General > General Technical Chat
Some really burned RIFA capacitors, but not X class
(1/2) > >>
cybermaus:
So my brother trades in old machinery, and he recently got two Bridgeport CNC routers from ' 89
Pretty early for CNC machines in fact. They had been in use until very recently at a university, where he picked them up himself.

Firing the first one up, after 20 minutes of fiddling with its mechanics, it released the magic smoke. And *lots* of it.
Being in an enclosure, it took some time to detect, and some more to respond and douse the fire, see for yourself.

Checking the 2nd machine, with a IR camera at the ready pointing at the same components, indeed, the capacitor themselves very quickly shot to 50C, and crawled to 80C in less then a minute. Normally not a really temp, but it was too quick, and nothing else was even getting warm yet.


Of course, it is always the question when finding a burned component: was this the cause, or a consequence.
Well, given this video and this thread, and the fact nothing else got hot, we think/hope this is the cause. They will be replaced for sure, all 12 of them.

Slightly worry however is that I measured the non-burned ones, and they measure a nice, exactly 0.1uF and 0mA leakage at 360V (that is as high as my leakage test setup goes)
So they measures very well. Except for the ESR, that seems a bit high to me, but I am unsure what is normal for a film capacitor like this.

So question: what would be a normal ESR value for a 80'ies RIFA 0.1uF non-X class film capacitor? Or any such film cap from that era?

I did check some equivalent modern ones, and they are lower, but I am not sure if I should compare 2010 devices to 1980 devices.

Thanks

Gyro:
The failure mode of Rifa MP (Metallised Paper) capacitors is moisture ingress due to the epoxy package cracking. As you would expect, this is most likely to occur after a period of storage. If there is visible cracking then you have no choice but to replace them as a matter of urgency.

With values as low as 100nF, it isn't really a matter of measuring ESR, but Q and dissipation factor [Edit: damp dielectric is likely to cause increased dissipation I suspect]. In this case, it looks as if the capacitors are being used in RC snubbers across the SCRs, where the resistor should swamp such factors. Metallised paper caps have a slightly higher dV/dt withstand capability than metallized film, but again it shouldn't matter in this case.

Either way, you're probably going to want to replace the caps with MKP Polypropylene film - although you do also have the option of using newly manufactured Rifa MPs which will probably be reliable for a few more decades of active use.
eti:
I see a smoking RIFA 😄 but as opposed to a reefer, one doesn’t get a choice, and it’s potentially far from worse for one’s health.
cybermaus:

--- Quote from: Gyro on October 07, 2021, 08:06:54 pm ---it looks as if the capacitors are being used in RC snubbers across the SCRs
--- End quote ---
correct.


--- Quote from: Gyro on October 07, 2021, 08:06:54 pm --- but Q and dissipation factor ... dV/dt withstand capability
--- End quote ---
Ok, got some studying to do, those terms are new to me (in the context of capacitors)

My reasoning of suspecting the ESR is/was (other then they measure a little high, but really not extremely so):
To get hot, there must be energy dissipated, and a high resistance in combination with current constantly flowing in- and out of the capacitor (either AC or spikes) gives a simple I2R

Anyway, obviously I already have polyfilm's on order (I only had 450V rated ones lying about)

Edit: relevant link
richard.cs:

--- Quote from: cybermaus on October 08, 2021, 06:29:01 am ---My reasoning of suspecting the ESR is/was (other then they measure a little high, but really not extremely so):
To get hot, there must be energy dissipated, and a high resistance in combination with current constantly flowing in- and out of the capacitor (either AC or spikes) gives a simple I2R

--- End quote ---
These normally fail the other way -  the leakage resistance in parallel with the capacitance causes enough current to flow for heating. The ones that measure no-leakage are probably OK for now, but will inevitably fail the same way at some point, this type are infamous for it.
Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod