Hi folks,

First a quick "hello" since I'm a long-time lurker but this is my first post. I'm not a professional EC but do have a degree in Electrical and Electronic Engineering (wow, dangerous!). I've been fascinated by electricity for as long as I know, and having my Dad working at a power station during my formative years certainly helped light the fire. I progressed to electronics and then to IT (boo, hiss!) At the moment I'm working for what remains of the UK computer industry, formerly ICL, where we still make a meagre income from Mainframe computers. This won't last forever though, and a change of career may be on the cards...

Ok, to my question. Over the past couple of days the lights here at home have been flickering briefly, like they sometimes do in a thunderstorm, sometimes several times in succession a few seconds apart, other times not for a while. Working on the computer yesterday I noticed that the UPS was switching to battery for a few seconds each time too, and decided to try and find out what was going on, since this is most unusual and we're in the middle of town (Manchester, UK) with no local overhead wiring.

Finally managed to capture a blip on the 'scope:



The blue top trace is the secondary of a small LV transformer, the flat tops are probably from all the computer gear running on this ring!

Looks like something is arcing over on the peak, then clearing by the next cycle, or could it be HV switchgear somewhere? Could it be a local fault (no sign of smoke coming from the panel!)

I'm getting slightly worried by it, since it's been happening for two days now. Any suggestions?

Ian.

hi jusme!
welcome to the ecn.

how old is your buildings electrical installation?

im not sure about the peak of the wave, it could be a local fault too imho. did you do more logs and always got the spike at the top of the sine wave?

PS - looks like a nice, damn expensive scope

[This message has been edited by :andy: (edited 06-21-2004).]

The center wave reminds me of a megger test on bad insulation, for what that's worth.

I'm going to take one stab at this... How old is your refridgerator? This could be a hard starting compressor, and if your not getting this wave form in other parts of the house, on the same phase, it would be within your house. Is it only affecting this one circuit? Track the freqency of the event, see if it co-incides with other electrical operations, of other eguipment.

Thanks for your comments. What I'd really like to know is whether it's an external problem (so I can just wait for something at the local substation to go pop!) or should I be getting the wiring inspected?

It seems to have stopped happening now - nothing in the last hour at least. Still niggling as to the cause though.

:andy:


It's "fairly" modern (rewired sometime in the 80's) and no obvious candidates for the photo gallery. No recent alterations either.


Yes, it always seems to be at the peak (either polarity), and always clears by the next cycle. Here's a couple more occurrences from last night:





(The red trace is just the trigger btw).

This is a graph from the UPS that shows when it's happening:

http://www.jusme.com/stuff/glitchups.gif (Not inlined because it's a wide image)

The downwards blips on the red (Line Min) line correspond to the glitches, also the (hard to see, sorry!) dips on the yellow line indicate each time it briefly switched to battery - normally this line is totally flat. Seems to start in anger around 5:30pm then continue through the evening, becoming infrequent (but not disappearing) through the night.

(Interestingly the line voltage here seems to be on the high side - this has always been the case and why the blue line, UPS out, keeps flipping between line and line-30V - "auto-trim" or something!)



Good call - a few years old (3-5 I think) not ancient. I'll try turning it off for a short period if the problem starts up again to see if that's it.



All circuits are on the same phase, it was affecting the lighting and the "office" ring at least (separate breakers). There were no other heavy electrical loads on in the house last night when it was happening frequently (no AC, cookers or heating) and it was happening while we were eating when all the office equipment apart from UPS and one PC was off. I'm wondering if it could be the UPS itself - switching to battery and glitching the mains when it does... ouch!

I'm trying to find out if the neighbours were seeing the same, but if this was a fault on only one phase am I right in thinking that the immediate neighbours probably wouldn't see it?

Ian.

Ian — Is the blue trace voltage upstream or downstream of the UPS? What is driving the red trace?

Bjarney:


The blue trace is the output of an ordinary low-voltage (9V) transformer plugged in upstream (before) the UPS.

The red trace is the same signal fed through a C-R differentiator to generate a trigger pulse for the 'scope on the fast edge caused by the blip on the AC.

(The transformer is just for isolation, and isn't powering anything btw.)

The lights flicker, the UPS switches to battery and the 'scope triggers simultaneously!

Ian.

Hi Ian and welcome to the forum.

Good set-up with the RC differentiator to trigger the 'scope.

Another thought that crosses my mind with the glitch always taking place on peaks is the possibility of an MOV being faulty.

Ok back from work now...
And if I weren't trying to buy a house, and save money, I'd be at the the bar now!

Anyway, I find it facinating... So The isolation transformer, although making an easier and safer voltage to read, may be obscuring other harmonic distortion in your reading on the scope. Changing the range slightly, you may still be able to see them in a half cycle. likewise zooming out, you may be able to notice longer duration voltage fluxuations, where the zero point may change.

This is load effect, (taht's a no brainer) but I think, exagerated due to a poor, or resistive connection. And it could be anywhere from that outlet, all the way back to the transformer. In last ditch effort to find those, I brake out my favorite geek toy. And, only on special occasions. A TDR, you may even have that cabability on your scope? Unfortunatly, I'm not familiar with the UK ring wirng. With a TDR you could read through the terminations all the way back as far as you can, on a dead circuit. And spot a bad one that way. But with the ring issue, you would have to break it in half, go one way then the other. Where as you're not an electrician, although obviously well edjucated, I suggest you call one for that type of thing. If it comes down to it you might have to call one anyway.

Back to the load issue... Recently, I have seen a spike simular to this! You mentioned other computer equipment being on this ring. I had a AFCI that would trip every day. And, after a buch of different searches though the wiring, I broke down and brought in the scope. As I was hooking it up, it happened, and I didn't need the scope to figure it out. Turns out to be his new flat screen monitor. When it went into "power save", it tripped the AFCI. Some sort of cap discharge when it went off. Something to think about.

Anyway, I hope you find this, and keep us informed.

Well, since lunchtime yesterday there hasn't been a single blip, which still leaves a niggling worry to what it might be and very little chance of locating it by process of elimination. I hate it when faults just disappear without knowing why - they usually come back to bite :-(



I didn't spot anything abnormal on a wider view, though I wasn't really looking much closer.



Very plausible, could be fun to find!


Thanks for all your suggestions, if it starts up again I'll have to do some methodical faultfinding or get the wiring properly inspected.

Ian.

Hi Jusme, To eleiminate the ring circuit wiring do a simple continuity test.
Disconnect the ring main at the distribution board and measure the resistance of the conductors L-L and N-N the readings obtained should be identical if all connections are sound. Any substantial difference will indicate a poor connection that could be contributing to your problem. That will at least save you searching at each outlet for dodgy connections.
Have you considered that the problem could be on the power co underground service cable. Have known older joints to give problems especially if there has been an alteration. They can arc across for some considerable time before giving up altogether. Had one not so long back where treee rootes had strangled a PME service cable against rubble and pipes in a trench, that was causing lights to flicker on a whole estate. It eventually capitulated and blew a hole in the cable and cleared the sub. Assume that as you have not posted lately that all is O.K. at the moment.
Good luck in your search

Just a quick update. No more flickering lights and no blips been noticed by the UPS for several days now, so I'm hoping the problem is (was) external and has gone away. I'm hoping to get the panel at least visually inspected soon though, just to be safe.

Plenty of things to try if it starts up again though, thanks for the advice.

Ian.

Hi we are having the same sort of thing here. Around 5 pm we get a regular pulsing effect a momentory dip happening several times in succession. It happened last winter too maybe its just that the supply transformer is just about at maximum load

If it is consistently at the same time, I would suspect a large load on a timer starting up. Are you on a dual rate system where the water heating cylinders and electric stoves are on a separately switched power source? If so, does the power switch on or off at 5 PM?

Where I went to college, we routinely saw transients at 7 AM when PF correction caps were switched on.

It mite be some large load switching on although this is a residential area with a few shops so I personaly think that its a stuck tap changer or just or maybe lumpy lectric

I've heard complaints about siginificant dips in one particular region of Austria at exactly 10 PM. They're fierce enough to confuse various electronics, e.g. EIB lighting controls and even set off alarms. It turned out that the power company switches the entire distribution to night rate (separate meter) at that time and that's what causes the brownout.

Thanks for all your reply on this I'm sure otd some sort of network switching that causes it. Although why it would happen 5 or 6 times in a row is a mystery. How many goes does it take to do a bit of switching