Correct on the GDIs, the one saving factor is that if they only trigger on an actual E-STOP event then the power is already switched off. That was my thinking at the time., that it would only be the HV spike produced by the shutdown, so that under normal operation they would not be triggered.. But certainly the required protection will be added .

 

See what happens in the simulation with a 1200V GDT fitted across the supply, immediately after the switch. 1200V should have sufficient headroom to guarantee no spurious operation.

 

I was thinking more of the GDT device being across the motor side, given that in an "E-STOP" condition the normal power would be switched off. Protection on the battery side will require not only a circuit breaker but also some series impedance in the DC line. THAT might be a rather "Big Deal" to add.
AND, I also would bee a bit uncomfortable around a 1000 volt battery pack. I have serviced 140 volt, 600 amp battery systems, replaced the batteries in them multiple times, and it is always a bit tense until the battery string is opened in a few places. ( The fues is 600 amps, the capability, for a short term, is closer to 1000 amps, at least for quite a few seconds.

 

The OP is of the view that the switch off voltage spike is not generated by the inverter/motor (but I have my doubts too) – if it is only the switching voltage spike that is at issue, why not place a GDT at both the switch and inverter input?

 

The OP is of the view that the switch off voltage spike is not generated by the inverter/motor (but I have my doubts too) – if it is only the switching voltage spike that is at issue, why not place a GDT at both the switch and inverter input?

The inverter input is stated to be a 1000 volt DC battery pack. Such a battery pack is quite capable of delivering a MASSIVE current for a few seconds, which would explode any Gas Discharge Tube device that I am aware of before any useful fuse could clear.
Of course, if the GDT were installed on the switched side connection then that would not be a shutoff issue,. But for that sort of application a discussion with the application engineering folks from the GDT maker will be needed.

Really, a discussion with the application engineer from the inverter maker should be the first step.

 

The inverter input is stated to be a 1000 volt DC battery pack. Such a battery pack is quite capable of delivering a MASSIVE current for a few seconds, which would explode any Gas Discharge Tube device that I am aware of before any useful fuse could clear.
Of course, if the GDT were installed on the switched side connection then that would not be a shutoff issue,. But for that sort of application a discussion with the application engineering folks from the GDT maker will be needed.

Really, a discussion with the application engineer from the inverter maker should be the first step.

Thank you both for your responses and recommendations. As you recommended, I will talk to the inverter design team.

Anyway, I will continue to investigate these voltage spikes in case there is something we have missed.