Ghost fixes lightbulb
 

Every once in a while something on my car will fix itself. That seems to have happened today.

For years I've been waiting to fix a left front turn signal problem. So today I have the bumper cover off and the light fixtures out to possibly replace the bulb. Inside the Civic's side marker light bulb are 2 filaments. With the light bulb plugged in and out of the headlight fixture I tested it again. The directional filament would not light up under any conditions and caused the parking light filament to intermittently shut off whenever I turned on the left turn signal. Hazards also worked improperly.

I did a "known good" test by swapping the left front turn signal bulb with the right front turn signal bulb. I didn't do anything else. I didn't clean anything or detach and reattach anything other than the bulbs. Suddenly both light bulbs worked normally. Then I put them back in their original places. And the bulb that has been failing consistently for two or three years of daily diving started working completely normally.

I am guessing that in removing the bulb I cleared some kind of minor corrosion or dust/dirt (the bulb housing was coated in fine road dust).

I suppose it might also be an intermittent open in the wiring.

Or a ghost in the machine.

http://ecomodder.com/forum/attachmen...1&d=1521416899

Her real name is Christine.

:)

"Fear on four wheels"

https://www.youtube.com/watch?v=9aU5l2e9YlQ

This the symptom of a bad ground.
-m

I wondered about some sort of circuit problem like that when I got the bulb out and could see that it's filaments were intact. But if it is a bad ground, then why did the problem go away without a "fix"? Did remove and reinstalling the bulb somehow clear a grounding issue? I did not clean any connectors. Should I expect the problem to come back if it is a ground issue?

I've seen broken filaments spot weld themselves back together. And then break again. Over and over. Try a new bulb!

I had a turn signal bulb act up intermittently for several days before it failed altogether. Because it was intermittent, it threw me for a loop. Literally wiggling the bulb socket and jiggling the harness could cause it to come back to life. Like a rube, I didn't even check the actual bulb until I couldn't get it to work again at all.

I have found the ground tab inside the socket is failing many times on many different makes of cars. I work on a large Ford fleet now and stock several different socket and connector pigtails. One problem and they get a pigtail, on 88-99 Chevy pickups it was a taillight plate. The plate was basically three sockets mounted to a $40 piece of plastic that the contacts melted and failed. Factory LED lighting has made the newer stuff more reliable.

As with many problems, searching for a common ground brings enlightenment :)

Then, I've had more than one lightbulb with a broken filament swinging about accidentally reattaching that with the other end and burning happily for a while, be it a few milliseconds, a couple or years or anything in between.

LEDs fail much more predictably.

After consideration, I am gonna clean the connector throroughly and use the 1998 bulbs. I will switch the sides they are on so if the failure happens again, I will get new data to disambiguate what is causing it (bulb or connector ground).

My "Christine" (1983 movie, +1 Frank Lee) apparently wants another chance to get me dead or ticketed and has demented my judgement.

Thx!

Microwave fixed LED - and lots of wasted ones
 

Yup, you read it right. I fixed a T10 LED by putting it in a microwave!
Repairing without touching, the microwave fills in the ghostery in this case.

When I got a new laptop at work the sysops told me there was this guy that bought laptops that had a screen failure or other problem but were physically intact. He'd disassemble the screen (or whatever the malfunctioning part was) and put it in a microwave for a few seconds; that often restored its functionality.

Well that sounded too good to be true.
But I had that collection of failed LEDs. What if...?

So I gave it a try.

Here's the before:
T10 crossboard lamp with 24 SMC LED's of which only 3 rows of 3 lit up.
https://i.imgur.com/SFvL563.jpg

Microwaved for about 2 seconds at about 900W, together with an identiical lamp that also failed partly. At 2 seconds it flashed and I killed the microwave within a quarter of a second.

After that, when tested one lamp lit up completely!
https://i.imgur.com/YhRY4Ag.jpg

Success! It works. The world is my oyster!

The other lamp had not improved. Gave it another burst in the microwave.
Sadly, that killed it.
But at least, one of the lamps was definitely repaired.

Spurred on by this success, I tried it out on a collection of misbehaving or completely failed LEDs and killed all but 1 of those.
Two T10 LED towers with ever so tiny SMC's that had faded to a shimmer.
1 second in the microwave restored most of those SMC's to original brightness. But not all.
A second dose than. Spark. Black spots on the LEDs. A writeoff.

One H11 fog light that had only 4 of its 10 high power SMC's lit up - or none at all, depending on the tide. Lit up briefly in the microwave. Restored those 4 LED's to continuous working, but not the rest.
Another dose then. Total fail.

A collection of housing LED's (E14 and E27 plugs, 230 Volts). The microwave either did nothing to them or killed them.

Conclusion:
A short burst in the microwave could restore your dim or failed LED's, but it works just in a minority of cases. Too short and nothing happens. Too long (2 seconds typically does it) and it is dead.
See them light up, then you have a chance if you kill the microwave at that very moment. See a spark, then throw it away; no need to even test it anymore.

In hindsight I might have put some buffer in the microwave with the lamps, like a glass of water.
I did not so I can't tell what effect it would have. My collection of doubtful LEDs is depleted now, so I can't do an encore.

Whatever you do, ventilate well.
Do not increase the time beyond 2 seconds; that will only bake the LEDs (they get black spots).

But it was fun trying :)

That is ghostly!

It is actually pretty simple.
These LEDs and many other electronic components are made in the cheapest possible way; that is, good enough to pass quality control and initial testing.
Little consideration is given to the stresses the repeated heating up and cooling down cause over time.
Hairline cracks appear in the soldered connections and the component fails.

When you put 2 metal objects very close together in a microwave, you'll see a spark between them. The hairline cracks are bridged, re-solder themselves and the component works again.

The problem is that other gaps can be bridged too - if they are not supposed to they'll fail. The PN layer in diodes, transistors and ICs is vulnerable.
Also anything metal quickly heats up by itself - and LEDs are notoriously heat sensitive.

So there's a minefield of possible problems.
But if you have a failed component that you can't or won't repair otherwise, good ventilation and it's not your microwave anyway, why not give it a try?

You can better step it up very gradually indeed or use a buffer like a glass of salted water (to aid conductivity).

Horn died while I was wiring radiator fan on/off. Tested horn itself by connecting directly to battery: loud AF. Tested relay with known good swap: no change. Checked fuse: perfect looking (perfect!).

Drove to LA for work, and on the way home horn started working again.

Thank you, ghost.

Half of the LEDs in the microwaved lamp have since died again :( It ain't much of a fix.
Bet the thermal stresses have no difficulty in causing new breaches.