UPS Capacitor replacement 2

[krisys]

We have some 25 kVA, 110V AC UPS in the offshore oil platforms. It is installed in the air conditioned room. These UPS are only just six (6) years old. Vendor has suggested that we should replace the UPS capacitors, for the better reliability.

I wonder, how frequently the UPS capacitors shall be replaced. In my earlier assignments, I have seen some UPS working for more than 10-15 years. No one have suggested to replace the capacitors there. The UPS was working fine.

I want to know, is it an over kill?

Please share your experience on this.

 

[itsmoked]

Systems with big electrolytic caps can start having problems with them in about 6 six years.

It's never a clear case. The life of electrolytic caps is directly tied to the temperature they live at. Depending on the design of the hardware it could have them aging out early.

At their maximum rated temperature most electrolytic caps are rated to operate reliably for between 2,000 hours and 5,000 hours. That's not even one year. But, as the operating temp comes down the life stretches out significantly.

We have no idea about your UPS's operation, is it a pass-thru or an on-line type? Do they get tested frequently?

A second issue with electrolytic caps is "forming". Left uncharged for more than two years results in them becoming "un-formed", if suddenly charged they frequently fail.


If the maker is suggesting this I'd probably follow their recommendation.

Lets say you decide to "let it ride" and a UPS fails when a demand is put on it. How will you look with the UPS maker saying "we warned you"?
If on the other hand the UPSs are tested weekly and a failure wouldn't be during a genuine need you could run them until a cap explodes on testing then replace them all.



Keith Cress
kcress -

http://www.flaminsystems.com

 

[VEBill]

"Vendor has suggested..."

Perhaps the vendor knows something... Contact them, or the OEM's Field Service Engineer, to get the complete picture. It's not likely to be just somebody's unfounded opinion; more likely this comment is based on something going on behind the scenes.

Perhaps it doesn't reach being classed as a recall, but just a known pattern worth preempting.

Historically, there have been issues with poor quality capacitors, but the worst were earlier than just 6 years ago. Ref Ref. Capacitor plague.

In my opinion, UPSs are troublesome in general. There's an amusing tidbit floating around the internet. Ref. UPSs compared to guard dogs

 

[IRstuff]

I've had the occasion to repair a high-end refrigerator controller board with a $3 electrolytic capacitor. Lucky for me, the capacitor was conveniently designed with a "pop-top"


TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[krisys]

Keith, VEBill and IRstuff,

Thanks for your valuable inputs. Specifically Keith's response is quite convincing. Nevertheless others inputs are valuable.

 

[EvilJeb]

The only thing I would add is, there are better and worse electrolytic capacitors. Some good ones are rated for higher temperature and longer hours. Also, even the worse ones will last longer if they're not subjected to a lot of heat.
The vendor may have an interest in NOT providing the best capacitors available, because if they did, they wouldn't have to service the product again in such a short time.

 

[krisys]

I am talking here about the world's some of the best UPS suppliers. In this competitive environment, vendors may not do this deliberately. If they do this they will be out from the market soon.

Yes, due to the inherent limitations of the oil filled capacitors, if this can happen, it is a common phenomena across the UPS vendors.

 

[electricuwe]

Some comments on technical and economical background:

A capacitor bank in any converter has to be designed for the targeted lifetime. For electrolytic capacitors the relevant parameters are temperature and current. Usually manufacturers assume a typical mission profile and design according to this. For a UPS designed for a datacenter 6 years might be a reasonable choice, for heavy industrial applications customers might expect longer lifetime.


On the competitive environment: You can design DC-links with electrolytic capacitor to survive 20 years or more. But this is expensive ! As in most competitive environments initial cost is more relevant than sustainability, many vendors try to save money by a DC-link design by using a very optimistic mission profile.

Some recommendations:

Check what is the target market of your UPS supplier ?
What is the output current related to the rated output current ? Ensuring margin here will increase lifetime of the caps.
Are the units single-phase units or three-phase units ? Three-phase converters need much less capacitance and ripple current capability for a given power and hence they are easier to design in regard to the lifetime vs. initial cost trade-off.

 

[IRstuff]

Capacitor materials are never perfect; I had a video card that sat in storage inside my house, so temperatures ranging from about 65F to 78F, yet, after 6 years, more than half of the electrolytics on the board had "popped."

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm