“I did not try to do any measurements at “working voltage” -”
Probably shorts at peak voltage.
Just replace it.
mrsmith: Of course the capacitor has to be replaced. And indeed, as I learned, self-healing capacitors eventually get to a point where further use at normal working voltages results in failure.
However, “All”, I darn well would prefer to not have this problem again (where I’m away, the main pump fails and a power failure drains the battery backup pump eventually.) The cost of a new pump is secondary to the grief incurred in replacing it, which is in turn vastly less than the problem of a flooded basement. Yeah, I could just pre-emptively replace pumps every 3 years or so, but that just runs against my grain, so to speak. A $250 (plus tax) pump failing due to the failure of a $3 part (once I know where to look for it) - and these failures are REALLY common - while the rest of the pump looks like it has another 10+ years in it, well.... My personality is such that I say - “Wait a minute: Can this be improved?”
Reading many reviews, actual pump failures seem rare, Usually either the capacitor goes out, or the sensing / on-off switching fails, or something trashes or blocks the impeller.
Unrelated to actually improving the pump, my first step, once I get my main pump operating again, is to install 2 of these. They are cheap enough that I don’t feel like building them.
https://www.basementwatchdog.com/watchdog-water-alarm/
https://www.lowes.com/pd/Basement-Watchdog-Battery-Operated-Water-Alarm/1005609
The “cool” thing is that the wire to the sensor can be extended almost indefinitely: I tried adding a 2k ohm resistor inline, with no effect on operation that I could detect. So, I can place one alarm “box” in the house (so my wife will hear it) and one in my shop. Some sort of Wi-Fi audio monitoring should also be fairly easy to implement. (Anyone have suggestions? A “baby monitor”?)
Back to the pump failure, though:
As part of my “journey” of investigation, I found this article, which (at least for me) was fascinating, AND, I think it has me pointed in the right direction to greatly extend the time before the next capacitor failure, and maybe catch the next one before it happens...
https://www.ecicaps.com/tech-tools/technical-papers/ac-film-capacitors-for-inverter-output-filters/
The article’s discussion of failures of self-healing film capacitors leads me to believe that the turn-on and turn-off transients of the switching to the sump pump motor gradually degrade the capacitor: Eventually these very brief but frequent* peaks degrade the self-healing feature of the dielectric to the point where avalanche failure occurs. In others, like mine, the capacitance itself is significantly degraded, fluid is expelled, etc.
*I should add that while I don’t know that it would be called “the water table”, this basement floor is actually below the level of “saturated ground” here, at least 1/2 the year... Improving surface drainage away from the building, some years back, had no discernible effect. The pump goes on and off a lot — one reason I went with a main pump with an electronic sensor.
Increasing the voltage rating of the capacitor should help, but:
All else being more-or-less equal, “real” (not just rated) significantly higher voltage capability requires a significant increase in the physical size of the capacitor. I can get a LITTLE higher voltage cap in there, but a cap the size of a soda can is a no-go. I am NOT going to rework the motor housing.
2nd, inductive (motor turn-on / turn-off / relay) voltage transients have a tendency to increase to whatever voltage they “need” to reach to find a conductive path. So, going from a 370 VAC cap to a 450 VAC cap may help a little, but it is likely not a real solution.
Instead, I believe the solution is to give that energy somewhere to go, and that somewhere is a transient voltage suppression circuit. Since my “main” pump has a control “box” that plugs into an outlet, and the pump plugs into the box, with a sensor wire running from the box down to the pit, I can suppress ahead of the pump relay, after it, and around it. (The parts order goes in to Digi-Key, tonight - heheh.) Much of this “could” be done with 2 high end surge suppressor outlet strips, but, heck, I like building my own circuits... Plus I can build in a provision to tap in and check motor turn-on current once in a while. That should clue me in as to how the capacitor is doing.
The main idea though, of surge suppression, is to tamp down the switch on/off transients, and let the varistors (MOV’s) take the abuse, rather than the cap in the sump pump getting whacked every time the pump turns on or off. The varistors will eventually wear out / be used up, but, if I put a couple thousand joules of MOV in each leg of the circuit, they should last quite a while.
A future problem is that the contacts of the relay in the control box will eventually degrade. The above transient suppression should help in this regard, but, relay replacement may be something else to look into. Maybe instead, that relay should itself power a moderately high current mercury wetted or solid state relay - I still have a couple stashed... somewhere...
Ah, what fun.
A control box with “soft” turn-on and turn-off characteristics would be a much better solution. If anyone knows of such...?
:-)