Guru
Joined: 26/07/2007
Location: AustraliaPosts: 1686
| Posted: 12:30am 28 Mar 2017 |
My observations are like this.
I originally used a power jack main board, and it had 50R resistors on the gates, and no diodes. Worked fine with the PJ drivers, but the 8010 using the 2110 blew up very quickly...... would run for a bit, but heat up very very fast.
So went to 5ohms, and the heat disappeared, and suddenly everything worked perfectly. When you can drive it up to 20kw to start big motors, you no longer worry about the drive. You can drive it all day at 4kw, and it does not care, with excursions to 8kw and back.... still runs fine.
Mad has had very good success with SOME of his tests, and has driven 8kw for long times I would not believe..... other times he has self destructed quickly, and at low loads.
Torroids take no prisoners, and this aspect is the most important I think, any tiny frequency drift, or loss of a wave form for any reason ( somehow turning off the 2110 for an instant) will kill the fets dead in one cycle.
I have had one unit die from idling for a few days, then dumping the house load on it. It had done a few MWH under harsh loading, but the idle seemed to be the killer. It idled for days, then at the first sign of load, blew up instantly.... could be the transfer switch too......
This does not look like a driver problem, but a back emf problem, between the chip logic and the load change after a long time doing nothing. This may never have happened if it was a E-I tranny.... or was it the transfer switch???
There are quite a few of these running happily driving houses around here with no sign of trouble.... but I do still wonder why the thing died on idle, and yet none have ever had problems at high power... they lap it up.
So I dont know why ... if the driver is so badly out of whack @ 5R, why it works fine so far.
I don't understand how tinker has had such a poor run... yes he has unorthodox layouts, but my experience so far has been that these things are pretty forgiving... if they work for me they should work for anyone..... and I have run these things hard for days with them spread out all over the place on the bench.
The drive info is salient, but I feel it is something else altogether.
The 10pn connector has very thin wire in there, and that will drop the current too.
The chinese boards using the egs002 boards have very low R in their gates too..... some just tie the gates together, and run a single 5r resistor... they still work just fine..
It just does not seem to be critical, much as the theory says it is.
I'm just not convinced so far.
oztules
Village idiot...or... just another hack out of his depth