Strange things in the early hours.

[Muhendis]

Some six months or so ago, I fitted a new LED ring shaped light to the landing ceiling as a replacement for the existing round fluorescent tube.

Last night around 02:00 it started flashing randomly about 20 times per second and never stopped. The flashes were very short and not bright but in the dark of night, very noticeable.

The light went on and off OK from the light switch but when supposedly off, the flashing kept on unabated.

In the light of day you needed to look carefully to see it still flashing.

Question is: Where did it get its flashing power from. The switch is wired on the live side and besides the fault cropped up after six months problem free service.

My only thought was that whichever internal component failed, it must have left the circuit in high enough impedance to pick up some energy from other nearby wiring and the internal power supply would charge up to a high enough level to trigger a discharge through the LEDs.

Any other Ideas?

[Crossy]

Hmmm, the usual reason for this type of thing is a switched neutral, you've not had any work done or a new meter recently have you? Can you verify that the polarity is indeed still correct just in case.

 

It's possible that something has been nibbling on your wiring and cause a High-Z leak, maybe some damp has got in with a similar result.

 

[Muhendis]

45 minutes ago, Crossy said:

Hmmm, the usual reason for this type of thing is a switched neutral, you've not had any work done or a new meter recently have you? Can you verify that the polarity is indeed still correct just in case.

Those were my first thoughts also but the switch is definitely in the live side. New meter was fitted over a year ago now and was absolutely correct. Would've been a bit of a bang if that had been wrong neutral being grounded earth as it is.

No evidence of water intrusion. I will do a bit more investigation tomorrow and see if the problem occurs with the light running on the bench. That'll be safer than climbing into the loft looking for rats.

Live is definitely where it should be.

Edited December 2, 2021 by Muhendis
new information

[talahtnut]

Failing all else you could fit a new switch, sometimes

they can give funny problems with leds.

[BritManToo]

41 minutes ago, Muhendis said:

Those were my first thoughts also but the switch is definitely in the live side. New meter was fitted over a year ago now and was absolutely correct. Would've been a bit of a bang if that had been wrong neutral being grounded earth as it is.

Nah, they wired my house with the wires from the meter reversed ...... no bang, but you have live wires all the time throughout your house no matter where the switches/fuse box switches are positioned.

Easiest way to diagnose is with the 20bht test screwdriver, touch it on the wires and see if they are live or not.

 

https://www.lazada.co.th/products/meta-100-500-v-ac-testing-screwdriver-i3019208762-s11101754832.html

 

I always touch every wire with one, when I'm working on the wiring, no point risking death.

Edited December 2, 2021 by BritManToo

[talahtnut]

Crafty Thai wiring dicovered in the missus's crib, picking up

strange x ray signals from Sagittarius A.

Do not be surprised at anything....

[bbko]

On rare occasions, maybe once every 6 months, I'll notice the round fluorescent light in one of the bedrooms flickering even when the lighted wall switch is in the off position. Lighted wall switch is dimly lit in the dark so it's easier to see at night, I just figured it's some kind of minute voltage leaking.

 

OP what kind of wall on/off switch do you have? 

Edited December 2, 2021 by bbko

[Led Lolly Yellow Lolly]

4 hours ago, Muhendis said:

Those were my first thoughts also but the switch is definitely in the live side.

 

4 hours ago, Muhendis said:

New meter was fitted over a year ago now and was absolutely correct

 

How are you verifying this? Flashing indicates flow of power. It's possible the neutral has been switched all along but a component on the board is now failing, disclosing the presence of volts. It's also often a gecko getting itself across a solder joint. The crossed wires could be anywhere, the could be on the meter outside, in your CU, in a box somewhere, somewhere in the roof, just anywhere. . .

 

The easiest way to verify is with a meter.

Neutral bar in CU to ground bar should be 0 volts (or close to zero depending on ground system employed)

Phase to ground bar OR neutral bar should be somewhere between 210 to 240 volts.

 

You can do the same test with your sockets inside. If you're careful you can do this test with light fittings.

Any deviation from the above you have a problem. Everything you describe points to a switched neutral, there isn't really any other explanation.

Edited December 2, 2021 by Led Lolly Yellow Lolly

[Led Lolly Yellow Lolly]

. . . and is this luminaire grounded? (either with a cable or unintentionally through screw fittings). . . You might also want to verify the integrity of your ground cable and ground rod. If the N-E side has somehow started hovering off the ground, this might cause what you are seeing. What kind of ground system are you using?

Edited December 2, 2021 by Led Lolly Yellow Lolly

[A1Str8]

It's a negatively charged poltergeist. You need to block the overflowing impedator. It's usually installed inside the alternating current deposit box. 

[bankruatsteve]

While not with the circular lamp, I have had similar problem with tube in aquarium when the switch gets rain on it (moisture inside) and with a bulb when it is near death.

[Crossy]

8 hours ago, A1Str8 said:

It's a negatively charged poltergeist. You need to block the overflowing impedator. It's usually installed inside the alternating current deposit box. 

 

Who ya gonna call??

 

Super Boondee of course.

 

https://www.nzherald.co.nz/technology/who-you-gonna-call-ghostbusting-gadget-launched/IKYJMXH677URUAQU54DUA6MSDM/

 

[Muhendis]

8 minutes ago, Crossy said:

Who ya gonna call??

TV Boast Gusters

[Muhendis]

OK so here is more on the problem.

Power off ≈ 18vac on the open circuit wires (light not in place).

Power off with 1kΩ fitted = 0v.

Likely a bit of leakage somewhere. 

The light fitting has an MT7813 ic which works at about 12v and peripheral components all high impedance.

The hunt is on, but later. Got to spend some money on car service today.

[Led Lolly Yellow Lolly]

On 12/2/2021 at 6:52 PM, BritManToo said:

Nah, they wired my house with the wires from the meter reversed ...... no bang, but you have live wires all the time throughout your house no matter where the switches/fuse box switches are positioned

Coincidentally I've just been dealing with reversed wires on a house we renovated. The new cabling was completed today and I threw the fuses in. within two or three minutes the RCBO tripped. Serious damage to internal cabling. . .

 

I corrected the polarity but the RCBO was still tripping. The only way to stop it tripping was to remove the neutral bus bar connection. Now, you'd think the ground cable would take care of shunting the current to the ground rod. That's not what happened here, even though the ground rod is three metres long and heavy caliber. The power actually appears to have shunted through a 2.5 mm cable to an aircon compressor case on the wall outside. The only thing I can think that would have caused that is the compressor case is making good contact with the building rebar through a wall anchor and this ufer ground provides a path less resistant than the actual ground rod ( @Crossy ???). . .

 

 

 

 

Edited December 3, 2021 by Led Lolly Yellow Lolly

[Led Lolly Yellow Lolly]

...... Fried all the way to the CU...

 

 

[Muhendis]

1 hour ago, Led Lolly Yellow Lolly said:

Coincidentally I've just been dealing with reversed wires on a house we renovated. The new cabling was completed today and I threw the fuses in. within two or three minutes the RCBO tripped. Serious damage to internal cabling. . .

 

I corrected the polarity but the RCBO was still tripping. The only way to stop it tripping was to remove the neutral bus bar connection. Now, you'd think the ground cable would take care of shunting the current to the ground rod. That's not what happened here, even though the ground rod is three metres long and heavy caliber. The power actually appears to have shunted through a 2.5 mm cable to an aircon compressor case on the wall outside. The only thing I can think that would have caused that is the compressor case is making good contact with the building rebar through a wall anchor and this ufer ground provides a path less resistant than the actual ground rod ( @Crossy ???). . .

 

 

 

 

 

Ooops!

[LoeiI]

15 hours ago, Muhendis said:

TV Boast Gusters

Sound a bit like the bloke who slurred to his Mrs "I've just been the toilet right and when i went in the light came on on its own and when i came out the light went off on its own do think I've got superpowers ?" she said "No you've Pi$$ed in the fridge again" ????  

[sometimewoodworker]

11 hours ago, Led Lolly Yellow Lolly said:

anchor and this ufer ground provides a path less resistant than the actual ground rod ( @Crossy ???). . .

That is probably the situation, along with the ground rod, though long and strong, being in dry soil. FWIW @Crossy has mentioned that his house rebar has lower resistance than his (placed in permanently damp ground) earth rod.

Edited December 4, 2021 by sometimewoodworker

[alanrchase]

This may help.

 

 

 

[easydoor]

Easy to solve the 'problem'....

Thailand works with a line (L) and Neutral (N).

Just have to change, (inverse) those two wires at the light and look you still have or not. 

[Gsxrnz]

Interesting thread. 

 

I've had similar problems to all those mentioned - and while my brain tells me that science is a universal fact, my heart tells me that electrickery in Thailand somehow manages to conform to non-scientific principles.

[millymoopoo]

Probably just picking up a little EMF (electro magnetic interference), LEDs are very very efficient, requiring very small amounts of electricity to light up.

Take a standard  fluorescent tube (old type, not LED) and stand under one of those high voltage tower lines at night, and the tube lights up, no wires connected. same at radio transmission sites, it's all the 'loose' electrical energy floating around.

Could well be a similar thing to that but obviously on a much smaller scale.!

[ballpoint]

17 minutes ago, millymoopoo said:

Probably just picking up a little EMF (electro magnetic interference), LEDs are very very efficient, requiring very small amounts of electricity to light up.

Take a standard  fluorescent tube (old type, not LED) and stand under one of those high voltage tower lines at night, and the tube lights up, no wires connected. same at radio transmission sites, it's all the 'loose' electrical energy floating around.

Could well be a similar thing to that but obviously on a much smaller scale.!

Yes. I take my wife to a high voltage transmission tower every christmas.

it's quite a long trip, but well worth it to see her eyes light up.

[bankruatsteve]

51 minutes ago, easydoor said:

Easy to solve the 'problem'....

Thailand works with a line (L) and Neutral (N).

Just have to change, (inverse) those two wires at the light and look you still have or not. 

What difference would that make if N instead of L is being switched?

[Skeptic7]

????

"phii"

 

 

 

Edited December 8, 2021 by Skeptic7

[sometimewoodworker]

1 hour ago, easydoor said:

Easy to solve the 'problem'....

Thailand works with a line (L) and Neutral (N).

Just have to change, (inverse) those two wires at the light and look you still have or not. 

That will not necessarily solve the problem. The problem is a poorly designed power supply.

[Muhendis]

2 hours ago, bankruatsteve said:

What difference would that make if N instead of L is being switched?

If Neutral is switched, the live is always connected to one side of the light. 

Neutral is earthed so there could be a current path of some sort bypassing the switch and making the lamp glow at reduced brightness.

Apart from that it's a potential hazard for someone.

Edited December 8, 2021 by Muhendis
new information

[khunPer]

On 12/2/2021 at 11:19 AM, Muhendis said:

Any other Ideas?

If you have a tiny LED indicator in the on/off switch it can make a LED bulb flash of glow at low brightness.

 

If you use induction from high power lines - or even some radio waves close to a transmitter - you might have free power for your lighting...

[bankruatsteve]

1 hour ago, Muhendis said:

If Neutral is switched, the live is always connected to one side of the light. 

That is why I made the comment - to the thought that switching the wires at the light would somehow fix the problem.