Infrared motion sensor, early failure

[jitter]

Recently I started a topic on the expected service life of mains filters. Mainly because I found an end of life MKT X-class cap which can also be found in mains filters.
Now I suspect another failed X-class cap, this time in an infrared motion detector.

Almost two years ago I installed two infrared motion detectors and recently found one was dead, it would no longer switch on when moving into the detection range. I decided to take it apart and measure some voltages. Basically it consists of two separate pcbs. One with the capacitive dropper and relay and the other with the detection circuitry. On the latter was a 78L05 which only got about 6 VDC on the input. I went to measure the other pcb, I found a relay that was supposed to work at 24 VDC (coil voltage), but nowehere near that voltage could be measured, the highest being 6 VDC. No wonder it didn't work anymore.

Before going any further I checked the receipt and the warranty period and found it to be within that period. I went back to the store and they gave me a new one.

I strongly suspect the 390 nF X2 cap in the capacitive dropper circuit had failed. In circuit, it measured nowhere near it, but I didn't desolder it in fear of voiding the warranty (no warranty void sticker in this thing, though  ).
Given that the DC voltage was so low, what are the odds that the X-class cap in the dropper had failed?

[bktemp]

That is a common failure, especially with cheap Chinese capacitors.
Whenever the plastic film fails because of a voltage spike, a small area surrounding the failure point vaporises and therefore removes the short:
http://www.wima.com/EN/selfhealing.htm
All foil capacitors show this effect, but cheap ones are often underrated and therefore lose their capacity even at normal operation without overvoltages in a couple of years.

[jitter]

Thanks for the reply.

If this is the case, then I'd expect to see the other detector fail soon too. Hopefully before mid-October, before the warranty will run out  .
Despite being from a German brand (Kopp) there was no country of origin stated on the package or the detector itself. The internals were neatly assembled, no problems there. Some of the bigger components had brands on them, but none of them were from the well known companies. Very possibly made in Asia.

Should the other one fail after the warranty period I'll first try a new capacitor.

[SeanB]

When replacing the capacitor use a 400 or 630V unit, and add a 270VAC MOV across the input terminals as well, this will ensure the capacitor does not get massive voltage spikes which will kill it. There should be a resistor of between 47-220R in series with the capacitor. If this is in the line side add the MOV after it, that way both will last longer and the resistor will be the fuse.

[wraper]

When replacing the capacitor use a 400 or 630V unit

Strange suggestion. ~250V X2 Capacitors already have higher DC rating than 600V. Usual capacitors with 400V DC rating will fail very fast in such use.

[SeanB]

Sometimes you cannot get the X2 unit, so a 400V or 630V film has to be used. The original probably used a plain polyester film unit of 250VDC rating. The 250VDC film will survive mains for a while as it has a puncture voltage of around 500V minimum, thus will work but will degrade fast. The proper approach is to use 2 630v devices ( twice the capacitance value) in series with some bleed resistors across them. That will last decades. Here you really do not want a class X cap that can fail short but want a class Y that fails open, but the class Y is limited to 47n commonly unless you get the ones for permanent installations.

[wraper]

Sometimes you cannot get the X2 unit, so a 400V or 630V film has to be used.

Definitely not 400V cap, it will eventually fail even without transients in the mains.

The original probably used a plain polyester film unit of 250VDC rating

I strongly suspect the 390 nF X2 cap in the capacitive dropper circuit had failed.

[jitter]

Thanks for the suggestions.

Space is very limited inside the device. If I am to replace the cap, it must be of exactly the same dimensions. There also won't be any room for a MOV.
There is what looks to be a 2-5 W resistor in series with the cap.
I checked the main board in the replacement detector before I installed it, and it looked identical to the old one. The X-class film cap was of a different brand, but that may not mean anything.

Wit a bit of luck, the unit that died had a weak cap in it and the other one and the replacement will be fine for years to come. I don't really look forward to having to repair these detectors every couple of years.

[NiHaoMike]

I rarely see X rated caps fail. I guess they handle 120V with ease but 240V is a bit too stressful for them...

[amyk]

Video on how these capacitors lose capacity over time:

[jitter]

Thanx for the link, I hadn't seen this vid from bigclivedotcom yet.

It looks to me like the capacitor he disassembled suffered from moisture damage, judging by how the metal was eaten away at the edges.
This shouldn't be the case with my motion detectors, they're hanging in a dry location that won't even get damp.

He also talks about the mottled appearance of the film elsewhere, I can imagine that's transient damage. Now I'm kind of hoping the other one fails after the warranty has run out... patience... 

[Zero999]

If this is the case, then I'd expect to see the other detector fail soon too. Hopefully before mid-October, before the warranty will run out  .

In the UK you may still be entitled to a replacement, even after the warranty period has past, especially if it's failled due to a design defect. Research the consumer law in your area.

[jitter]

If this is the case, then I'd expect to see the other detector fail soon too. Hopefully before mid-October, before the warranty will run out  .

In the UK you may still be entitled to a replacement, even after the warranty period has past, especially if it's failled due to a design defect. Research the consumer law in your area.

Time to revisit this old thread, because the second detector has failed.
The first one just failed dead, this one was stuck in a switch on and switch off loop. It would keep turning off after the set time immediately followed by it turning on again, regardless of whether or not there was any motion.

Obviously the first thing I went for was the X2 cap in the capacitive dropper, and this one too had worn beyond usable capacity. Of the 390 nF less than 82 nF were left. This resulted in a drop of the nominally 24 VDC to about 6.9 VDC every time the relay kicked in. Probably enough to upset the 78L05 in the detector board, hence the odd behaviour.

With this one well past the warranty period, I might have gone to the store and see if I could get a new one nevertheless, because I believe the same laws as quoted apply here too. But to be honest, it just didn't seem worth the hassle and I wanted to do the repair myself. I found an old Epcos 220 nF 250 V~ - 305 V~ (what's up with the multiple voltages?) X2 cap in my stash and tried it. Worked fine, the voltage stayed well up above 23 VDC with the relay switched on and went to nearly 24 VDC when switched off. Good enough for me, so I didn't even bother sourcing a new 390 nF X2 cap.

If the design really is flawed, I should see the other detector, that I did get a warranty replacement for, fail within a year or two. If that one won't fail in that short time period, it was probably just a bad batch of X2 caps used in that production run. If it does fail, I might contact Kopp, the manufacturer of the motion sensor.

To be continued?

Edit: hey, I just noticed in the picture that the value of the the gray X2 cap reads 280 V~, but under the UR logo it says 250 V~... again, what's up with the multiple voltages?