SilverStone Olympia 1000W Power Supply 7 Year Redux @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,599
SilverStone Olympia 1000W Power Supply 7 Year Redux - Many people ask about long term computer power supply testing, and simple truth is that it is too expensive for HardOCP to do in-house as it would require hundreds of thousands of dollars' worth of resources. However, we can give the inquisitive a non-scientific look at how well a personal PSU does in our testing 7 years later.
 
Last edited:
I noticed some degradation in a CoolerMaster power supply that came with a case I purchased in the Q6600 days. As I upgraded from system to system and component to component over the years little quirky things kept happening. Then every so often I would blue screen and I would chalk it up to my overclock. Then one day I replaced the power supply because I got a deal on a "newly" used case and the Corsair power supply came with it. All the quirks stopped and everything was good again. I still keep the CoolerMaster around for worst case scenarios like my current power supply just flat out dies and the nearest computer store is a 2+ hours drive away.

My Asus Sabertooth 990FX v1 motherboard is getting old also. I ordered it the day after [H]ardocp reviewed it. This would have been around June 2011 I believe. It's been running 24/7 every since and it has gained some quirks like not wanting to start up if kept off for more than 30 minutes. It was the 2nd best motherboard available at the time in which I purchased it. It has held up really well over the years and looks like it might even get to see SteamOS and retail Windows 10.

Thanks for the article! Great work as always!
 
Thanks for this review of a well-used PSU, it's a good insight into what we MIGHT EXPECT after several years of service.

Even if it's still working fine a PSU that long in the tooth tends to lose some performance. It would be nice if you guys could make this a regular thing, selecting PSUs for powering your test benches (or reviewer';s home systems) so you an compare performance X years later. Not for statistical information (getting that many units and that many testers would be incredibly expensive), but to give a bit more insight to what sort of other issues you might see (e.g. if you can sample form multiple PSU builders / OEMs).

One test gives us some insight, but leave a whole bunch of questions unanswered.
 
Interesting read. I am still using a BFG 1KW PSU that has seen damn close to 24/7 use since 2009. I've often wondered if it's still OK to use. For now though, it still seems to be working fine. Next time I do an overhaul of the system it's powering, I might finally put it to rest.
 
Modred189 said:
To which part?
The system specs for the various systems this PSU supported ('cause I read the article and re-skimmed it now and I don't see it) or the long term test? (Which was sarcastic/facetious, hence the smiley.)
Click to expand...

On the Conclusion Page:
Hardocp.com said:
Paul's Thoughts:

Today, we took a look back at the SilverStone Olympia 1000W because people have asked about this kind of issue a number of times and we have never really been able to give any kind of real answer. We were able to do this today only because we had data on this exact sample from its original review and I retained control of the unit and its use. That is truly the extent of the why we did this article. This is not a "hit piece" or meant to be about SilverStone specifically. After our initial review, I found this unit to be of sufficient quality and performance to put it into service in my personal desktop, in a file server, and in a few other backup systems with the unit running for me for the last 7.5 years.

I pulled this unit from the review pile years ago because it was a more robust design than lots of other products on the market and I needed something that I could use and not worry about. In that regard, this unit has slaved away for 7 years without me doing much to be nice to it other than cleaning it out with canned air every so often. So, after breaking down the last system that this unit was serving in and selling off (or recycling) the parts of that system we had an opportunity to address a simple curiosity; "How does that Olympia 1000W perform after 7 years of service relative to when it was brand new?"
Click to expand...

I'm thinking it's safe to assume this unit got near daily use for 7 years straight and fairly heavy use for several of those years when it was in his personal desktop and file server.
 
Honestly? I thought this was an amazing article. I certainly wouldn't hold this against Silverstone any more then I would any other OEM. The stress tests they put things through in a lab to represent life cycles honestly in no way represent actual use. This actually would make a case of what to expect out of any PSU over the years. I'm much the same in that the PSU usually doesn't get upgraded unless it is giving me problems. However I can say that the results here pretty well mirror my experience with 5+ year old PSU's in that you do start seeing problems especially when loads get heavier. I don't really have anything of value to add here other then I really sincerely enjoyed this type of article.
 
I have a 1200W PSU I just retired after 6 years use. It is still going string, but I got a Corsair for a good price. It is now on standby.
 
Based on my hobbies, I'm unsurprised that a power supply containing vented electrolytic capacitors has degraded and presents more ripple. The electrolytic capacitors are essentially perishable products, and seeing more ripple with age is very common in even high quality electrolytics. I'm guessing about how much the supply was used (guessed "a lot") but they're definitely getting into the "spontaneous failure" period of their existence. I've replaced so many electrolytics over the years if I get something with old ones at this point it's faster and easier just to replace them than to test them.

This PSU seems to have held up fairly well imo.
 
I noticed the same thing with an ST-1000 that I bought to run my 2900XTs around that time period. When I initially got my GTX 465's it couldn't boot an X58 system with 2 of them, 1 it had no problem. These units definitely degrade, but it can still power a moderately spec'd systems after 5+ years.
 
This article is perfectly timed for me personally. My ~8 year old PSU crapped out on me this weekend and it took quite a bit of troubleshooting to narrow it down to the PSU. Fortunately I was able to poach my wife's old rig of it's power supply and my machine is back up and running, but it's just not the same anymore. I don't get those FUCKING RAD lights making a particularly familiar glow in my gaming space. It was the only part in my machine that I cared about looking good (old danger den acrylic case let the glow out)

I've been in a bit of a funk mourning it's demise all weekend and today. I've been sadly crawling around various computer parts shops online trying to find something that might fill the whole in my heart. Reading this article has helped with just accepting it. These things age the fastest it seems, and I should have known the reaper was coming for it.

For the curious... it was a Tagan BZ Series 1100W PSU. I've had it powering an XFX 790i mobo with a QX9770 and various cards in SLI for years. Never really felt a need to upgrade anything but the ram and the occasional graphics card bump.

whimper
 
Interesting, I've got a few PSU's kicking around of various ages. One x58 i7-920 machine has developed some "quirks". Might be the power supply. I forget the brand on it even, it's bright red and built like a tank.

Always made it a point to use higher quality PSU's in my builds. Cheap ones are usually the source of many problems.
 
Great re-review! really like the idea behind this. At lower loads this PSU is still in spec, so overbuying on wattage isnt really as bad as people say. My 1200w supply is probably only pushing 800-900w leaving breathing room and extending the life of the components with less heat generated. And as the components do degrade I will hopefully still be in spec at ~75% load. Again great review piece! Now lets see how these digital power supplies do in 7 years
 
Excellent article!

Now the next question is probably, "Will this be a regular feature even in its non scientifically rigorous form just to provide examples?" The answer to that is, right now, probably, no. Unfortunately, I only have three more power supplies that we have reviewed over the years that I have retained and are still functional so there are not a lot of options to do this with regularly.
Click to expand...

Any chance we could learn which psu's have died over the years? And what their lifespan was?

Myself, I had an Enermax Galaxy 1000w unit that I got 4 years, 4 months out of. Running another enermax 1350w unit since 7-2011, so far so good. My units are on pretty much 24/7.
 
That was a very interesting article. I've got a couple of PSUs knocking about of around that age.

However, I would have liked to have seen much more info on how the device had been used on the first page. In particular how it had been used in the past year. This would have given a better background to the rest of the article. It would be too much to expect exacting detail, but whether (and how long) it had spent time unused, was it used in a high-stress system, a low-stress one, or what.
 
cyzzledyzzle said:
Based on my hobbies, I'm unsurprised that a power supply containing vented electrolytic capacitors has degraded and presents more ripple. The electrolytic capacitors are essentially perishable products, and seeing more ripple with age is very common in even high quality electrolytics. I'm guessing about how much the supply was used (guessed "a lot") but they're definitely getting into the "spontaneous failure" period of their existence. I've replaced so many electrolytics over the years if I get something with old ones at this point it's faster and easier just to replace them than to test them.

This PSU seems to have held up fairly well imo.
Click to expand...

I would be very interested in seeing the caps replaced in this power supply & retesting it.
 
Makes me wonder if my random lockups aren't video card related anymore. My TX1000 is definitely getting long in the tooth.

Every now and then when I am gaming, say a little over a couple hours at a time, my screen will randomly black out and my monitor will act like it's not getting signal anymore. Other times, the screen will do that, but then it will come back and I just have weird stripes for a screen.

The blacking out happened on a previous build and my current one. The striping is new-ish and happened with my current build.
 
I wonder how much performance you would gain back by just replacing all the electrolytic capacitors :cool:
 
Meaker said:
I wonder how much performance you would gain back by just replacing all the electrolytic capacitors :cool:
Click to expand...

That is an interesting question. If you have the soldering know-how, would it be worth it to replace the caps on an old high end PSU instead of buying a new high end PSU? Especially with modern power supplies having very little room to get better nowadays.
 
CaptNumbNutz said:
On the Conclusion Page:


I'm thinking it's safe to assume this unit got near daily use for 7 years straight and fairly heavy use for several of those years when it was in his personal desktop and file server.
Click to expand...

Yes, but that information needed to be on the front page, not the the last!
 
Awesome article. It would be cool if the unit could get some forensic tear down to understand where the degraded perf originated. Old caps? Who knows?
 
This is very eye-opening. Thankyou Kyle for doing these tests.

I was considering replacing my PSU which is a Corsair HX520 from 2008, when I do my next refresh, but was kind of on the fence, considering Corsair's quality and how stable my system has always been since I set it up in 2008.

I am now leaning towards replacing my PSU....which is a damn shame.

I've only had to do 2 upgrades over the years (videocard, and CPU).....well 3 if you count the SSD add/upgrade. And I've been putting a full upgrade off for quite a while, since my system still runs pretty well. However, when I tried a newer game such as Bioshock Infinite.....I couldn't play it in 1080p, so it is time to upgrade again.
 
Very interesting article. I'd love for the H to do more of these. Especially regards to overclocking long term effects.
 
Quartz-1 said:
Yes, but that information needed to be on the front page, not the the last!
Click to expand...

When you do the next power supply review on your site, please make sure and link me in so I can read it. ;)

ianken said:
Awesome article. It would be cool if the unit could get some forensic tear down to understand where the degraded perf originated. Old caps? Who knows?
Click to expand...

It would be nice, but I just am not going to spend the money and time on it. Our guess would be caps.

trick_m0nkey said:
Very interesting article. I'd love for the H to do more of these. Especially regards to overclocking long term effects.
Click to expand...

So would I.
 
Loved this article. I know it takes time and resources that are often consumed by reviews of new products, but I love seeing articles about old products and how they stack up in the modern world. It really helps inform people better about where their old system stands and if its worth it to upgrade. This look back at an old PSU, while not exactly scientific (as noted multiple times in the article), is very informative for [H] users who are probably wondering how long they should keep these things... especially with high end PSU's shipping with 7+ year warranties these days
 
Tsumi said:
That is an interesting question. If you have the soldering know-how, would it be worth it to replace the caps on an old high end PSU instead of buying a new high end PSU? Especially with modern power supplies having very little room to get better nowadays.
Click to expand...

My semi-educated guess (I'm an electrical engineer) is that you'd probably be able to get back most of the measurable performance losses by doing this, and it might be a worthwhile pursuit if you have a really high quality PSU and are comfortable working with high voltage electronics. It would be really interesting to see someone test this out, I'd try it myself if I had a suitable PSU for a test subject.
 
I've got an old BFG 1200 EX PSU in my main rig. It is from 2009 and was used when I bought it.

I have been suspecting that it is starting to have issues, and I have actually had it apart once and found that the filter cap on the power input plug had come loose due to the solder cracking. It would crackle every time the plug was removed or inserted into the power supply.

One of the ends on one of the modular SATA power cables has also gone bad. That was not fun to figure out.

Been contemplating replacing it with a newer power supply and expect my system would be less quirky if I did so.

If I could get the correct caps for a good price, I would probably try recapping it first to see if it would help, but all the higher end caps seem to be pretty pricey if ordering them and there is no place local that I could even buy them from.

The other problem is finding caps that are not only the correct rating, but the same diameter. There have been quite a few power supplies that I have replaced blown caps on and I usually end up having to shoehorn different size caps in them because the originals were way overrated.

I have even had to run jumper wires from the board to a cap or two in some power supplies because there is no way a replacement cap that will actually do what it is rated for would fit in the original space.
 
Interesting article.

I think one of Paul's notes at the end is really important.

n addressing that simple curiosity, and our observations, a few caveats should be noted. While this was a fun look at what has happened to one particular unit over time, this observation is of a completely uncontrolled design and produces only an observation not scientifically rigorous data. Why? Well, we had only a single Olympia 1000W, not the 3 minimum for statistics or the desired 10 for some rigor. I have no idea what the actual loads were over the years on the unit or how long in hours it has actually run, but probably "a lot." While this unit was running the environmental and electrical conditions (temperature, humidity, AC input, etc) were also uncontrolled in the same manner and to the same extent that we do in our load testing. There is no way to actually say if this unit has experienced a representative form of "normal" usage or not

I don't say this to insult Paul; the article was really useful. But the other thing I would note is that the way a power supply ages may be due in part to the OEM too. Who is to say whether Seventeam's design would age the same as Seasonic, Channel Well, or Super Flower? Probably the one thing I would be willing to posit is that lousy OEMs units age worse than quality ones.

Environment can be a factor too. I built pretty hardy systems for relatives going to the Philippines. Despite high quality components, sometimes even with gold-plated connectors on some, the salt air of the Pacific caused corrosion that aged electronics in an incredibly short time.

I'd love to see how tests like this would fare from multiple OEMs over time to see if there's a difference between one over another. I agree that all PSUs age. I'd just like to see the degree to which some do more than others.

Thanks, Paul.
 
Agreed. I have specc'ed many *very* high-end power supplies in gov't systems (up to 50 kw), and have had to plan periodic capacitor replacement as part of the logistics support. All electrolytic filter caps wear out- the more ripple current, the faster the wear. It's really a big deal that this supply lasted over 7 years.
 
Nice article!

I have a SeaSonic S12-600 from 2006 or 2007 (can't remember :D). I've used it through a number of builds, all overclocked. I replaced it with a beQuiet! Straight Power E10 600W couple of weeks ago not because of performance problems but because it was too loud. It was never too quiet and with age surely never got quieter.
I did some stress testing (furmark + prime) just before replacing it and it was happily pulling 440W off the wall. It's retired now. Maybe I'll keep it as a emergency spare? Dunno, it might not be safe after a couple years more.
 
I have the original Silverstone 300W Nightjar which i first built into a PC...........

2014 for my 5820k (Silverstone 750 Gold GS modular)
2013 for my i5 4400 (Silverstone 600 Silver Modular)
2010 for my 1090T (Silverstone 750 Bronze Modular)
2008 for my Q6700 (Silverstone 500)
2006 for my E6600

.......... over eight years ago with a cramped Silverstone LC16M case along with a Core2Duo and 7950GT passive.

and i have just rebuilt that PC using the same case, and the PSU is still going great guns in spite of the ceaseless thermal and electrical abuse.

all of my Silverstone PSU's are working fine.

two rules when building I work to:
1. buy a quality motherboard
2. buy a quality PSU
 
Very interesting article. Where was the load failure point more precisely -- 800, 850, 900, 950, 1000W? For instance, if it still does 900-950W that's great imho, but if it's puking at 725W, not so great.
 
Tsumi said:
That is an interesting question. If you have the soldering know-how, would it be worth it to replace the caps on an old high end PSU instead of buying a new high end PSU? Especially with modern power supplies having very little room to get better nowadays.
Click to expand...

The thing is that the capacitors aren't the only thing that goes bad. When the capacitors start to fail, they cause other components around them in the circuit to work harder from the increasingly bad ripple and voltage regulation. Eventually components will get so hot that they'll start an infinite loop of more heat and more resistance until something burns/explodes. If you open up your PSU and notice that the PCB has brown/black spots on it, you know there's a problem.

So if you want to recap a unit, you need to catch the capacitors when they start going bad, which is hard if they don't start bulging from the top/bottom. I've recapped dozens and dozens of units over the years, and the ones that didn't have failure of other parts have worked like a charm for years after the capacitor swap.

If you don't factor in your labor cost in time, the parts to recap your average PSU generally run between $15-$30 from Mouser. If the PSU had polymer caps then all bets are off because those are damned expensive. A single cap can run $5.

The units that I recap the most are Antec PSUs. For years they used those garbage "Fuhjyyu" Taiwanese caps with the stolen electrolyte formula. The earliest I've seen was a unit from 2004 and I still occasionally come by Antec units with a couple of those in them. Antec units are also the worst to recap because they solder on big daughterboards at funny angles that are an absolute pain to work around.
 
xorbe said:
Very interesting article. Where was the load failure point more precisely -- 800, 850, 900, 950, 1000W? For instance, if it still does 900-950W that's great imho, but if it's puking at 725W, not so great.
Click to expand...

Well we ran the exact same tests that we ran ~7 years ago so it was somewhere between 800W and 1000W. It would start at 1000W at 120v AC input but it would not run the test.
 
Kyle, it is very important that you realize the importance of providing customers with the performance of your PSU. It is important that you guys also take into account the various environmental considerations while logging all this data as well. For example if the PSU is used in an environment with high humidity values and high temperature, the capacitors might not last as long as they would in a cold environment and that is true for the performance of the PSU in general as well. I’m very interested to what you guys conclude.
 
Last edited by a moderator:
jackperez said:
Kyle, it is very important that you realize the importance of providing customers with the performance of your PSU. It is important that you guys also take into account the various environmental considerations while logging all this data as well. For example if the PSU is used in an environment with high humidity values and high temperature, the capacitors might not last as long as they would in a cold environment and that is true for the performance of the PSU in general as well. I’m very interested to what you guys conclude.
Click to expand...

That is not going to happen for a product in use for 7 years. Sorry.
 
Interesting review!

Have you noticed that companies like Seasonic and EVGA are now offering 10 and 12 year warranties on their flagship PSUs?
For instance: 12 year warranty
https://seasonic.com/prime/


Maybe buy a unit now. Test it, install it on a file server, and re-visit in 2027.

If we haven't been blown up by Iran or North Korea, or by some other random world event, I'd be interested in seeing the results.
 
You must log in or register to reply here.
Back
Top