exploding capacitors

[rgermain]

This thread will be similar to a recent one regarding condensing fan motors, but since that one hasn't received a conclusive answer, I'm reluctant to hijack his thread.
Application is a two pump skid. Pumps are in parallel with a common inlet header and a common outlet header. Only one pump will operate at any given time in lead-lag configuration. There are check valves on the discharge side of each pump to prevent back flow through the pump that is down. Tech (knowledgeable tech) switched pumps and reported watching the de-energized pump motor spin slowly in reverse direction and suspects that the check valve is leaking by causing the pump to turbine backwards. As he was watching, the cap violently exploded. The contactor was open. The motor is an inductive single phase motor with start and run caps. AC caps aren't supposed to have a polarity. I have to investigate next week. Anybody seen this before?

 

[MikeHalloran]

How sure is the tech that the motor was actually spinning _slowly_, and not spinning fast while being illuminated by AC lighting, so the apparent slow rotation was an illusion?

I have seen centrifugal pumps being backdriven at impressive speeds. The reported symptoms included reduced bearing life, on a succession of brand new replacement motors. I have not seen exploded capacitors in similar circumstances, but I surely haven't yet seen all the ways in which a capacitor can do something spectacular.

Consider multiple check valves (of different types that will hopefully fail in different ways at different times), or power operated switching valves driven by the skid controller.
... Or reverse rotation sprags/brakes.




Mike Halloran
Pembroke Pines, FL, USA

 

[itsmoked]

You could have the motor generating as it is being driven and ultimately you can have the voltage (with no where to go) exceed the capacitor voltage rating. KaPOW.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

Did it happen during commissioning?
Was it first time this configuration and componentes were used by your organization?
Or is it a standard design that you have used before?

In short - is this a possible component weakness? Or is it some kind of design oversight?

I ask because there have been exploding AC capacitors in my life and they could be traced back to batches that didn't quite meet expectations. Some manufacturers do not test all their products 100%


Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[OperaHouse]

I have a trick to make 240V single phase muffin fans work on 120V AC. I put a capacitor in series. When LC
is at resonance the voltage on the capacitor and inductor (motor) is many times higher than the supply. I
choose a vale that is off resonance that runs the motor at about 180V for quieter operation. On RE sites
I see pump motors being used as poor boy hydro generators. This leads to two theories for an electrically
isolated motor.

If the motor can reach an over speed condition, the frequency may be close enough to LC resonance that the
run capacitor sees over voltage and fails.

Start capacitors are not designed for continuous operation. If the motor stays in an under speed condition
and the centrifugal switch never opens the start capacitor overheats after a period and explodes.

I have another theory but it involves aliens.

 

[rgermain]

Thanks guys,
No, several people corroborate the spinning slowly version. It was during commissioning of the other pump being replaced due to smoke. I designed the panel and have used it at a half dozen sites or so over the years.
I initially considered the motor generating concept (still my best guess) but an AC cap is able to change polarity 60 times a second. My next thought was voodo, but aliens sound more reasonable. Perhaps the cap was bad, but it exploded with AC power removed. What I plan to do is explore the possibility of a short causing errant current somewhere. Incidently, I misspoke when I stated that the pumps employed a lead-lag strategy. It's actually a 100% redundant system using one pump as a spare.
Thanks again guys. Generally it's the fundamental things that are overlooked.

 

[waross]

Changing polarity is not an issue. Changing frequency is. As the back EMF frequency (closely related to the motor speed) changes, the impedance of the capacitor and the impedance of the motor winding change in different directions. If the motor is properly disconnected, the only source of voltage will be the back EMF. I suspect that several people were misled as to the motor speed due to the strobe effect of the lighting.
Regardless of the actual speed, the evidence strongly suggests that the speed was sufficient for the combination of the frequency and the magnitude of the back EMF to destroy the capacitor.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[jraef]

Whether or not the speed was or was not faster than reported may not be quite as relevant as what OperaHouse and itsmoked said, that the Start caps are not designed to be in the circuit for long, and then with the contactor open, the voltage on re-gen from the spinning motor had nowhere to go, built up to beyond the capacitor rating and made it pop.

Now it comes down to WHY a motor that is disconnected from the line was regenerating at all, because in theory, there was no grid excitation of the windings and without excitation, no generation. Two possibilities on that:

1) Residual magnetism in the motor was sufficient to make it BEGIN to regenerate, then the Start cap kept it going until such time as it popped the cap. There are several website devoted to the "home power gen" market wherein people seek out and find single phase motors with high levels of residual magnetism, then set up a capacitor circuit to resonate with that and make a self-excited island generator out of an AC motor. Not all motors will have sufficient RM to accomplish this, but clearly some do.

2) The operator "selectively removed" the part of the story wherein he energized that pump by mistake, immediately turned it off, then energized the other one. That scenario might explain WHY there was back flow because if the check valve was dampened, it had not yet closed when pump 2 was started. The on top of that if he performed that changeover too quickly, the centrifugal switch in Pump 1 had not even engaged yet, so the Start cap was ALREADY charged and in the circuit supplying excitation, then when Pump 2 was energized and created back flow through the still open valve on Pump 1, it didn't take a lot of regen to exceed the capacitor value and pop the cap.

I vote for #2. Leaving out parts of a story is a common way of avoiding blame. He may not have understood WHY that happened, but he likely knew that he SHOULDN'T have done that.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[OperaHouse]

It would certainly help the story if we knew which capacitor went bad.

 

[rgermain]

Don't know yet. I'll let you know next week.

 

[rgermain]

Examined pump motor yesterday. 7-1/2 HP, 240 VAC, 1 phase motor. Both start caps were blown. When I opened the discharge butterfly valve, the motor turbined in reverse approx. 60 - 80 RPM. Big dirty spot on shaft made it easy to see. Using a Fluke 787, I recorded approx. 1 volt being generated while it was rotating and repeated the exercise several times to verify repeatability. When the valve opened, voltage occurred. When the valve was closed, voltage immediately stopped. Label on caps is attached photo. Motor winding resistance is 3 ohms. This is contrary to everything that I thought I knew.