Author Topic: DIY Battery Pack BMS MOSFET failure question  (Read 272 times)

0 Members and 1 Guest are viewing this topic.

Offline djQUAN

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: ph
    • My DIY website
DIY Battery Pack BMS MOSFET failure question
« on: May 31, 2020, 05:26:39 am »
I have a LiFePO4 battery pack used in my car (about 10Ah 4S 12.8V) with a 30A BMS under the front passenger seat. Along with a supercapacitor bank (2.85V 3000F x6 in series with active balancing) in the engine bay for supplying the cranking current. It has been in use since January without problems and works really well until the other day I went to start it and the cap bank was at 8V. Battery pack was at 13.2V and the BMS died for no reason.

I investigated and the discharge fets have 0V on the gates. Charge mosfets are fine. Control ICs are fine. Could not find what is pulling the gates low (Nch fets) so I hooked gate to source to my bench supply set at 1A. Used the FLIR to find which part is getting hot to see which is pulling it low and narrowed it down to one mosfet shorted gate to source, drain to any other terminal is open. Cut off gate terminal and now it works fine with one less mosfet on the discharge side.

The mosfets are AOD472 rated 55A, Vds 24V, Rds 6mohm. There are 5 in parallel with the BMS current limiting specced at 30A with two 5mohm shunt resistors in parallel for current sensing. At cold starts I have only seen 14A charging current into the battery at most but it quickly tapers down after a few minutes.

My question is how could the gate insulation be damaged? The only reason I could think of is over voltage on the gate be it a spike or sustained. I doubt it would be a spike as the large cap bank holds the voltage pretty steady. And I have a meter in the dash so I can see if it gets pretty high for a while. Voltage stays at 14.2V engine cold, 13.6V hot and does not fluctuate.

Also, the way the mosfets are driven in the BMS is that it has a 1Mohm pull up resistor to B+ and an NPN transistor pulls low to turn it off in fault conditions. There are 5 mosfets in parallel but only one got damaged. A spike may not be able to get through the 1M pull up against the 5 mosfet gate capacitances in parallel, could the miller capacitance pull it to dangerous levels with a fast current spike? Or just a random weak part from a chinese manufacturer failing prematurely? I do have a lot of ST STD95NH02L mosfets in the same package (80A 24V) to replace them if more pop and fail again.
 

Offline m3vuv

  • Frequent Contributor
  • **
  • Posts: 609
  • Country: gb
Re: DIY Battery Pack BMS MOSFET failure question
« Reply #1 on: May 31, 2020, 11:17:33 am »
ive , a homebrew ebike running 72 volts with a 100amp bms,its from china,i had to buy 3 before i got one that worked,my guess is its down to chinees crappy bms units.
 
The following users thanked this post: djQUAN

Offline djQUAN

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: ph
    • My DIY website
Re: DIY Battery Pack BMS MOSFET failure question
« Reply #2 on: May 31, 2020, 01:38:21 pm »
I have made a few battery packs now with various size and spec BMS's mostly with cheap modules whichever I could procure and various chemistries too. This is the first time I experienced one suddenly failing for (what appears to me) no reason that was why I was curious. If it happens again maybe a zener across gate to source would be a wise addition.
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3105
  • Country: fr
Re: DIY Battery Pack BMS MOSFET failure question
« Reply #3 on: May 31, 2020, 01:50:58 pm »
Or just a random weak part from a chinese manufacturer failing prematurely? I do have a lot of ST STD95NH02L mosfets in the same package (80A 24V) to replace them if more pop and fail again.

Could be even a fake part not matching the spec of the original, MOSFETs (along with capacitors) are notorious for being counterfeited.
 

Offline djQUAN

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: ph
    • My DIY website
Re: DIY Battery Pack BMS MOSFET failure question
« Reply #4 on: June 01, 2020, 03:03:46 pm »
Or just a random weak part from a chinese manufacturer failing prematurely? I do have a lot of ST STD95NH02L mosfets in the same package (80A 24V) to replace them if more pop and fail again.

Could be even a fake part not matching the spec of the original, MOSFETs (along with capacitors) are notorious for being counterfeited.

If it pops again or I get a little extra time, I'll take them all out and replace them with the ST devices and then test the parts originally installed in the BMS for Rds and Vds breakdown to compare with its datasheet.
 

Offline shakalnokturn

  • Super Contributor
  • ***
  • Posts: 1041
  • Country: fr
Re: DIY Battery Pack BMS MOSFET failure question
« Reply #5 on: June 01, 2020, 06:12:16 pm »
Same thoughts as you had either voltage spike due to inductance, you should be able to scope that or weak part Alpha/Omega are not the worst around but I've seen a lot more failed than ST/NXP/ONsemi...
 

Offline djQUAN

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: ph
    • My DIY website
Re: DIY Battery Pack BMS MOSFET failure question
« Reply #6 on: June 05, 2020, 05:13:40 am »
I replaced all the mosfets on the BMS.

As I suspected, the miller capacitance might have something to do with the unexpected damage. The original parts had a 280pF reverse transfer capacitance while the new ST devices I put in only has 90pF even though it is rated higher current (80A vs 55A) Ciss is nearly identical at around 2000pF.

I did some static Rds testing at 10V Vgs. The old parts measured at 6.1mOhm and the new ST parts are 3.52mOhm.

Total resistance between input and output of the BMS before the mod was 6.45mohm,
after changing the mosfets, it was 5.25mohm (includes PCB traces and current shunt resistances).
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf