Sub-second power outages only on some circuits

Talk Electrician Forum

Help Support Talk Electrician Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

tim.fmf

Member
Joined
Sep 13, 2018
Messages
9
Reaction score
0
Hi Everyone

I work in a building in Hammersmith in London which has several circuits in it where each circuit has its own trip switch in the fuse box.

Approximately three or four times a month, we get a sub-second power outage which only effects some of the circuits in the building and it does not cause any of the switches in the fuse box to trip. This causes our server and another computer on the same circuit to restart but the other computers in the building and the lights remain switched on.

Does anyone here know what could be causing these sub-second power outages and what can be done to stop them?

Thank you very much.
 

 
Might be worth having a look with a heat camera in the distribution boards for any hot spots / loose connections.

That and periodically checking everything is in fact tight! Do on a shutdown etc.

 
probably a loose connection somewhere
Thank you, Andy,  very much for your reply.

To be honest, I might be wrong but I am doubtful if it is a loose connection because, as I mentioned, it only happens three or four times a month and I don't know what would vibrate the building enough as seldom as that. The building is nowhere near an earthquake zone. The reason I think this is relevant is because as I understand it, a loose connection requires shaking in order to become temporarily disconnected.

Might be worth having a look with a heat camera in the distribution boards for any hot spots / loose connections.
Thank you, Onoff, very much for your reply.

Having said what I said to Andy, you might be right in suggesting looking for hot spots.

One idea I have thought of is that it may be the case that different circuits in the building react differently to spikes of electricity from the supplier. Is this a possibility?

Does anyone have any other ideas?

 
Only additional idea is that knowing which circuits suffer and which don't, coupled with knowledge of the building distribution system, should greatly narrow the field for a more in depth check.

I too think that a thermal camera could prove valuable.

 
All joking aside I've known mechanical fixings loosen over time due the building being sited over a Tube line. 

Electrical connections get stressed under load so go "hot & cold". This leads to cycles of expansion & contraction. Don't forget screw terminals aren't considered maintenance free.

Just make sure everything's tight (but not too tight), theoretically to the right torque according to the MIs.

 
All joking aside I've known mechanical fixings loosen over time due the building being sited over a Tube line. 

Electrical connections get stressed under load so go "hot & cold". This leads to cycles of expansion & contraction. Don't forget screw terminals aren't considered maintenance free.

Just make sure everything's tight (but not too tight), theoretically to the right torque according to the MIs.
Hi Onoff

Thank you very much for your informative reply.

 
To be honest, I might be wrong but I am doubtful if it is a loose connection because, as I mentioned, it only happens three or four times a month and I don't know what would vibrate the building enough as seldom as that. The building is nowhere near an earthquake zone. The reason I think this is relevant is because as I understand it, a loose connection requires shaking in order to become temporarily disconnected.


You are incorrect thinking a loose connection requires vibration to cause problems. Electricity always generates some heat as it passes through a circuit. Higher power means greater heat and loose connections also increase heat. As power increases and decreases during normal usage, the joints will be heating and cooling over time. A termination that has never been correctly tightened will appear to work fine for a period of time, then start to deteriorate as arcing and sparking across the joint reduces the integrity of the electrical connection. Sensitive low power devices like IT equipment may be the first to detect such deterioration compared to say lights appearing to dim or flicker.  There could be a poor joint, (or multiple joints), at an accessory,  a junction box, at the MCB inside the CU. I would be starting with verifying the CU  connections and continuity integrity of the whole circuit(s) that are experiencing problems. It could also be a switch contact inside an accessory. Note: It is not uncommon to find the result of poor quality installation work showing its head a few years after something was rewired. A product of sub-contracting cheap labour on tight time deadlines to install some wiring.

Doc H. 

 
You are incorrect thinking a loose connection requires vibration to cause problems. Electricity always generates some heat as it passes through a circuit. Higher power means greater heat and loose connections also increase heat. As power increases and decreases during normal usage, the joints will be heating and cooling over time. A termination that has never been correctly tightened will appear to work fine for a period of time, then start to deteriorate as arcing and sparking across the joint reduces the integrity of the electrical connection. Sensitive low power devices like IT equipment may be the first to detect such deterioration compared to say lights appearing to dim or flicker.  There could be a poor joint, (or multiple joints), at an accessory,  a junction box, at the MCB inside the CU. I would be starting with verifying the CU  connections and continuity integrity of the whole circuit(s) that are experiencing problems. It could also be a switch contact inside an accessory. Note: It is not uncommon to find the result of poor quality installation work showing its head a few years after something was rewired. A product of sub-contracting cheap labour on tight time deadlines to install some wiring.

Doc H. 
Hi Doc H

Thank you very much for your very informative reply.

Please excuse this newbie question but what do MCB and CU stand for?

One thing to note is that the server and the other computer on the same circuit are always effected at exactly the same time. It is as if the whole circuit is being completely switched off and on again momentarily. I have also been informed by a colleague that the lights sometimes (but not always) go off and on when this happens. (Last time this did not happen.) Does this extra information provide any further clues as to what may be going on?

 
Is there anything else in the building scheduled to happen at those times of the month, generator test etc?
No there isn't. When I say it is three or four times a month, it is actually somewhat random.

 
Last edited by a moderator:
What about any heavy industry near by that could be starting say a big compressor or pump which may cause the voltage and frequency to momentarily dip, this dip is getting picked up by sensitive kit like IT?  Does it coincide with any external influences like high rain fall (so pumps have to start) or high or low ambient temperature spikes relating to HVAC systems having to work hard.

But as above I would be getting the FLIR camera out and looking for switchboard hotspots.

 
Last edited by a moderator:
Hi Doc H

Thank you very much for your very informative reply.

Please excuse this newbie question but what do MCB and CU stand for?

One thing to note is that the server and the other computer on the same circuit are always effected at exactly the same time. It is as if the whole circuit is being completely switched off and on again momentarily. I have also been informed by a colleague that the lights sometimes (but not always) go off and on when this happens. (Last time this did not happen.) Does this extra information provide any further clues as to what may be going on?


MCB:  Miniature Circuit Breaker,  the modern day switch type replacement for traditional wired fuses.  (fits into a CU). 

CU:  Consumer unit or Distribution board where all of the circuit cables connected to the MCB's to obtain electrical power for the circuit.

The items that you called the switches in the fuse-box. 

The extra information confirms what I previously stated, you need to get the whole circuit checked, mechanical terminations and any thermal checks.  As a further related thought, how critical are your servers? and what provision do you have in the event of a genuine power failure?  A UPS could be a useful addition if power fluctuations are an external problem. https://en.wikipedia.org/wiki/Uninterruptible_power_supply 

Doc H.

 
Last edited by a moderator:
Is it always the same circuits go out? It could be an issue in an HV or LV switch room on a particular phase. You really don't want to be messing in there.

Is there not a facilities maintenance team on site who investigate/deal with this? 

Has the building ever had a periodic inspection? Often done on a "shutdown" over say a bank holiday weekend. 

 
probably a loose connection somewhere


Concur ^^


I think the first two post point to the most probable cause...

And hence the first course of action, investigation needed...

Until all circuit connections have been verified you are, as the phrase goes..   "pi55ing in the wind" ...

trying to second guess other possible causes.

Guinness  

 
Hi Everyone

Thank you very much indeed for all your replies so far. It therefore looks as if we will need to get our wiring investigated. In the meantime, I have another question:

How long can a loose connection exist for before there is a permanent failure? Can it exist without getting worse for more then two years? (I am asking this because that is what seems to have happened in this case.)

 
Top