains that the stabilized temperatures showed 63C on the LTC and 51C on the STM32.

It’s about 100C on the Mosfet and the shunt resistor. I would call that “toasty”, but I suppose this is without being really able to radiate or naturally convect any heat away. Don’t know if it would be worse if there was a bit more internal space, but the ambient temperature was a little higher. I’d say use bigger wires, to conduct more heat away from the BMS and lessen the extra heat generated in the wires.

I might make a 3d-printed enclosure/box to do some more thermal tests.

10 Likes

anyway you could get a heatsink made for it

soemthing real simple like a thin metal piece to bolt on under?

Thank you! Yea, pretty toasty but expected as this is pretty close to a worst case scenario. It’s about as hot as I’d want to see the case of a 150°C junction rated FET get up to.

4 Likes

Simple to do but a thin piece of metal will quickly “heat soak” and rise to almost the same temperature unless the heat can be moved somewhere else.

4 Likes

Talking about a thin piece of metal.

I tested a TO-220 heatsink with a sticky thermal pad on the Mosfets.

10 Amps

12 Amps

13 Amps

13 Amps, different angle

3 Likes

attached to the enclosure wiht a hole so airflow?

:scream: wow…
But I don’t think we could set it up that way in an enclosure. Out in the open it can get the convection currents going. Inside and enclosure convection is quite limited and the ambient temp is higher. It would get banged around pretty good inside an enclosure too. :slightly_smiling_face:

Heck of a temp difference though!

1 Like

I just wanted to try it out, so I shared the test results, not that I was gonna start shipping those heatsinks, or recommend using them.

2 Likes