[hub]

I found the RCCB, it's a Moeller single phase. Connected it to the machine wiring and seems to work. It's a "G" model with a 10ms delay.
Will see what happens next time after a long power off. Looks good so far

[wizard]

We may have a communications issue here, I know that electrical systems can be implemented differently than in the USA so terms may differ.

Thank you for the reply

The breaker is:
ABB Vikavirtasuojakytkimet - Vikavirtasuojakytkimet (Moduulikojeet)
Yes, it's the main panel breaker.

I don't think it has a current rating, except for the 40A/400V..

That really doesn't add up. When I refer to the main panel breaker that would be the master breaker for the entire panel box. Usually in the USA that is a breaker rated from 150 to 400 amps depending upon the house. It sounds like you have a branch circuit breaker tripping. If that is the case an inrush load might trip it. I'm still leaning against an inrush trip though.

I'm pretty sure it's leakage that's triggering it.. but somehow caused by drivers powered off for a "long" time, just my best guess though..

A power supply with a bad capacitor could cause this to happen. Other part could also potentially fail in a power supply. This is fairly easy to isolate out as you simply disconnect one or more power supplies and power up to see what happens.

I don't want to completely dismiss an inrush current problem either, multiple power supplies turning on at the same time could be an issue. It really depends upon how fast the breaker is. However you would see an inrush trip after a few minutes of a power down.

I'll try your isolating suggestion, just will take some time since the 16h+ wait in between.

Thanks again wizard,
Hub

[wizard]

If you mean a RCCB then no. I guess that could be a solution, to add a dedicated one to the machine, easy to test as I have an extra somewhere if I can find it..
I only have an automatic fuse on the machine side.

Also, my VFD has a separate connection, it's not powered on with the rest. (It's connected to the same phase though).

Here in the USA it isn't considered good practice to have GFI's wired in series. My understanding is that GFI's are similar to RCCB so having two RCCB connected as you are suggesting might be considered bad practice or even illegal in your local.

Adding a breaker, even a plain breaker, might not do much for you in this situation. The speed of the breakers can be a factor determining which one trips. That being said I prefer to see all power supplies, in a machines controller, fused or protected by a circuit breaker.

[hub]

We may have a communications issue here, I know that electrical systems can be implemented differently than in the USA so terms may differ.


That really doesn't add up. When I refer to the main panel breaker that would be the master breaker for the entire panel box. Usually in the USA that is a breaker rated from 150 to 400 amps depending upon the house. It sounds like you have a branch circuit breaker tripping. If that is the case an inrush load might trip it. I'm still leaning against an inrush trip though.

A power supply with a bad capacitor could cause this to happen. Other part could also potentially fail in a power supply. This is fairly easy to isolate out as you simply disconnect one or more power supplies and power up to see what happens.

I don't want to completely dismiss an inrush current problem either, multiple power supplies turning on at the same time could be an issue. It really depends upon how fast the breaker is. However you would see an inrush trip after a few minutes of a power down.

Thanks wizard,
If you mean the main switch by main power breaker/master breaker, (3 phase on/off switch), I can't see it's rating without opening the panel. The input is 20A 230V / phase to the main panel. (3 phases). The input has 20A fuses for each phase.

It's the RCCB that's tripping (leak current). It "monitors" leakage between L & N. and L & PE and also N & PE if I remember correctly. (For each phase)

I'll continue to test / isolate devices to see if I can find out which is causing this. Or if it is a combination of many...

[hub]

Here in the USA it isn't considered good practice to have GFI's wired in series. My understanding is that GFI's are similar to RCCB so having two RCCB connected as you are suggesting might be considered bad practice or even illegal in your local.

Adding a breaker, even a plain breaker, might not do much for you in this situation. The speed of the breakers can be a factor determining which one trips. That being said I prefer to see all power supplies, in a machines controller, fused or protected by a circuit breaker.

It's wired in parallel. But I'll check if that's legal here, haven't connected like this before.
Everything is fused. I have a 10A fuse for the drivers & power supplies, which will give me max 2.3kW @ 230V single phase. It's all I need, I don't expect the servos to draw more power than that at any time.
All together; drivers, PSUs & VFD are fused with a 16A. (Both fuses are C-curve)

[hub]

I got a mini spindle (dremel type) 170W for high RPM stuff like engraving, grinding, sanding etc. It will do 8k to 35k RPM.
The "main" spindle is sometimes limited by its' 7200RPM max.
It's installed on the machine next to the main spindle, will test it later to see how it works.

[hub]

"Machined" some steel flat bars as they aren't exactly straight.. (Grinding with the newly installed mini spindle). It was SLOW... but with good results. Got a few done, still some more to do.

Attachment 281974

[wizard]

"Machined" some steel flat bars as they aren't exactly straight.. (Grinding with the newly installed mini spindle). It was SLOW... but with good results. Got a few done, still some more to do.

You might want to consider the use of the larger spindle with a larger wheel for your grinding operations. It should be a bit stiffer. Either way you need to keep an eye on generated heat as you can easily distort a price like that due to the one edge getting hot.

Of course this means going out and buying a larger, 2" maybe, grinding wheel. You also need to make sure it is speed rated.

[hub]

You might want to consider the use of the larger spindle with a larger wheel for your grinding operations. It should be a bit stiffer. Either way you need to keep an eye on generated heat as you can easily distort a price like that due to the one edge getting hot.

Of course this means going out and buying a larger, 2" maybe, grinding wheel. You also need to make sure it is speed rated.

Thanks for the tips Yes it would be a lot stiffer. I think that's what I'll do for grinding, sounds like a good idea.

Still can use the mini high RPM thing for engraving though.

I used a diamond grinder bit just as a test. Worked ok. Not much heat, warm to the touch. But I did it very slowly with medium RPM.
The flat bars were a bit curved/distorted. Maybe it's how the were or maybe it was me when cutting them with my angle grinder, they DID get hot when cutting to length.. I didn't check them before cutting.

[hub]

I think I have figured out the RCCB problem.. If I connect all but the 1.5kW servo drive & filter there's no problem. The second RCD at the machine didn't help much.
So I plan to get some sort of delay switch for the servo drive. (First, still have to confirm it's causing it, and that it wasn't "random luck")

[wizard]

That is good news! You mentioned filter (hoping you mean line noise filter) and that has me wondering if the filter has any internal caps connected to ground. You may want to try start up with the filter bypassed.

I think I have figured out the RCCB problem.. If I connect all but the 1.5kW servo drive & filter there's no problem. The second RCD at the machine didn't help much.
So I plan to get some sort of delay switch for the servo drive. (First, still have to confirm it's causing it, and that it wasn't "random luck")

Don't knock luck! I'm not sure a delay switch will do a lot for you but you can try. I'd avoid buying one and try a toggle switch with the deal created by a human.

[hub]

That is good news! You mentioned filter (hoping you mean line noise filter) and that has me wondering if the filter has any internal caps connected to ground. You may want to try start up with the filter bypassed.

Don't knock luck! I'm not sure a delay switch will do a lot for you but you can try. I'd avoid buying one and try a toggle switch with the deal created by a human.

Seems it's really the problem, and that it's repeatable. Just powered on the machine in the same way after a long(ish) power down.
Works fine A separate switch for it will be just fine for me.

Yes it's a line filter. The other drives has line filters too, different models though. I'll try without the filter to see how it works. Do you think the filter might be defective?

[hub]

This is the filter for the 1.5kW servo driver:

[wizard]

Seems it's really the problem, and that it's repeatable. Just powered on the machine in the same way after a long(ish) power down.
Works fine A separate switch for it will be just fine for me.

You wouldn't want the separate switch in the circuit long term, I was thinking of it as a debug aid.

Yes it's a line filter. The other drives has line filters too, different models though. I'll try without the filter to see how it works. Do you think the filter might be defective?

Defective??? Maybe or maybe the filter is designed to work that way. The big variable here is that you are telling us that it takes awhile before a power up trips the breaker. That would cause me to lean towards a defective device.

The problem is in the past and even to some extent these days, it was common to use capacitors connected to ground to try to suppress electrical noise on an AC line. The problem is the capacitors can cause significant amounts of current leakage to ground. This is enough to trip a GFI. It can be an issue of bad capacitors or it can simply be the result of the capacitors designed into the circuit. I don't know about the specifics of the filter you are using but bypassing it and trying different power up scenarios should tell you if the filter is causing the problem.

I don't know about this specific filter but I have had experience with old temperature controllers that passed considerable current to ground via capacitors connected to AC circuits. Generally this seems to be less of a problem on newer hardware. By the way these where new out of the box controllers, we aren't talking failed caps here. Line filters vary a bit with respect to their internal design so you can't say for sure if the device has failed without the technical specs.

[wizard]

This is the filter for the 1.5kW servo driver:

Whoops I should have opened page 7!

The schematic on the device confirms capacitors connected to ground. That is one piece of evidence. The second piece is that they have stamped it to indicate high leakage currents. My guess is that this device is working as expected by the manufacture.

I'm hoping everybody reading this takes note that grounding is very important in industrial equipment as devices like these filters can produce high ground currents, high enough to kill in a machine not properly grounded.

[hub]

It's funny I never noticed the "high leakage current" print on it until you mentioned it. Doh!
Thanks for your valuable input, Wizard.

Hub

[hub]

Finished the steel & AL parts today for the X & Z axis. All the parts are now made. I hope tomorrow I'll be able to disassemble X & Z and assemble the new parts.
Well, at least disassembly.. and some assembly in one day.

[hub]

Got some new parts assembled today:

Attachment 282652

Attachment 282654

[hub]

More assembly today. More than half way through I guess... Takes lots of time to get everything straight & square

[hub]

Got most of the mechanical parts assembled I'm happy with how rigid it turned out. Might even work long-term..