Log in to the online community

Want to post a reply? You'll need to log in
An interesting problem, loose connections and LEDs.
Question
I wonder how many of you have seen the interesting problem of LED lamps and loose connections?
A friend of my Partner recently moved into a new flat. She replaced all the lamps with LED options and a new and before invisible problem showed up, two of the new LEDs flashed on and off at perhaps 0.5 Hz, sometimes, but noticeably when the CFL outside light was switched on and they were already on.

Changing the LED lamps made no noticeable difference so B&Q got away with that! Fitting filament lamps made the problem go away. I wonder what was happening?
Carefully trying to measure the supply voltage at the ceiling rose gave a clue, when catching the lamps flashing (only sometimes, the rest of the time normal light) indicated about 40-50 volts, and 240 ish when they worked normally. Cable continuity measured OK, there was a fairly complex set of 2 way switches and looping in to disentangle, but removing the switches indicated a number of loose screws, and one connection where two live loop wires entered but only one was caught by the screw. The start current of the CFL was sufficient to make a good contact of this loose wire, and sometimes it connected, but at other times it was several thousand Ohms! With the higher current of normal filament lamps the connection made good contact, with about 50mA of LED it was not reliable.

With about 50V the lamps would start, but then as the smoothing capacitor discharged they had insufficient voltage and went out, the time constant being about 2 seconds. I have now seen this twice, and it is at first look very odd indeed. There are many installations out there with loose connections which work fine with larger loads, but as the LED currents are small the contacts no longer spark slightly and self-clean. AFDDs would be useless, and the situation is not dangerous because of the low current.

 
5 Replies
Zoomup
2916 Posts
davezawadi (David Stone):
I wonder how many of you have seen the interesting problem of LED lamps and loose connections?
A friend of my Partner recently moved into a new flat. She replaced all the lamps with LED options and a new and before invisible problem showed up, two of the new LEDs flashed on and off at perhaps 0.5 Hz, sometimes, but noticeably when the CFL outside light was switched on and they were already on.

Changing the LED lamps made no noticeable difference so B&Q got away with that! Fitting filament lamps made the problem go away. I wonder what was happening?
Carefully trying to measure the supply voltage at the ceiling rose gave a clue, when catching the lamps flashing (only sometimes, the rest of the time normal light) indicated about 40-50 volts, and 240 ish when they worked normally. Cable continuity measured OK, there was a fairly complex set of 2 way switches and looping in to disentangle, but removing the switches indicated a number of loose screws, and one connection where two live loop wires entered but only one was caught by the screw. The start current of the CFL was sufficient to make a good contact of this loose wire, and sometimes it connected, but at other times it was several thousand Ohms! With the higher current of normal filament lamps the connection made good contact, with about 50mA of LED it was not reliable.

With about 50V the lamps would start, but then as the smoothing capacitor discharged they had insufficient voltage and went out, the time constant being about 2 seconds. I have now seen this twice, and it is at first look very odd indeed. There are many installations out there with loose connections which work fine with larger loads, but as the LED currents are small the contacts no longer spark slightly and self-clean. AFDDs would be useless, and the situation is not dangerous because of the low current.

 

Interesting and helpful comments.

Edit, add. I recently had to fault find where lights in three locations in a house worked intermittently or flickered. Three plate wiring. The fault was a loose neutral screw in a hidden batten lampholder in a cupboard.



Thanks,

Z.

mapj1
3555 Posts
Interesting, but not a total surprise as it sort of ties in with the great difficulty to sustain an arc with soft metals like copper, that tend to either be a dry joint, and near open circuit, or with a bit more current to microscopically spot weld to be a good contact. We also see this effect  with low current tests of Zs and so on, that there can be an apparently high resistance connection on the low power meter where the brass is a bit oxidised, but comes a 25A pulse or two and remeasure, it is either blown clear and easy to find as a single thread of flex has been vapourised , or more likely som eoxide has been vanquished and it is all back in spec again.
I have seen circuits for switching audio and for crackling 'phone lines that use a burst of  'contact cleaning current' that are meant to fix faults or make them permanent so they can be found more easily.
M.
AJJewsbury
2580 Posts
I have seen circuits for switching audio and for crackling 'phone lines that use a burst of 'contact cleaning current' that are meant to fix faults or make them permanent so they can be found more easily.
That brings back old memories - I recall a 'track cleaning' circuit for OO-gauge model railways (which suffered a lot from oxide on the surface between the loco wheels and track) - did something like raise the voltage when it saw the current drop.

   - Andy.
mapj1
3555 Posts
The electronic track cleaner is still around though not so useful with the model trains with more complex electronics on board  (electronic track cleaner) as far as I can tell, it does not really clean the tracks  as such, just allows it to run for longer perforating the oxide, until eventually it builds up to the a thickness that  it stops at. Presumably great potential for long and medium wave RF interference as well, unless the rise times are controlled. Anecdotally I have heard of some home brew versions shocking people as well, presumably from the design stable of if a few volts are good, more must be better..

M

Share:

Log in

Want to post a reply? You'll need to log in