Opinions sought

Here wizards, magicians, sorcerers and everybody can rest a bit and talk about anything they like.

Just remember to respect the rules.

Opinions sought

Postby jiml8 » Nov 26th, '21, 21:18

Over the course of 2021, I have made the largest upgrade to my workstation that I have made in 11 years, and it has been pretty expensive. I am very very pleased with the performance of this system.

But now I am looking at the power supply.

It is a Coolermaster RS750 supply that my records tell me went into service in this box in September 2010. It is a high efficiency supply (rated 80+ gold back then) running somewhere around 90% efficiency I suppose. At the time, it was well-rated by those who analyze power supplies. Its performance has been flawless.

Over the course of its life, it has always drawn power from a UPS (never plugged directly into the wall), and in various iterations of this workstation got to maybe as much as 75% of its rated capacity in terms of the amount of power draw - and that was at a time when I had 7 hard drives in this box, and 6 of them were SCSI drives.

At the present, it is operating at roughly 15%-20% of its rated capacity to power my current system. It may run a bit higher than that when I have large compiles underway, but I doubt it is ever getting past 30%. At least, I have not observed that.

So.

Should I replace it? On the one hand, flawless performance and no trouble now. Never been overpowered or seriously strained. Well reviewed as a very good quality supply when it was new. Lightly loaded now.

On the other hand, more than 11 years old and powering a system that I just spent more than $1800 USD to upgrade.

I don't want to spend money for no good reason, and I cannot decide whether replacing this supply is a good idea or not.
jiml8
 
Posts: 1212
Joined: Jul 7th, '13, 18:09

Re: Opinions sought

Postby benmc » Nov 26th, '21, 22:00

just my 2 cents, from a related field.

Power supplies are at their greatest stress at power up from fully discharged.
Inrush currents and capacitors getting their initial charge are the usual killer of power supplies.
Add over voltage / spike in the mains supply and that covers about 90% of failures.
Some glues used in pcb assembly, over time, become conductive as they break down, and can cause a catastrophic failure.
Low quality capacitors also leak their electolyte out the mounting pins and can do serious damage through corrosion..
As with all things, though, random component failure take up the balance of failures.

If the power supply has been on 24/7 via a UPS, the likelihood of complete discharge power up and over voltage supply are minimal.
if you can, take the lid off the power supply and look for domed capacitors, or if the plastic wrap around capacitors has shrunk, or note leakage, it means the capacitors have been stressed in some way.
E-dispose of it and replace it.

if you do not observe stressed capacitors, or discoloured glue, it should be good for another 10 years (fingers crossed)

(ok, 5 cents)
benmc
 
Posts: 1015
Joined: Sep 2nd, '11, 12:45
Location: Pirongia, New Zealand

Re: Opinions sought

Postby morgano » Nov 27th, '21, 01:47

Yes it is all about quality of design, components and assembly.

I sometimes service a system at a customer and one of the largest machines is always powered on, since end of the eighties, PLC power supply never changed, just works. 35+ years now. OMRON. Another power supply on same machine lasted less than a year, what crap. Also, never updated software (but i added another small PLC to extend functions twenty years ago. The maker went bankrupt some years earlier and I have no source code for original program - but have read out a binary backup)

Another example of lifetime https://en.wikipedia.org/wiki/Voyager_1

Our microwawe oven is about 30 years. 12 years ago i replaced a relay and fuse, and proactively some electrolyte capacitors, gave it a drop of oil on fan bearings, cleaned out dust. Last week the magnetron broke. Crazy how hard it is to find replacement on a just 30 y old machine - and incl transport cost like a cheap oven, but far from a quality oven. Common parts like this should really be more standardised! This is a stainless steel chassis machine, can easily be maintained hundreds of years if we just replace a part at a time :)
Mandriva since 2006, then Mageia since 2011 at home & work. Thinkpad T43 & T400, Dell M4400, Acer Aspire 7. Workstation using LVM, LUKS, VirtualBox, BOINC
morgano
 
Posts: 902
Joined: Jun 15th, '11, 17:51
Location: Kivik, Sweden

Re: Opinions sought

Postby jiml8 » Nov 27th, '21, 23:14

Electrolytic capacitors have been a reliability problem...since forever. And I recall a time when electrolytics were failing frequently. If I remember correctly, that time was around the time when I purchased this power supply, or maybe somewhat before then. I do know (since I looked it up) that when this supply was available in the market, Coolermaster was advertising it as having Japanese capacitors (it was largely Indonesian capacitors that were failing, while japanese capacitors had a reputation for reliability). As I recall, the key turned out to be the electrolyte the Indonesians were using; apparently it wasn't thermally and/or chemically stable over time.

And I have had to replace electrolytics in the UPS that services this workstation; the battery charging circuit failed due to a leaking electrolytic cap.

So, probably the best idea, as suggested, is to remove the cover from the supply and perform a visual inspection. I might try that, if I can get the supply out of my workstation without disconnecting everything. The cables on this supply are not modular, and I don't know if they are long enough for me to get the power supply far enough out of the box to open it, without disconnecting it. And disconnecting it...nah. I don't want to do that. Those cables snake everywhere.

But, then, you both make good points; a properly designed, properly sourced, and not over-stressed supply should last indefinitely. And this one is properly designed and properly sourced, and I have no knowledge of it ever being over-stressed and being on a UPS would guard against that anyway.

So, perhaps I am unduly worried.
jiml8
 
Posts: 1212
Joined: Jul 7th, '13, 18:09


Return to The Wizards Lair

Who is online

Users browsing this forum: No registered users and 1 guest