Weird RCD Issue...

[marconi]

6mm 4 core flex has a vd of 6.7mV/A/m. So 15m of it has a neutral conductor vd of 0.1V/A or 1 Volt for every 10A. For 20A through each RCBO and SP supply vd neutral is 3 x 20/10 x 1 = 6V

It has a CCA of 32ishAmps. Might the neutral be undersized when operating from a single phase supply? 3 x 20A = 60A?

http://www.aeicables.co.uk/literature/CurrentRatings.pdf

Do you have a DVM with capture mode to examine the N-E voltage at power up. Some RCBOs use a rise in N-E voltage to detect loss of neutral and initiate a trip. That is why they have a functional earth.

 

[pc1966]

Do you have a DVM with capture mode to examine the N-E voltage at power up. Some RCBOs use a rise in N-E voltage to detect loss of neutral and initiate a trip. That is why they have a functional earth.

Most DVM would not see that, unless they are fancy scope-style models.

But your observation about the 1/3 size neutral in the single phase case, coupled with the big inrush from the loads listed above, could well be the explanation!

Temporarily putting the RCBO earths on neutral as well ought to answer that point, and you suggestion for a timed start-up of the phases might also save the day longer term. As an aside, some computer rack supplies offer sequenced start up for similar reasons.

 

[Lucien Nunes]

Marconi and I seem to have slightly different interpretations of the description, perhaps the OP can clarify: Notwithstanding the comment:

still these two RCBOs that trip

I understood the problem to be that the overall RCD feeding the shoreline in the workshop trips, if these RCBOs are connected to loads. Do the RCBOs also trip?

 

[marconi]

Seriously - I like a bit of ambiguity in a problem description - adds to the fun

PC1996 - I don't have one - but don't some of the Fluke DMM have the peak or min capture modes for this type of investigation?

 

[Strima]

What happens when you bypass the contactors and link the phases directly for single phase?

 

[pc1966]

PC1996 - I don't have one - but don't some of the Fluke DMM have the peak or min capture modes for this type of investigation?

Quite a lot of meters have peak-hold and similar. But they often have an underlying response time in the tens or hundreds of ms as part of the usual noise filtering (i.e. rejection of AC ripple on DC, etc). So if it is a surge like motor starting over several AC cycles they would pick up on most of it, but I'm not so sure about a half-cycle charging of filter caps, etc, where they might underestimate by a significant degree.

I am sure you get kit for that job, but it is easier to had a sales feature to an existing measurement system than to engineer in fast capture capabilities!
[automerge]1589969112[/automerge]
Another thought I had to reduce the switch-on surge was to use a set of three solid state relays (on an adequate heat sink!) that have zero-crossing switching.

The first SSR is switched by the incoming single phase supply (so L1 load only comes on at the next zero-crossing of the supply), L2 is switched by L1 load coming on, and L3 is switched by L2 load coming on.

Unfortunately I don't have any to play with, but my hope is you would get the three-phase load coming on as:

L1 - (10ms delay) - L2 - (10ms delay) - L3

Also as they are not coming on at a voltage peak the dV/dt is less so hopefully reduced surge on any filter capacitance L-E.

 

[static zap]

Am I interpreting this right ?
a) 1 Phase present
b) Tripping happens if RCBOS on non present phases are left Active. (ie turned on).
c).. Feels like leakeage at some notional Unbalanced wye point ..
d) Is the correct action , to leave the the two non present breakers off.
e) (an automatic fix - may be possible - But it will basically achieve the same as leaving 2 RCBOs turned off)

 

[marconi]

-MAtt: Good morning Sir. This was an interesting problem. I just wondered if there was any more you could say on it since you last posted? I think we would all like to know what racing machine it is for too? Pictures are always good

M