Need tips in fixing APC 1500VA UPS having buzzing noise

[vsmGuy]

I have a year 2005, APC UPS and one of the caps - an SIEMENS EPCOS 20uF 150V Capacitor - started making a buzzing noise.

This seems to be a AC, non-polarized capacitor and most local stores have it priced at ~$15 + tax!

They look like this: http://i.stack.imgur.com/CTGU5.gif

I found caps of similar rating (20uF 250V) that are $3 each: http://www.bgmicro.com/24uf220vac.aspx

Do these look like a good replacement? Please let me know if there is something better suited (but perhaps more expensive) parts.

I also had a few questions:

0. Could I also replace these caps with these ones: http://i.stack.imgur.com/FZBrg.jpg

1. Is the buzzing because of the capacitor is bad or that a different component could be failing and having a secondary, audible effect on the capacitor?

2. Should not both these caps be interchangeable?

3. What could be the function of this SIEMENS EPCOS 20uF 150V Capacitor?

To act as part of a LC filter that puts out the sine wave?

4. Should I also change the MOVs on this thing and recap parts of it "just to be on the safe side"?

5. If replacing the MOVs is recommended, what surge rating/part number am I looking for?

Any tips/hints to kepping this nice UPS running would be greatly appreciated

This post also has some relevant information: http://electronics.stackexchange.com...-buzzing-noise

[budm]

Is this UPS a true sine wave inverter or step sine wave type? Do you have full model number?

[vsmGuy]

Originally Posted by budm

Is this UPS a true sine wave inverter or step sine wave type? Do you have full model number?

UPS is a true sine wave inverter; model number being SUA1000VA. I never use step sine wave types :-)

The noise seems to be from the EPCOS cap, but also could be from the chicklet brown cap to its left: http://i.stack.imgur.com/6v3u5.png

[budm]

This UPS looks like it uses high frequency power supply inverter that is modulated with 60Hz then the high frequency gets filters out for L/C circiuts.
Basically it is AC to DC to high frequency AC modulated by 60 Hz then high frequency get filtered out to get AC since wave output. You can tell if that is the circuits being if it does not have really big transformer inside for the AC out put.
You should try to look up the spec of the original cap to see if it is non-inductive type, high pulse type, low ESR type. just match the new cap spec with the old cap spec.

[vsmGuy]

budm, according to my newbie knowledge, these UPS are designed such that it reuses the big input transformer for output too via cleverly designed relay network.

This is the datasheet of the cap: http://www.epcos.com/inf/20/20/db/fc...MKT_B32232.pdf

Unfortunately, I can't buy it a piece from the manufacturer and everyone else is charging $15+ for something inferior, and my bet seems to be to buy it from an online catalog store OR get those fan starter caps like these: http://i.stack.imgur.com/FZBrg.jpg

I have no idea what the specs for those fan starter caps could be.

[budm]

I do not see 20uf rating on that spec sheet. It is a general purpose Metal Polyester film, page 6 is the one you need to look for in other cap spec.

[vsmGuy]

Originally Posted by budm

I do not see 20uf rating on that spec sheet.

I know - that's because EPCOS makes them no more. See the model number on the picture I attached. Even if they did, their MOQ is 4k pieces.

Originally Posted by budm

It is a general purpose Metal Polyester film, page 6 is the one you need to look for in other cap spec.

In all probability, the capacitors I am going to buy are going to be generic ones with no data sheet. That's why I asked the questions in my first post to begin with and would very much like your feedback on them.

[Toasty]

Please attach pics to posts. Links to external site disappear or become broken, and the info provided by you is lost.

That being said, Epcos B32232 sheet attached. Metallized Polyester Film.

SUA1000VA is not a 1500VA UPS...? 1000va = ~670 watts / 1500va = ~980 watts

Noise may very well be from another component. Usually, by time these caps start buzzing, it's too late. I've never seen one fail. The orange drop caps nearby (also MPF) I see fail when high spikes on the line occur. Telltale is a change in color or a crack/burn in the casing.

Yes, it's used for "smoothing" the output when in "running" mode.

Be nice to see a pic of the whole area. I'm assuming the pic in #3 is your unit.

Could also be noise from another device on the line this is connected to. Cheapo Chinese fluorescent ballasts radiate that and can be feeding it onto the line. Try moving it to another circuit.

You have enough "noisemakers" nearby to probably discount these as failures. I see 6 relays and several coils/chokes nearby. That entire bank of rectifiers/FETs can also be noisy at times.

DigiKey & Mouser carry some of this Epcos line, but not enough. $20 sounds about right going by their other caps in the line. That is a big mother.

T

PS: Make sure any output side power strips are -NOT- surge protected. They -say- sine wave, but spikes do occur and will quickly take out the MOVs, especially if lightly loaded.

.

[vsmGuy]

Toasty, quite an insightful post

Originally Posted by Toasty

That being said, Epcos B32232 sheet attached. Metallized Polyester Film.

Yes, I posted the source link to this here: http://www.badcaps.net/forum/showpos...75&postcount=5

Originally Posted by Toasty

Noise may very well be from another component. Usually, by time these caps start buzzing, it's too late. I've never seen one fail. The orange drop caps nearby (also MPF) I see fail when high spikes on the line occur. Telltale is a change in color or a crack/burn in the casing.

The caps are light brown, and hence, a shade darker than orange, but perhaps they were that color to begin with?

Originally Posted by Toasty

Be nice to see a pic of the whole area. I'm assuming the pic in #3 is your unit.

Correct, and I posted closeup shots here: http://electronics.stackexchange.com...-buzzing-noise

Also, a user on this forum did a video of him repairing a similar unit : youtube.com/watch?v=mpUzHO3MGCI

This has a similar layout as mine and would give you an idea what the board is laid out like

Originally Posted by Toasty

Could also be noise from another device on the line this is connected to. Cheapo Chinese fluorescent ballasts radiate that and can be feeding it onto the line. Try moving it to another circuit.

I was under the same impression, but the noise also appears when I take the unit off the line and run it off batteries on heavy load (the noise wen running on batteries increase with load with almost no noise when lightly loaded ~11W)

Originally Posted by Toasty

You have enough "noisemakers" nearby to probably discount these as failures. I see 6 relays and several coils/chokes nearby. That entire bank of rectifiers/FETs can also be noisy at times.

Agreed. However, the noise manifests in the region around the EPCOS.

So the noise could be from either:

1. The EPCOS cap itself
2. The chicklet caps around the EPCOS
3. Another component that is making these caps noisy (This is what I am most afraid of - that even after replacing these caps, the noise remains, as the root cause is somewhere else to begin with )

Originally Posted by Toasty

PS: Make sure any output side power strips are -NOT- surge protected. They -say- sine wave, but spikes do occur and will quickly take out the MOVs, especially if lightly loaded.

Hmm - what you are saying is - trust the UPS with surge protection and just have power taps from the UPS output?

Otherwise, the UPS could likely create spikes itself in the output, which can take out MOVs downstream?

[vsmGuy]

Originally Posted by budm

You should try to look up the spec of the original cap to see if it is non-inductive type, high pulse type, low ESR type. just match the new cap spec with the old cap spec.

Originally Posted by Toasty

That being said, Epcos B32232 sheet attached. Metallized Polyester Film.

I found caps of similar rating (20uF 250V) that are $3 each: http://www.bgmicro.com/24uf220vac.aspx

1. However, these are 12 years old (made in October 1998) - should it be good to go or it's too old to be of use?

2. I am assuming Metallized Polyester Film, specified Metallized Mylar should be better quality being a brand name and not a generic Polyester cap, or is Mylar and Polyester used pretty much interchangeably

[Toasty]

Correct. Surge protectors on INPUT to UPS, not on output.

Having them on output side can also present a load/short to the UPS which it tries to compensate for by boosting output. A vicious circle.

I think the unit is fine and just gone a little noisy. You could replace them and the problem will remain. Because the noise is load variable, I would be scoping the output for spikes or looking further back in the circuit.

Try a paper towel roll as a "stethoscope" to pinpoint the noise. Try moving components with a wood or plastic tool to see if they're just resonating from the switching circuit. It's why PSUs have that glue/caulk around coils and caps.

They make noise when making power. Nature of the beast.

T

PS: I'm now seasick from watching that video....

Some folks have -NO- camera skills...

[vsmGuy]

Originally Posted by Toasty

Correct. Surge protectors on INPUT to UPS, not on output.

Having them on output side can also present a load/short to the UPS which it tries to compensate for by boosting output. A vicious circle.

Things like this make me say "ignorance is a bliss". All the while I had been buying surge protectors and I have only one power tap I got at Walmart. Now I either need to remove the protection PCBs from the strips I have or buy new taps

Originally Posted by Toasty

I think the unit is fine and just gone a little noisy. You could replace them and the problem will remain. Because the noise is load variable, I would be scoping the output for spikes or looking further back in the circuit.

I was afraid something like this might be up, but read on (I don't want to put in $50 in parts + a weekend of my time behind a lost cause)

Originally Posted by Toasty

Try a paper towel roll as a "stethoscope" to pinpoint the noise. Try moving components with a wood or plastic tool to see if they're just resonating from the switching circuit. It's why PSUs have that glue/caulk around coils and caps.

They make noise when making power. Nature of the beast.

I gave it a closer look today - the EPCOS is surely making th noise - I touched it with a plastic rod and I could feel it vibrating. I touched the other components around the region with it and could feel nothing.

Surely this confirms the source of the noise once for all?

Originally Posted by Toasty

PS: I'm now seasick from watching that video....

Some folks have -NO- camera skills...

Actually I PM'ed the guy for his feedback, so we should see him on this thread

[severach]

I remove surge parts from all my strips. They do what they say which is the opposite of what you want. You want them to stop lightning. What they do is offer lightning more paths to damage your equipment.

Lately I've been removing the power switches too. By accident is the only way they get turned off. I like almost hospital grade plugs for next to nothing.

[vsmGuy]

Originally Posted by Toasty

I think the unit is fine and just gone a little noisy. You could replace them and the problem will remain. Because the noise is load variable, I would be scoping the output for spikes or looking further back in the circuit.

Toasty, I had hoped you were wrong and sunk ~$30 in parts to replace the caps (new, exact to original specs).

Noise remains - perhaps a bit louder than before

Originally Posted by Toasty

Try a paper towel roll as a "stethoscope" to pinpoint the noise. Try moving components with a wood or plastic tool to see if they're just resonating from the switching circuit. It's why PSUs have that glue/caulk around coils and caps.

The noise is very well from the 20uF cap we were discussing about - I can feel it vibrate when I touch it with a wooden stick.

Replacing the 20uF cap did not help - the noise remains, as I noted above.

I really like this unit and would hate to trash it, if it has become unsafe.

Any ideas if the unit is still good to power my Computer and Audio Receiver?

(these babies cost me $2k+ and I found it fit to put this APC workhorse to cover their back.)

I still wonder - why the noise? noise comes from the cap when the batteries are charging or when I make them discharge - could it be something because of the batteries themselves (but what?)

[Toasty]

The title says 220vac, you're looking for 250vac. The capacitance is fine.

It's -only- a 12% difference in voltage, and for a trial to see if the noise stops, it's worth it. I would not leave it in place permanently.

Sealed caps like these are generally fine for a long time on the shelf. I would apply voltage slowly to make sure it is stable. A lamp in series with the power line. Or, try just feeding the cap through a lamp externally by itself, before installation. That would be a safer way, in case it decides to go *bang*!

If the lamp goes bright and stays that way, the cap is shorted.

Caution is advised.

T

[vsmGuy]

Originally Posted by Toasty

The title says 220vac, you're looking for 250vac. The capacitance is fine.

It's -only- a 12% difference in voltage, and for a trial to see if the noise stops, it's worth it. I would not leave it in place permanently.

Thank you Toasty for the advice -
The original cap that is making the noise (and the one I intend to replace) is SIEMENS EPCOS 20uF 150V Capacitor.

Given that, these seem to fit the bill just fine: http://www.bgmicro.com/24uf220vac.aspx

It's the age that worries me.

Originally Posted by Toasty

Sealed caps like these are generally fine for a long time on the shelf. I would apply voltage slowly to make sure it is stable. A lamp in series with the power line. Or, try just feeding the cap through a lamp externally by itself, before installation. That would be a safer way, in case it decides to go *bang*!

If the lamp goes bright and stays that way, the cap is shorted.

I can do that - the lamp acts as a current limiter (in case the cap is shorted)? Would a regular 40W incandescent work for this?

[selldoor]

Originally Posted by vsmGuy

I can do that - the lamp acts as a current limiter (in case the cap is shorted)? Would a regular 40W incandescent work for this?

Yes that would be ok - see here
http://antiqueradio.org/dimbulb.htm

[Toasty]

Did you replace the (orange) poly cap next to it? They are wired in parallel. The poly cap should be around 2µF in value. Together you get ~22µF.

T

[vsmGuy]

Originally Posted by Toasty

Did you replace the (orange) poly cap next to it? They are wired in parallel. The poly cap should be around 2µF in value. Together you get ~22µF.

T

Thanks Toasty for the reply.

I replaced both the caps around the area and the noise remains.

Since the noise comes when the batteries are charging or when I make them discharge - could it be something because of the batteries themselves?

I saw these posts on the internet that suggest that could be a possibility:

The whine is usually caused by the charge circuit trying to push the charge to the damaged cells in the battery.

They screech like that when the battery in them needs to be replaced

It would be interesting to find out if APC UPS emit buzzing noise when their batteries are dying.

I hope APC UPS users can pitch in sharing if the heard buzzing noise from their UPS when their batteries are dying.

[vsmGuy]

Attached is a clip of the sound that comes from the 22uF, 150V bipolar yellow capacitor when charging the batteries/online.

At 45s into the recording, I turn off the line power - notice how the noise goes away completely when it's running off batteries in the 1000VA (the one which was running when I recorded the noise). It returns/increases when I connect a heavy load to the UPS output.

I am sure you will agree this is very different from the gentle hum of a transformer?

If this is a characteristic sound of APC UPS when the batteries are dying, I would like to know if there's a battery I can buy, from say Walmart or Home Depot etc just to see if a new battery makes the noise go away.