Author Topic: Looking for mechanism by which motor run capacitors would be slowly destroyed  (Read 2113 times)

0 Members and 1 Guest are viewing this topic.

Online johansenTopic starter

  • Super Contributor
  • ***
  • Posts: 1182
Standard off the shelf cheap motor run caps, 40uf at 370 or 440volts.

They filter a saturatable inductor at perhaps 8khz pwm to produce a 240v sine wave.

Nominally the capacitor will have 4 amps at 60hz and perhaps 4 amps rms at say 8khz. But during 150 amp overload, the capacitor may have my guess is 20 or 40 amps rms at 8khz.

Failure is gradual, but once the capacitance drops too far the customer complains, for obvious reasons.

Only idea i can come up with is thermal expansion cyclical stress in the capacitor.
Any other ideas? Over voltage isnt possible.

 

Offline Psi

  • Super Contributor
  • ***
  • Posts: 10385
  • Country: nz
Maybe just spec a 650V cap.
How are you sure overvoltage is impossible?

Maybe the motor has a minor fault, or is just very inductive, and is producing high current pulses which are causing higher than normal voltage spikes over 375/440V.  Or could be spikes coming in from external dirty power at the factory.

Or could just be cheap china motor run caps, have you tried a better brand?

Could also check what temp the cap is running at in their installation. Maybe it's just getting too hot and failing from accelerated old age

It's a bit messy, but you can take the faulty cap apart and unroll it to see if its full of areas that have been blown away from overvoltage or not.
« Last Edit: November 01, 2023, 04:38:42 am by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline NiHaoMike

  • Super Contributor
  • ***
  • Posts: 9323
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
Try using a pair of 80uF in series?
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline woofy

  • Frequent Contributor
  • **
  • Posts: 377
  • Country: gb
    • Woofys Place
Maybe heat and/or poor quality capacitors. These are non-polarized aluminium electrolytic capacitors and if the electrolyte drys out they will degrade.

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 10174
  • Country: gb
Only idea i can come up with is thermal expansion cyclical stress in the capacitor.
Any other ideas? Over voltage isnt possible.

Motor run capacitors are metalized Polypropylene film. Under transients they self heal by evaporating the metalization around any film puncture. This leads to a loss of capacitance over time. It could be that if the capacitors experiences currents of 20A+ the current density in the metalization becomes high enough to vaporize or disconnect areas of it. This would account for the reduction in capacitance that you're seeing.

Several manufacturers, eg. Epcos, produce ranges of film capacitors that are specifically designed for high frequency, high current / high pulse applications. They will be nowhere as cheap as bog standard motor-run capacitors though.
Best Regards, Chris
 

Offline Circlotron

  • Super Contributor
  • ***
  • Posts: 3366
  • Country: au
Maybe heat and/or poor quality capacitors. These are non-polarized aluminium electrolytic capacitors and if the electrolyte drys out they will degrade.
You're thinking of motor start caps, not motor run caps. Different beast.
 
The following users thanked this post: wraper

Offline DavidAlfa

  • Super Contributor
  • ***
  • Posts: 6403
  • Country: es
Just bad quality.
Been told a story where heavy machinery needed fridge-sized panel full of caps!
It was a LOT of money there! They had options A (Let's say $10K, unknown brand) and B (maybe 25 or 35K, top tier brand).
Meh the 10.000 will do it! 6 months later it failed, most caps where shot or heavily degraded.
Guess what, they put the good ones that time, lasted several years, the guy left the company to go on its own, and they were still working.

Few weeks ago I went to check a friend's water pump.
The cap he starting cap was dated 1974, rated for 125V, and had been working at 230V maybe 35 years, still strong. That's crazy!
« Last Edit: November 01, 2023, 01:04:53 pm by DavidAlfa »
Hantek DSO2x1x            Drive        FAQ          DON'T BUY HANTEK! (Aka HALF-MADE)
Stm32 Soldering FW      Forum      Github      Donate
 

Offline woofy

  • Frequent Contributor
  • **
  • Posts: 377
  • Country: gb
    • Woofys Place
Maybe heat and/or poor quality capacitors. These are non-polarized aluminium electrolytic capacitors and if the electrolyte drys out they will degrade.
You're thinking of motor start caps, not motor run caps. Different beast.
You are correct, my bad.

Online johansenTopic starter

  • Super Contributor
  • ***
  • Posts: 1182
Maybe just spec a 650V cap.
How are you sure overvoltage is impossible?

It's a bit messy, but you can take the faulty cap apart and unroll it to see if its full of areas that have been blown away from overvoltage or not.

Not enough voltage on the dc bus to destroy a 370vac rated film cap. Yes if i can get my hands on some failed caps i will take them apart.

Motor run capacitors are metalized Polypropylene film. Under transients they self heal by evaporating the metalization around any film puncture. This leads to a loss of capacitance over time. It could be that if the capacitors experiences currents of 20A+ the current density in the metalization becomes high enough to vaporize or disconnect areas of it.

It seems to me it would take a lot of amps beyond just 10x nameplate nominal expected conditions to do that , but yes i agree its possible.

Just bad quality.
(...)

Few weeks ago I went to check a friend's water pump.
The cap he starting cap was dated 1974, rated for 125V, and had been working at 230V maybe 35 years, still strong. That's crazy!

Not abnormal. The duty cycle is so short an electrolytic start cap will reform and survive. Also many motors have a 125vac rated start cap that is connected to only half the motor windings, when used at 230vac. (The motor becomes an auto transformer).
I just replaced a 170vac rated start cap on a 120vac refrigerator a few weeks ago . Replaced it with some motor run caps big enough to start the motor. Compressor failed again week later. Turns out simply unplugging and plugging in the compressor start relay was enough to loosen the connection and it burned up. Luckily it didn't release all the refrigerant.
« Last Edit: November 02, 2023, 02:03:03 am by johansen »
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Got any failed parts?  Take 'em apart and see (mind the goopy oil; not a hazard, just messy, they haven't used PCB in DECADES, even the cheaper countries know better than to touch the stuff... I think?..).  The large current sounds like a good way to melt the connecting straps.  Note the internal construction is a foil-film jelly roll with a couple tabs crimped into it; the tabs are just floating in air, headspace above the roll, before the sealing plate and terminals.  Or maybe they're immersed in oil, but the power dissipation won't be much better in that case (very little circulation).

As a matter of design review -- mind, not knowing anything about what this is, other than, I guess, some manner of mains inverter -- it sounds like inappropriate type capacitors were chosen, perhaps simply because motor-run types seem familiar enough for mains use, or are cheap and abundant; but I would consider several points against them, the fairly weak connections for one, but also the large inductance of same (figure >20nH per cap, plus wiring!), the low frequency range (intended 60Hz use only), whether MFG has any data on higher-frequency operation (almost certainly they don't provide a V or I vs. F rating/plot), and whether shopping revealed any better/more suitable alternatives (axial tape-and-fill, and boxed style, esp. X1/X2 type MKP, are much more likely to handle the current and frequency, and aren't particularly expensive).

I also don't know whether this is actually a design scope question or repair/maintenance, but it at least seems likely that attention should be brought to the designers if nothing else.  Maybe it's also very old and that's N/A, I don't know.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline jonpaul

  • Super Contributor
  • ***
  • Posts: 3656
  • Country: fr
  • Analog, magnetics, Power, HV, Audio, Cinema
    • IEEE Spectrum
Motor caps are made to JUST meet the spec and lowest cost.

Both film and oil filled caps are available as motor/run/start and lighting ballast/PFC.

Most are rated ONLY for mains freq ripple current.

At 8 kHz or HF, the ESR/ESL will cause higher losses and stress the terminations to foil connections.

The SMPS/HF rated caps have multiple tab to foil and end caps to increase the HF current rating.

Consult the cap mfg specs if HF use above mains Freq is needed.

Jon

An Internet Dinosaur...
 

Offline CJay

  • Super Contributor
  • ***
  • Posts: 4136
  • Country: gb
If I understand your first message, you're using run capacitors designed for continuous duty at 50/60Hz in a non-motor application at 8KHz and asking them to supply 20-40 amps?

Have you considered they just might not be suitable for that use?

Have you measured their temperature under load?

 

Online bdunham7

  • Super Contributor
  • ***
  • Posts: 8178
  • Country: us
Any other ideas?

Use a capacitor that you have complete specs for and is designed for the purpose?  Actually measure the current through the capacitor under regular and "overload" conditions and then contact a product rep and ask them to help you select the right unit?

KEMET C4AQIBW5400M3OJ is a 40µF DC-Link capacitor rated for 28.7A @ 10kHz and it costs $6 in bulk quantities.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Online johansenTopic starter

  • Super Contributor
  • ***
  • Posts: 1182
If I understand your first message, you're using run capacitors designed for continuous duty at 50/60Hz in a non-motor application at 8KHz and asking them to supply 20-40 amps?

Not my product, but its an established company that made some changes recently, one of which is they now use toroidal inductors instead of E cores with wide open air spaces. this is leading me to think the inductors saturate much more harder faster and this is why the capacitors are failing much quicker.

But yes the ripple current they have to pass is going to be rather small during the nominal load. product is designed to handle 5 or more times the nominal load for 2-5 seconds, and the only way to do that economically is to just let the inductor saturate.
 

Online johansenTopic starter

  • Super Contributor
  • ***
  • Posts: 1182
KEMET C4AQIBW5400M3OJ is a 40µF DC-Link capacitor rated for 28.7A @ 10kHz and it costs $6 in bulk quantities.
interesting, that's cheaper than many motor run caps.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7681
  • Country: ca



Film capacitors that fizzle and fade, going low value- are common and due to the metallization failing from corona or corrosion. You get discharges that cut and slice up the metalization and reduce the plate area. Hurrah "self-healing" lol.
Metallized Film Capacitor Lifetime Evaluation and Failure Mode Analysis
SEM Lab Failure Analysis of Film Capacitors

It would be worthwhile to do an autopsy of your caps. If they're oil-filled I think that is a different bird, as run caps they can take much higher ripple current.
 

Online bdunham7

  • Super Contributor
  • ***
  • Posts: 8178
  • Country: us
It would be worthwhile to do an autopsy of your caps. If they're oil-filled I think that is a different bird, as run caps they can take much higher ripple current.

I think they're typically resin-filled and an autopsy would be a challenge.  Edit:  There are oil-filled as well, but I found none with detailed specs. I don't know about them taking much ripple current, they are made to be physically rugged more than anything else.  Most of these don't have very detailed specifications, but let's look at a KEMET C870CF35400AA0J.

https://www.mouser.com/datasheet/2/447/KEM_F3063_C87-3316450.pdf

*   420/470VAC, depending on the life you would like. 

*   Specifically rated for 50/60Hz, implying a current rating (not explicitly stated) of about 7 amps.  Maximum permissible current is listed as 1.3X.  That says to me that the cycle current should never go over about 9A.

*   dV/dt is 15 or 20V/µs, it appears to me that in the OPs circuit 240VRMS @ 8 kHz he'll have a dV/dt of about 17V/µs. 

So we'd need actual measurements of the ripple current in his application.  This capacitor might be marginally OK or it might be not acceptable under normal conditions.  It would seem to not be adequate for the overload conditions if his description is accurate.  The original capacitors they are using may be of much lesser quality since he describes them as 'cheap'.  He doesn't say what sort of life they are getting out of the current models, I'd be curious if they are getting 1000 hours or 10,000 hours or what.

« Last Edit: November 03, 2023, 05:00:57 am by bdunham7 »
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline chinoy

  • Regular Contributor
  • *
  • Posts: 95
  • Country: in
    • RDDreams
I had this problem. In the end it boiled down to the wave form that was being feed to the motor.
The generator / UPS that would power it was not putting out a clean sine wave. So I was changing caps every 3 months till we cleaned up the input power.

Other tricks I use is buy more UF and voltage than you need. That way as the caps ages it lasts longer.
Next mount the cap in such a way that the heat of the motor does not impact it. I had a compressor that kept burning thru caps. As soon as I installed the cap off the compressors heat and vibrations cap life was never an issue again.
Good thoughts, Good deeds, Good words.
All the religion you need in life.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf