Type A or AC RCCD for a PC computer PSU?

[Mike Lee]

Hi, I have a small problem of a type AC Disjoncteur differential (I believe it is an RCCD in english parlance) tripping occasionally, it happens about twice a month. I believe it to be caused by my computer power supply, I have come to this assertion due to it being the only thing switched on when the RCCD trips....although I could be wrong! My question to those much more enlightened than I is should I put the circuit that the computer is on into a type A RCCD? would that reduce or even remove the tripping problem or should I be looking at spending more on a PC PSU? I have, in the past had numerous tingles from PC power supplies (which are all switched mode) being employed in the computer hardware repair business. Your comments would be most welcome.

 

[Bobster]

What is the mA rating of the RCCD you are using?

Is there anything else on this circuit?

Computer PSUs naturally have quite a high amount of current leakage to earth. 3-5mA is what I have experienced. I know that 10 and even 5mA RCCDs are used in France. So this may be the reason.

 

[Mike Lee]

Thanks for your reply Rob, it is a 30mA RCCD, I thought they were all 30mA...certainly the norm. There is quite a lot of stuff hanging off the wiring in my mancave but almost all of it is either off or in standby mode - however - the rccd trips occasionally. I thought the theory behind an A type rccd was that it could deal with a device chopping the sinusoidal mains like a washing machine speed control or an electric cooker hob and I was wondering if a switch mode psu might be put in the same category. If the general feeling is that it will make no difference I could either try replacing the rccd or the computer psu..both costing enough to want to make sure doing so would bring about the favoured results.....

 

[Bobster]

Your correct in some sense, that they deal with an imperfect sinusoidal waveform. However, this problem doesn't cause them to trip, the opposite in fact. It can saturate the device and prevent it from tripping even during a fault.

Even in standby mode there will be normal earth leakage from things like PSU's.

A 30mA RCCD can trip at as little as 23mA. My guess would be you have a number of items with natural leakage and it's just on the threshold of the RCCB.

Is there anyway of splitting the devices you have powered onto a different RCCD? Or disconnecting a few and seeing if the problem still stands.

There is also the fact that there could be a genuine intermittent fault somewhere.

 

[Mike Lee]

Thanks again Rob, There is only one feed into my mancave, and I am mindful of the fact that I have a 3 phase supply to the house and would not want to risk running an extension feed from another part of the house in case I inadvertently select one from a different phase.. .
I think I will have to just disconnect stuff and, as the tripping occurs so infrequently, leave it disconnected for quite some time to try to prove the fault - at least until I get bored with that and cough up for a new rccd to try.... actually... I could swap the two type AC's I have on the plateau (distribution board)
There's no chance of a genuine intermittent fault somewhere..I did the re-wiring!!

 

[static zap]

....I have, in the past had numerous tingles from PC power supplies....

Since all PC DC rails in supplies are referenced to the earth case ,
a) I would PAT test some of your EURO leads , or do some simple earth continuity tests .
(Unless you still have plenty of CRT monitors and a dry enviroment! )

b) ..and maybe wave a "Magic-Lies" volt stick about for any metal hardware with high leakage potentials from trapped wiring ...

 

[Mike Lee]

Hi Static, thanks for your input. My comment about receiving tingles from PC cases was in the past from my days as a hardware engineer and to be sure there were then instances of the building earthing being 'sub-optimal' to the extent that the network cards in the PC's would not work in some rooms which was traced to the neutral floating at 30 odd volts to earth and yes.. it was a revelation to me. I do not think it applies to my current problem (pun intended) as I have never detected 'tingles' here in my man cave so I am led to believe my earthing to be good and the problem lies with the leakage and not the earthing. I fear the simple continuity testing of my EURO leads will I fear reveal nothing as the tripping is quite infrequent and if one or more of the leads were faulty would that not cause continual problems? I do not, unfortunately possess a PAT tester. No CRTs on site but I do run a 4 LED monitor flight sim (but not on the PC I suspect of causing the tripping). I had not tried a volt stick, so I did so...I had a aliexpress special lying around and I find the case of the errant PC indicates 36v on my old and cruddy BUSE(?) (I cant read it properly as the legend has worn off!). I tried the same test on my Flight Sim PC and that reads zero on the psu case......I think I may just have my culprit Thank you static....looks like a new order is coming CCL computers way Which is good because I don't now have to disturb the wiring in my distribution board. Thanks again.
As an addendum I should perhaps mention that I have performed continuity checks between the PC PSU ground and other devices grounds and the heating copper piping in the room so I am confident the earth to the computer power supply is sound. Also it may be interesting to note that my house boiler is on the same rccd as my mancave and I have had a 230/24v transformer fail in my boiler TWICE following rccd trips, once the primary was open circuit and once the secondary was open circuit , now this may be un-related... but TWICE? I am of a mind to move my mancave onto a different rccd as well as replace the computer PSU

 

[Bobster]

So you've not only got the man cave on the RCCD but part of the house too?

 

[Mike Lee]

Indeed...it cost a bleeding fortune as it was...I couldn't afford any more disjoncteurs differentials or a bigger board. Luckily whenever I am in my man cave the rest of the house is usually empty so I can keep the wheel in the meter from tipping the house on its side due to centrifugal forces... on really sunny days I can even get the meter running backwards with the solar input....thats really gratifying even if it will never turn me a profit.

 

[static zap]

...I had a aliexpress special lying around and I find the case of the errant PC indicates 36v on my old and cruddy BUSE(?)....

I'm hoping we have found your issue ,
though I have found a few mobile phone chargers guilty of 50V with a similar technique.
( an open circuit voltage with little current )
{Laptop PSU's have previously been"figure 8"--(50/50-if-unplugged there!)--> to earthed "clover leaf"-mostly}

 

[Mike Lee]

Ouch!.. the Chinese have never been too bothered about safe electrical design on the cheap as chips knock off chargers have they..

I have ordered a new computer PSU and will update this thread after it arrives and I have installed it.

 

[Lucien Nunes]

I find the case of the errant PC indicates 36v

I have performed continuity checks between the PC PSU ground and other devices grounds and the heating copper piping in the room so I am confident the earth to the computer power supply is sound

These two are contradictory. When you say the case 'indicates 36V' I would infer 'with respect to earth.' But then you say it is well connected to earth, in which case the voltage between them will be negligible, a fraction of a volt.

If a power supply that is supposed to be earthed is not, its case will float up to a voltage somewhere between line and neutral, often around 110V w.r.t. earth due to the symmetry of the suppression capacitors. Having 36V w.r.t. earth present does not in any way indicate that it is faulty, merely that it is not earthed. I am not saying this PSU is OK and not to blame for the tripping, rather that it is not a valid conclusion that the voltage indicates a PSU fault.

When you measured the 36V, where were the two meter probes?

 

[Mike Lee]

Hi Hucien, I was not using a meter, the 36v is indicated in the display of a volt stick - with the pointer end on the psu case and my finger on the capacitively coupled pad on the top. I am reasonably sure of the earthing of the psu through the mains wiring as I have done a continuity test using a test meter between the psu case and
the mains leads earth socket (EU plug) - and -
to the copper pipework going to the rooms radiator, which I know is earth bonded back to the distribution board - and -
to the shielding around my network switchs' rj45 connectors - and -
to the earth connector on my scopes front panel.
All were less than one ohm.
So, whilst I agree that the two statements I made appear contradictory I assumed it is because I do not fully understand the nature of how a volt stick works with its capacitive/inductive coupling.....
I have a second PC in the man cave, the volt stick measured nothing on the case of that PC.
Thinking a bit more about it, it could be possible that the leakage (if there is leakage) could be from another component in the PC case, perhaps I should isolate the PSU from the case and see if the 36v is still there with the volt stick but to do that I have to post this first....

 

[sparksfly]

Insulation resistance test your leads, I identified an issue whereby customer would experience RCD tripping every 3 days. It was a combination of loose neutral wiring accessory connection and very bad IR reading on a PC lead.

 

[Lucien Nunes]

could be possible that the leakage (if there is leakage) could be from another component in the PC case,

No. If the case reads live w.r.t. earth, it's not solidly connected to earth. By default, the PSU will leak enough to read on a meter, directly through its case or via anything connected to the PSU which will have an earthed DC common rail.

Voltsticks have their uses but can give confusing indications unless you know exactly how they work. Any voltage capacitively coupled to the tip is compared with the voltage on the touch pad. You're big and fairly conductive, so you capacitively couple the touch pad to your surroundings quite well. These are assumed to be earthy, so it's comparing the voltage near the tip with earth. If you are not close to earth potential, the voltstick has no reference and it may be hard to interpret the indication without a reasonable knowledge of physics and electronics.

For example, suppose you stand near a fridge that is not earthed. Normal leakage in the fridge may raise the casing to say 100V, but the current is so small you probably won't get a tingle, so you're not aware it's not earthed. The floating fridge couples capacitively to you and if the capacitance from fridge to you is half that from you to earth, you'll end up at 33V. Again, very high impedance, tiny currents, you won't know anything about it. But when you point your voltstick at an appliance that is earthed, it sees the 33V difference between you and that, and lights up. It has no way of knowing which end of itself is earthy.

So I think you should ignore what the voltstick says at this stage although it might be trying to tell you something about your surroundings. Many regulars on this forum detest voltsticks because of this ambiguity, and the tendency for non-electricians to use them for safety critical purposes which should never be done. Search for your leakage using calibrated instruments!