XFX Black Edition 850w ATX PSU
Simulated Load Testing
Published: 12th October 2009 | Source: XFX | Price: £159 |
Simulated Load Testing
To provide accurate and consistent results in all of our PSU testing, Overclock3D uses professional grade DC electronic load equipment capable of placing a sustained load of 3690w across a total of six rails (including +5vsb and -12v) on the PSU! This is achieved by using a combination of SunMoon and Analogic electronic load equipment which allow us to adjust amperage loads in increments as small as 0.01A while also measuring voltage and wattage readings on-screen.
During today's tests, we will be placing the XFX Black Edition under 25%, 50%, 75% and 100% load levels inside a hot box regulated to a temperature of around 50°C. Additional 'Cross Load' and a 'Max Load' tests will also be performed under these conditions to simulate how the PSU reacts to heavily uneven loads as well as running above its specified output.
XFX Black Edition 850w Results @ 50°C | ||||||||||
| +3.3v | +5.0v | +12v | +5vSB | -12v | AC Watts / DC Watts | Efficiency | Intake / Exhaust | Δ Temp | ||
| Test 1 (Low) | 4.37A | 4.37A | 14.00A | 0.75A | 0.20A | 242w / 213w | 88.01% | 50.4°C / 47.4°C | -3.0°C | |
| 3.40v | 5.12v | 12.16v | 5.00v | -12.05v | ||||||
| Test 2 (Med) | 8.75A | 8.75A | 28.00A | 1.5A | 0.40A | 471w / 425w | 90.23% | 51.4°C / 55.4°C | 4.0°C | |
| 3.39v | 5.10v | 12.10v | 4.96v | -12.06v | ||||||
| Test 3 (High) | 13.12A | 13.12A | 42.00A | 2.25A | 0.60A | 705w / 636w | 90.21% | 49.5C / 55.7°C | 6.2°C | |
| 3.38v | 5.08v | 12.06v | 4.93v | -12.11v | ||||||
| Test 4 (Full) | 17.50A | 17.50A | 56.00A | 3.00A | 0.80A | 948w / 849w | 89.55% | 51.0°C / 59.9°C | 8.9°C | |
| 3.37v | 5.07v | 12.00v | 4.89v | -12.17v | ||||||
| Test 5 (x-load) | 17.00A | 16.00A | 1.00A | 0.00A | 0.00A | 182w / 151w | 82.96% | 51.6°C / 49.3°C | -2.3 °C | |
| 3.39v | 5.11v | 12.14v | 5.01v | -12.60v | ||||||
| Test 6 (x-load) | 1.00A | 1.00A | 70.00A | 0.00A | 0.00A | 939w / 850w | 90.52% | 50.8°C / 59.3°C | 8.5°C | |
| 3.40v | 5.14v | 12.02v | 4.98v | -14.33v | ||||||
| Test 7 (MAX) | 22.00A | 22.00A | 61.00A | 4.00A | 0.80A | ----w / ~950w | --.--% | --.--°C / 52.0°C | -.-°C | |
| -.--v | -.--v | --.--v | -.--v | --.--v | ||||||
XFX Black Edition 850w Performance Overview | |||||||
| +3.3v Diff. T1-T4 | +5.0v Diff. T1-T4 | +12v Diff. T1-T4 | +5vSB Diff. T1-T2 | -12v Diff. T1-T2 | Avg Effic. T1-T4 | Noise Rating | |
| 0.88% | 0.97% | 1.31% | 2.20% | 0.99% | 89.50% | Low | |
Before we get to the big red part of the chart I first want to draw your attention to the standard tests 1-4. These best reproduce the kind of load that the PSU will receive when installed inside a PC. As we can see from the results and indeed the performance overview chart above, the XFX Black Edition 850w has some of the most stable rails I've ever seen in any PSU tested before on OC3D. The +3.3v and +5v rails are especially solid with less than a 1% drop in voltages from a low to full load which even bests our current favourite PSU, the Corsair HX850w. Efficiency was also extremely good with a high of 90.23% being seen at a 425w load in test 2 and an overall average of 89.50%.
However, all is not well. Part of OC3D's PSU testing procedure involves a "MAX Load" test in which the PSU is placed under a continually increasing load which takes it outside of its rated specification. This not only identifies the true capabilities of the PSU (for example if it is a 750w PSU overclocked to 850w, or a 1000w PSU underclocked to 850w) but also ensures that the OCP (Over Current Protection) is fully functional on the unit. Unfortunately this is where the XFX failed in a BIG way.
During this test the load was increased to 22A on both the +3.3v and +5v rails and 62A on the +12v rail. This combined with the small load on the +5vSB and -12v rail equated to around a 970w load on the PSU (120w above rated output). At first, everything appeared to be working fine. The ripple results were perfectly acceptable, the exhaust temperature was around 65°C and the voltages were all well within spec. Around a minute later the PSU powered off. Nothing unusual there, OCP had kicked in and saved the day. To find out whether the unit would be capable of sustaining a slightly lower load, the +12v rail load was reduced to 61A and the PSU powered back on. What happened next warranted a change of underwear...
An EXTREMELY loud BANG (heard even two floors down!) was emitted from the PSU accompanied by a bright flash. Yes, the XFX Black Edition had failed in a potentially dangerous way. Luckily none of the OC3D equipment was damaged by the elaborate death of the XFX PSU, but it is extremely worrying that the OCP protection on the PSU didn't kick in for a second time. Of course this could be put down to the unit being from a faulty batch, but as the unit tested in the video was the 2nd one to suffer this exact fate at the hands of our MAX LOAD tests, something does indeed seem wrong.
XFX Black Edition 850w Scope Results @ 50c | ||||
| +3.3v | +5.0v | +12v | ||
| Test 1 (Low) |  |  |  | |
| Test 2 (Med) |  |  |  | |
| Test 3 (High) |  |  |  | |
| Test 4 (Full) |  |  |  | |
| Test 5 (x-load) |  |  |  | |
| Test 6 (x-load) |  |  |  | |
| Test 7 (MAX) | ||||
Finally we come to the +3.3v, +5v and +12v ripple results as recorded using a Rigol 25Mhz 400MSa/s oscilloscope during the load tests further up the page. Once again, the standard test 1-4 are exceptional with the ripple hitting a high of 30.4mV on the +3.3v rail, 22.4mV on the +5v rail and 36mV on the +12v rail. Even the cross-load results come back all-clear with a maximum ripple of 60.8mV on the +12v rail during test 6. Of course, the big let-down is once again Test 7 where I didn't have a chance to record any results before the unit met its untimely demise.
Now let's move on to the conclusion where I try to sum up everything seen today...
Most Recent Comments
We have heard nothing back from eitherside yet, I will try and get some information next week for you dude. All we know for now is it was a big concern, and XFX had decided to do their own testing with Seasonic the OEM manufacturer.
Well after wondering what to do regarding this PSU, I decided to contact XFX personally and this is the entire history from my conversation with one of there technicians, over the course of yesterday and today.
[ 7/11/2009 8:02:39 PM] I have read that when tested this power supply blew up and that it was a major issue for you at the time of the review and that you were looking into it! The review was read on overclock3d.net and has been discussed at length here hxxp://forum.overclock3d.net/showthread.php?p=354173#post354173 Can someone please get back to me and let me know what has been done about the stability of this psu (to stop further potential problems) and if the problemthat caused the psu to blow up has now been rectified?
[SHAUN_T 9/11/2009 11:38:10 AM] Hi, thanks for your messages. Please note, it was the weekend when you posted your support ticket, therefore there were no technicians available to answer your question at the time. There have been many XFX PSUs sold recently and the customers have had no issues whatsoever, so there are no known faults or issues with the product. Have you tried installing it into your system? If you have any problems with it, please do let us know. Kind regards, Shaun.
[ 9/11/2009 1:28:23 PM] No I haven`t tried installing it in my system as yet because of the review I read, and i`m not sure whether or not to just rma it for a refund because, now I`m not sure if they deliberately overstressed the psu or not but they are claiming that the protection that`s supposed to be built into the psu to stop it blowing up or exploding only kicked in once and the 2nd time it blew up and they claim it happened to 2 out of 2 of the psu`s, can you do me a favour and read the review via the link I posted then report back to me what you make of it so I can have some sort of peace of mind and a genuine honest opinion from yourself. The review is located here hxxp://forum.overclock3d.net/showthread.php?p=354173#post354173 Thank you
[SHAUN_T 9/11/2009 2:54:49 PM] Hi, thanks for your message. Like I said, there are no common faults/batch issues etc with the XFX PSUs, we have had many customers using them without a problem. I have scanned over the review vaguely but we cannot pass much comment as we do not know the system setup etc etc. Why they had these problems I do not know, but there isnt much more I can do than to assure you there are no problems with the PSUs. Kind regards, Shaun.
[ 9/11/2009 3:30:14 PM] So is it wrong to assume that you knew about this issue because they mentioned they had informed you (well not you personally) but XFX about this issue and that you knew about it, and that XFX were very concerned about this and were going to run some tests to verify there was no problem because you seem to be side stepping the issue completely instead of assuring me the units had been tested after that review and that those 2 out of 2 units must have had a fault. You have done nothing to ease my mind as you have clearly stated that there are no common faults with xfx psu`s, but how would you know if you hadn`t done some tests after that review quite clearly stated you had been informed and they had no response from xfx regarding the issue. Thanks for replying.
[SHAUN_T 10/11/2009 11:36:55 AM] Hi, thanks for your message. If there was an issue with the PSUs, for example, a batch issue they would have been recalled. Obviously, other users have had no issues with this product, and we have not had any support tickets complaining about anything. Therefore, as far as we know there are no problems with these PSUs. If you do not wish to use the PSU, it may be best for you to return it to your reseller whilst you only bought it 4 days ago. Kind regards, Shaun.
[ 10/11/2009 12:26:07 PM] I just figured if they did let you know then maybe you would have done your own tests to either substantiate there claims or prove them to be false, I haven`t made my mind up what I`m going to yet but I do appreciate your replies and you seem to have been fairly honest in your responses. Thank you.
[SHAUN_T 10/11/2009 12:27:23 PM] Hi, thanks for your message. No problem, we`re not here to lie to you, so if you do have anymore questions or encounter any problems in the future, please let us know. Kind regards, Shaun.
whilst they are not lying, they are doing what all media people are trained to and only put forward arguments they appear to answer the question but are actually sidestepping. I have to say it was very interesting to read their response.
Seems that they would rather have people blown up before doing anything - which includeds tests themselves.
Seems that they would rather have people blown up before doing anything - which includeds tests themselves.
I would not worry you will never reach the cut out point with your rig, safe as houses mate.
Hi, thanks for your message. No problem, we`re not here to lie to you, so if you do have anymore questions or encounter any problems in the future, please let us know. Kind regards, Shaun.
Frankly I am not surprised that XFX haven't received many customer complaints or particularly high fairlure rates. The Power Supply Unit should function up to and including the rated power output. Jim's testing has proven however that once pushed past it's limits the overvoltage protection is nigh on non existant and so the Black Edition 850W reacted in this way.
For a start, I can only think of a small list of typical system configurations that would even push a PSU to 850W let alone higher and so chances are (at this point in time) even those with the upper end systems aren't suffering from any problems. This however does change as time goes by. By nature, a PSU like any electronic device will degrade over time and in theory that threshold at which this unit will go bang will lower accordingly. Even if I'm being pedantic because the rate of degradation could be argued to be low, it's the principle of it. Would you expect a £160 unit to be capable of shutting itself down before it causes itself or other components damage? Based on Jim's testing it seems as though it cannot.
Whether your system commands 850W and over is irrelevent. If the unit cannot save itself once pushed too far, why settle for it? I don't for a moment find this acceptable.
(...)
Whether your system commands 850W and over is irrelevent. If the unit cannot save itself once pushed too far, why settle for it? I don't for a moment find this acceptable.
Those were my thoughts exactly. A PSU with this kind of price tag should not have this kind of failure. It's intolerable. That's why I returned my XFX 850w (without even using it once) and brought home a Corsair HX-750.
Those were my thoughts exactly. A PSU with this kind of price tag should not have this kind of failure. It's intolerable. That's why I returned my XFX 850w (without even using it once) and brought home a Corsair HX-750.
TBH I think you did the right thing, I love my HX850
I dont like the green fan , and the PSU case look strange ugly , and to be honest the modular feature is not necessary because I have a case with holes to hide the cables at the back of the mobo tray , this is the reason I bought my lovely PC Power & Cooling 910W
I dont like the green fan , and the PSU case look strange ugly , and to be honest the modular feature is not necessary because I have a case with holes to hide the cables at the back of the mobo tray , this is the reason I bought my lovely PC Power & Cooling 910W
They are your own preferences mate. :p
I for one love the green fan, the PSU case and would love a modular PSU!
I think XFX are clearly denying the issue until enough genuine end-users have or complain about this problem, which could be a while.
It's a shame, as the PSU appears to be very good except for that, and the styling is different, and would undoubtedly sell well in the enthusiast market.
What would be best for XFX at this point would be for them to acknowledge the issue and offer customers an RMA for a revised model (if indeed one has been made) if they wish. This would save their reputation (in this new field to them) somewhat, and perhaps they will produce some more good quality PSUs in the future, instead of being forced out the market for a small issue. More competition is usually a good thing for us customers.
As it stands, I feel like the company is burying their head in the sand somewhat and sidestepping the issue as already said. This is pretty much what I would expect from a big company of their type, but they have an opportunity to now prove us wrong.
I do feel that the tech support chap was being honest though. As far as he knows there probably never was an issue brought to his attention, and he was just doing his job (correctly!).
All in all, I wouldn't buy one until this is addressed or recognised, as it gives me little faith in the brand, or indeed the RMA procedure (vital with PSUs in my experience). Swap it for a Corsair and you won't be disappointed (and I say that after having two fail on me, but that's a different story).
It's a shame, as the PSU appears to be very good except for that, and the styling is different, and would undoubtedly sell well in the enthusiast market.
What would be best for XFX at this point would be for them to acknowledge the issue and offer customers an RMA for a revised model (if indeed one has been made) if they wish. This would save their reputation (in this new field to them) somewhat, and perhaps they will produce some more good quality PSUs in the future, instead of being forced out the market for a small issue. More competition is usually a good thing for us customers.
As it stands, I feel like the company is burying their head in the sand somewhat and sidestepping the issue as already said. This is pretty much what I would expect from a big company of their type, but they have an opportunity to now prove us wrong.
I do feel that the tech support chap was being honest though. As far as he knows there probably never was an issue brought to his attention, and he was just doing his job (correctly!).
All in all, I wouldn't buy one until this is addressed or recognised, as it gives me little faith in the brand, or indeed the RMA procedure (vital with PSUs in my experience). Swap it for a Corsair and you won't be disappointed (and I say that after having two fail on me, but that's a different story).
CSS Valid
Which yeah technically is only a refund, but companies like to take the long way round (I guess) as it has to be rma'd first which is the point I was trying to make, even though technically it's not an rma because i'm returning it to the reseller not the manufacturer of the psu.
Thanks for your replies