Page 1 of 1

Check the gate signals with a scope

PostPosted: Wed Jul 31, 2013 7:48 pm
by lancefraterUC
The failure to drive the IGBTs correctly is catastrophic. When testing with a scope and assuming most are atleast dual channel both gate signals must be checked. If the toroid is wound incorrectly the signals will not be conjugate (opposite) and both IGBTs will turn on at the same time (letting the smoke out). The gate signal are isolated from the low voltage ground so the ground probes must be moved to the gates. If the IGBTs aren't installed the grounds can be attached to the negative side of the DC bus and center point between IGBTs, the signals can be measured at the toroid outputs. The gate drive signals should appear opposite; otherwise one set of gate wires should be switched

This is a great chance to teach people how to use a scope in isolated circuits this way.

Lance
University of Canterbury, NZ

Re: Check the gate signals with a scope

PostPosted: Fri Aug 02, 2013 1:11 pm
by Bayley
Phasing is strongly emphasized in the user manual. The chances of switching the phasing is minimal with a little care during assembly.

We chose not to include a "check your phasing" step in the manual since, even without IGBT's installed, one slip of the scope probe shorts the outputs of the isolation transformer, resulting in damage to the UCC's. This damage is extremely frustrating to diagnose, since often it manifests itself as increased impedance in the output stages of the UCC's, loss of control on the ENABLE pin, or other modes which result in a partially working IC. (I speak from extensive experience with this particular chip here!)

And needless to say, scoping the gates with the coil operating is extraordinarily dangerous and should be avoided when possible. Our manual has no steps in which the user has to troubleshoot an energized board.

Re: Check the gate signals with a scope

PostPosted: Sat Aug 03, 2013 4:41 pm
by lancefraterUC
Bayley wrote:Phasing is strongly emphasized in the user manual. The chances of switching the phasing is minimal with a little care during assembly.


It is evident that with the number of fuse blows, diode failures and IGBT shorts reported in the forums that mistakes are happening or the signalling at the IGBT gates is not correct (large over shoots). I still think that with much care the gate signals should be checked. The loss of the IGBTs and diodes is much more costly to the user than failure of the UCCs. Are the series gate resistors not suitable to limit the UCCs and avoid damage?

Bayley wrote:And needless to say, scoping the gates with the coil operating is extraordinarily dangerous and should be avoided when possible. Our manual has no steps in which the user has to troubleshoot an energized board.


I agree

Re: Check the gate signals with a scope

PostPosted: Sat Aug 03, 2013 8:58 pm
by Bayley
lancefraterUC wrote:
Bayley wrote:Phasing is strongly emphasized in the user manual. The chances of switching the phasing is minimal with a little care during assembly.


It is evident that with the number of fuse blows, diode failures and IGBT shorts reported in the forums that mistakes are happening or the signalling at the IGBT gates is not correct (large over shoots). I still think that with much care the gate signals should be checked. The loss of the IGBTs and diodes is much more costly to the user than failure of the UCCs. Are the series gate resistors not suitable to limit the UCCs and avoid damage?

Bayley wrote:And needless to say, scoping the gates with the coil operating is extraordinarily dangerous and should be avoided when possible. Our manual has no steps in which the user has to troubleshoot an energized board.


I agree


The gate resistors limit the output current to about 2.5A, which will overheat the UCC's if shorted for too long.