Floating Neutral with Single Phase Inverters in machinery

With no N. connected inside the machine, Voltages will vary and may reach up to 400 Volts on the individual L1, L2 and L3 terminals. The connected N. provides an essential reference point. Correct wiring should protect the machine from damage. No regs. required.

Edit. Add.

Three Phase Mystery Solved, No Neutral Required in a Balanced Load? - Bing video

Z.

Thanks for the reply Zoomup. I do agree with the correct wiring which was at fault. However I wouldn't of thought the single phase devices would go up in smoke due to the disconnected neutral.

However, we will conduct further checks in the future to prevent this happening again. 

Sy 

https://www2.theiet.org/forums/forum/messageview.cfm?catid=205&threadid=107966

Z.

Lost N causing a overvoltage on single phase connections is a well known problem - probably one of the reasons behind the preference for 3P rather than 4P switching for 3-phase, where leaving the N solid is permitted.

Other than the requirement for each circuit to have its own individual N, I don't know of any regulation that's intended to prevent this sort of situation - it's certainly permitted to omit N where it's thought not to be needed. Likewise there's no stipulation that 2300V single phase loads shouldn't be connected across phases - I guess a degree of common sense is presumed.

I am a little surprised that L-N voltages appeared correct if no N was supplied - unless there were some other loads connected on the feed side of the isolator and some of the N conductor was present.

The much worse situation is where the N also serves as a PE conductor (CNE or PEN) - in which situation you can find exposed-conductive-parts sitting at 230V indefinitely and even RCDs won't trip to save the situation.

   - Andy.

Thanks Andy for the response. 

That was the misunderstood situation where by we had L-N correct voltages. What happened is the Power side to the Servo Drives charged upon power being applied, this is thought to have created the over voltage down the N. Basically creating 230v on the single phase inverter & 400v on the N side to the inverter. 

I thought it was very strange to have had this outcome, Its can only be the Servo drives with them having L1 - L2 - L3 having no where for the voltage to go. So in theory everything that should of been single phase become 2 phase. 

Sy 

More accurately, the correct voltages were generated by a star of near identical impedances - with equal currents from all 3 phases to a floating  centre point there is no need for a neutral conductor to get the correct 'zero-centre' .  And when all 3 are on standby, the current through the mains filters and wiring capacitance is presumably near equal enough to fool you. But as soon as one of the 3 loads is not quite  the same as the other two, it all goes wrong, as without a neutral, the most heavily loaded phase sees a falling voltage, the star point moves closer to that corner of the triangle if you like,  and in true comedy see-saw fashion, the other two phases are then horribly over-volted. Of course as the voltage changes, so does the current drawn, and some loads will disconnect more or less impressively as the voltages go too far out of bounds, so the effect can be a few seconds of destructive psychedelics of flashing lights and motors buzing and stalling an then restarting and things blowing.
To test for a high impedance neutral, having passed the balanced with no load test then put a load you do not care about too much onto one phase only - a cheap  electric heater element is a good candidate. and see if the 3 phases remain more or less equal voltage. IF they go off piste, check those connections before switching on the expensive stuff.

Mike.

In my view, installing a FIVE PIN Ceeform outlet without a neutral is simply wrong. May not be specifically prohibited by regulations, but what about common sense, or accepted practice ? It is nearly as bad as fitting a 13 amp socket without an earth because only double insulated appliances are to be used.

If no neutral is required then the four pin type should be installed.

And, the installer of the machine should have read the manufacturer's installation requirements for the machine, and also noted the machine terminal markings and supply needs. Was there a neutral terminal in the main terminal box? We may all have been fooled by the lack of Neutral at the supply socket if it had an apparent  Neutral provision, and the supply Voltages seemed good. Testing at the machine with a neon Voltage tester or a sensitive digital multi meter may have not revealed the true picture. A tungsten test lamp or a small load may have revealed the lack of a Neutral as Mike has detailed above. An artificial internal Neutral may have been created by internal filters or other components.

Z.

Thanks Mike, great technical insight to what happened there. Pretty much exactly what happened. It fired up for 5 seconds ok, then suddenly went as the phases shifted. 

thanks for the suggestions below, anything that’s not fixed wired in future will be checked all the way to source. We don’t know if the controllers are damaged or anything yet until we replace the PSU and fire it all back up again. Luckily the PC is still ok.

thanks

sy

This was my point exactly, and the shear laziness of not connecting one core which was just coiled up is beyond belief. If there is five pins, there is five wires. No excuses on this one. If we didn’t require the N it would of been left out on the plug side. Not the socket side. 

thanks

sy