Gap distance between main bus bar

[SAMGO]

Hi to everyone, hope you are all doing well,

I'm now in the process of building a new DRSSTC and I have a question about my inverter design.
My inverter is based on two GD600HFY120C2S, see datasheet in attachement, running at around 560VDC from 3 phase supply. For the bus capacitance I'm using a 2s2p configuration of 4700uF 400VDC caps. The input of the 3 phase 80A rectifier is protected by 10A type C circuit breaker for added safety.

The gap between + and - it's now 2mm, this is okay for 560VDC but I was wondering if this could be a problem because of transients voltages.

If a transient comes it will be limited at the voltage necessary to fill the gap between + and - . In this way I can protect my IGBTs, maybe at the cost of loosing my 3 phase rectifier but it's cheaper and easy to change. The problem is that in this case the arc will also make the bus caps discharge and I have the feeling that is is not good and could kill the IGBTs in some strange way   

I'm now trying to decide between leaving it this way or putting a insulator layer / milling away some copper to enlarge the space between rails.

What do you think will be the better solution ?

Here is my Inverter so far, it's still missing the heatsink:




I also  made a small and simple circuit to safely discharge the main caps. The basic idea is to have an emergercy stop switch that will both open the 3 phase supply and put a power discharge resistor between + and - , this will also work if the 230V driver circuit it's un-plugged.

You can check it here:

[TMaxElectronics]

With the width of your bus bars you could probably increase the spacing quite easily by drilling one of the mounting holes off-center. That would allow you to move the bus bar on the outer terminal further out, increasing the spacing considerably.

You will also need a pre-charge resistor for the dc bus. The inrush current of charging up such a large capacitance directly from the mains is massive, and will very likely cause problems.

I also think that a 10A breaker for such a high power coil is a little small. I managed to trip the breaker on my 3 phase 16A outlet in a matter of seconds with my skm400 coil and I suspect your inverter might do the same
And a type C breaker won't help do mucht to protect the semiconductors. if anything goes wrong the bus capacitance is likely enough to kill them by it self. I'd recommend just not installing a breaker on your coil directly, since all it would protect is the cabelling and that is already protected by the breaker in the distribution box.

[SAMGO]

With the width of your bus bars you could probably increase the spacing quite easily by drilling one of the mounting holes off-center. That would allow you to move the bus bar on the outer terminal further out, increasing the spacing considerably.

You will also need a pre-charge resistor for the dc bus. The inrush current of charging up such a large capacitance directly from the mains is massive, and will very likely cause problems.

I also think that a 10A breaker for such a high power coil is a little small. I managed to trip the breaker on my 3 phase 16A outlet in a matter of seconds with my skm400 coil and I suspect your inverter might do the same
And a type C breaker won't help do mucht to protect the semiconductors. if anything goes wrong the bus capacitance is likely enough to kill them by it self. I'd recommend just not installing a breaker on your coil directly, since all it would protect is the cabelling and that is already protected by the breaker in the distribution box.

Thanks TMax that's actually a good idea, I'll take note and see what to do when my fat heatsink arrives. Actually the holes are already off-center but I forgot that the widht of the busbars was changed last minute bruh

For the circuit breaker it's true that it doesn't add a lot of protection. I think I will go with the 10A one for the test phase and that evaluate to skip it when everything work as it should if necessary.
 
I agree with you about the pre-charge resistor, there is no way it will work if I attach it to 400VAC directly. For now I will be using a variac to be able to run it at a lower bus voltage. When everything works I will put 3 pre-charge resistors that get shorted after some seconds by a relais and regulate power via on-time.

[davekni]

The gap between + and - it's now 2mm, this is okay for 560VDC but I was wondering if this could be a problem because of transients voltages.

If 2mm is air gap only, shouldn't be much risk.  A 2mm surface path (creepage distance) would be concerning.  Even for air, 3 or 4 or 5mm wouldn't hurt for comfort.  I'd think the risk is less from transients and more from any fine metal whiskers or other conductive debris falling or blowing across the gap and starting an arc.

[klugesmith]

>> I'd think the risk is less from transients and more from any fine metal whiskers or other conductive debris falling or blowing across the gap and starting an arc.

At work we had a problem with low-voltage busbar parts made of copper and plated with pure tin.   Tin whisker growth is a real thing, even today.

[Duane B]

A 2mm gap might be troublesome with thermal expansion, dirt and debris over time. You might be better off increasing the distance of the gap and installing some type of MOV across the bus.

[klugesmith]

You can look up clearance and creepage rules for your working voltage.
https://www.protoexpress.com/blog/importance-pcb-line-spacing-creepage-clearance/#Calculating_PCB_creepage_and_clearance

I would never depend on the 2 mm spacing as an intentional spark gap to protect components from an overvoltage transient.  So why not insert an insulating sheet to keep debris away from gap?

[SAMGO]

The gap between + and - it's now 2mm, this is okay for 560VDC but I was wondering if this could be a problem because of transients voltages.

If 2mm is air gap only, shouldn't be much risk.  A 2mm surface path (creepage distance) would be concerning.  Even for air, 3 or 4 or 5mm wouldn't hurt for comfort.  I'd think the risk is less from transients and more from any fine metal whiskers or other conductive debris falling or blowing across the gap and starting an arc.

A 2mm gap might be troublesome with thermal expansion, dirt and debris over time. You might be better off increasing the distance of the gap and installing some type of MOV across the bus.

You can look up clearance and creepage rules for your working voltage.
https://www.protoexpress.com/blog/importance-pcb-line-spacing-creepage-clearance/#Calculating_PCB_creepage_and_clearance

I would never depend on the 2 mm spacing as an intentional spark gap to protect components from an overvoltage transient.  So why not insert an insulating sheet to keep debris away from gap?

Thank to all of you for the suggestion! As a solution I will use a FR4 1.5mm spacer that is around 30kV of insulation so that I can first eliminate the debris problem (thanks I didn't thinked about that) and also augment the lenght of the creepage distance.


bdw Tin whisker growth it's very magical and spooky ahahhahahah those are the hard to find problems ....

[TMaxElectronics]

bdw Tin whisker growth it's very magical and spooky ahahhahahah those are the hard to find problems ....

especially if they immediately get blown to pieces when they make contact and disappear

I will use a FR4 1.5mm spacer that is around 30kV of insulation

Yeah that will probably improve the situation without too much effort. But I'm not so sure about the 30kV for 1.5mm FR4. I've seen normal mains (230V) cause tracking from one side of the board to the other before. But that might have been from heat/contamination with smoke machine fluid.