How Hot Do BMS Fets Get?

rg12 · Oct 10, 2019

I have this new BMS I use that has an aluminum plate on both sides with one side sinking the smd resistors for the balance circuit and the other for the main bms fets.
Now, I chose to mount it so the balance sink is facing outside since I know from experience that this side gets pretty hot on a regular basis.
But, I never tested a BMS that has the fets on a separate side to see how it behaves under high currents (let's say 150A for a 72V pack), does it get hot?
I need to know how much to pad for distance away from the cells.

amberwolf · Oct 10, 2019

There have been threads here with posts showing that under the wrong conditions, FETs get hot enough to desolder themselves from the BMS PCB.

So...you'd have to test under your conditions how hot they get, with no airflow, and presumably the cells themselves up against that side (which will heat the cells it contacts, and the heat from the cells will heat the FETs).

rg12 · Oct 11, 2019

amberwolf said:
There have been threads here with posts showing that under the wrong conditions, FETs get hot enough to desolder themselves from the BMS PCB.

So...you'd have to test under your conditions how hot they get, with no airflow, and presumably the cells themselves up against that side (which will heat the cells it contacts, and the heat from the cells will heat the FETs).

What do you mean wrong conditions?
I'm talking about normal use within specs

I heard somewhere that they only work when they do the the on/off switch and the rest of the time current just passes through them but nothing is converted like you would use them in other devices that use them to turn current up or down and on and off a million times like a controller

fechter · Oct 11, 2019

If you know the Rds specification on the FETs, how many are in parallel, and your current drain, you can calculate how much heat is going to be dissipated. But temperature will also depend on how much cooling the BMS gets.

rg12 · Oct 12, 2019

fechter said:
If you know the Rds specification on the FETs, how many are in parallel, and your current drain, you can calculate how much heat is going to be dissipated. But temperature will also depend on how much cooling the BMS gets.

I never really noticed the fets getting even warm and was told that they don't really get hot since they only pass current and don't really "control" but I need someone to strengthen that theory.

flippy · Oct 12, 2019

if you have the 300A 32S bms from the bms thread then it wont get hot under 100A. if you are going to do a 100+ amps continous you will see "uncomfortable" temperatures. but those can be set in the firmware so it simply cuts out so you dont blow the fets.

if you go beyond 100A continous i might recommend simply replacing the mosfets with higher quality ones.

fechter · Oct 12, 2019

rg12 said:
I never really noticed the fets getting even warm and was told that they don't really get hot since they only pass current and don't really "control" but I need someone to strengthen that theory.

Your temperature test is a good way to know. Run it hard then stop and see how hot it gets. A FET that is turned on still has some resistance and will generate some heat, just a lot less than one switching on/off like in a controller. Knowing the FET Rds and the current, the exact amount of heat can be calculated.

rg12 · Oct 12, 2019

flippy said:
if you have the 300A 32S bms from the bms thread then it wont get hot under 100A. if you are going to do a 100+ amps continous you will see "uncomfortable" temperatures. but those can be set in the firmware so it simply cuts out so you dont blow the fets.

if you go beyond 100A continous i might recommend simply replacing the mosfets with higher quality ones.

I have the 24S version with 300A mostly using them for 100-120A but now building a pack I'm going to allow 180A.
Non of these numbers are constant since it's an ebike, it's a pull here, pull there, not even one minute of full throttle most of the times.

fechter said:
rg12 said:

I never really noticed the fets getting even warm and was told that they don't really get hot since they only pass current and don't really "control" but I need someone to strengthen that theory.

Click to expand...

Your temperature test is a good way to know. Run it hard then stop and see how hot it gets. A FET that is turned on still has some resistance and will generate some heat, just a lot less than one switching on/off like in a controller. Knowing the FET Rds and the current, the exact amount of heat can be calculated.

I don't know anything about what fets are in there but even if I did, I still have to know the average amp draw as it's going to be used in an ebike and I have no idea how the guy is going to ride it since everyone rides different and use different controllers (some programmable...)

flippy · Oct 13, 2019

open the bms and look at the fet numbers.

and ask the guy what currents he will be using.

rg12 · Oct 13, 2019

flippy said:
open the bms and look at the fet numbers.

and ask the guy what currents he will be using.

He will use an ASI BAC4000 with max 180A but It's never constant so I need to guess the average amp draw somehow...

How Hot Do BMS Fets Get?

rg12

100 kW

amberwolf

Administrator

rg12

100 kW

fechter

Administrator

rg12

100 kW

flippy

1 MW

fechter

Administrator

rg12

100 kW

flippy

1 MW

rg12

100 kW

Similar threads