Guru
Joined: 02/02/2017
Location: AustraliaPosts: 1464
| Posted: 04:48am 01 May 2021 |
Mike,
my results with -5V bias is here
It was a power board that used IR2110 to drive the gates, not an onboard BJT totem-pole
The test pulled Gate low to -5V but still the high side oscillated.
Of course there is much more to this.
(A bit of my Hacker mentality follows)
The home inverter is near 2 years old. If it has ran for 1/2 that time due to low input voltage cut-off then that is 1 year continuous.
how many switch cycles?
365 days x 24 hours x 3600 seconds x 20,000 PWM rate = 6.3E10 cycles, and 1/2 that for each half bridge.
3.2E10 switch on/off events with this shoot through.
I'd love to use some simple changes to control this but right now I am happy with how
it runs. The air compressor start up test I did with it put 11kW
through 3 FETs. This has them close to SOA limits.
The shoot through must reduce the safe margin further.
It's only a matter of time..
My benchmark for inverter longevity is the 3kW Victron that lasted 2 years.
It failed when under 500W load. Not an 11kW event.
2 years so far with a board that costs $50?
That delivers 3kW regularly? 4 FETs per side too so it's easy to repair and cheap too.
I can live with that.