I've been told that it's a bad idea to gate SCRs on when reverse-biased. Is that accurate? I'm reasonably confident that the SCRs won't actually conduct reverse-biased whether gated or not; are there circumstances where that is not true? Are there other problems with gating a reverse-biased SCR? Increased losses? Decreased lifetime?
In a former life where I worked in power electronics we would not gate the SCRs on in reverse conduction because it reduced the reverse blocking capability of the SCR. It would therefore be easier for an inductive spike or other transient to force the SCR on and ultimately lead to its destruction.
I don't have any evidence to back this up, it was "tribal knowledge" from our greybeard engineers. We were also working on medium voltage (up to 4kV) and high voltage (13kV and up) systems where we had a number of SCRs in series and it was very important to maintain reverse blocking voltage.
From an Infineon app note, page 19, 188.8.131.52.
Thyristors shall only be pulse triggered during the forward off-state phase. Positive trigger pulses during the reverse off-state phase will lead to significantly increased off-state losses due to the transistor effects caused. These losses adversely affect the functionality and may lead to destruction. Exception: For light triggered thyristors control pulses during the reverse off-state phase are permissible.
Infineon has told me that they can't really quantify the additional losses because it depends on reverse voltage, junction temperature, part number, and the exact unit you happen to have.
I also received this from SanRex today:
When the SCR is reverse biased and a gate signal is present, reverse leakage current will go up typically by 10 times or more. This increased reverse current will cause over temperature on certain areas of the die, which will deteriorate the die itself.