Problems with the second Corsair Rm1000X

[Spytek]

Hi All.

I apologize for the bad English used but I'm Italian

Kindly, I need your advice because it is the second Corsair Rm1000X power supply that is giving me problems.

The first has been replaced for excessive coil whine and the second has low voltage on both the 12V and 3.3V lines.

By default the power supply delivers 11.980V and 3.296V; under load these voltages drop to 11.78V and 3.26V.

My system is composed by Intel I7 9700K, Ram 16 gb, Asus Z390 Strix, Evga 1080 Ti, SSD 2x 1 Tb Corsair Pro Raid 0.

I opened a ticket to Corsair to report the problem, pointing out that it is very strange to have low voltages at default and that the first Corsair Rm1000x (replaced for excessive coil whine) had higher voltages (12.1V and 3.3V at default). Corsair replied by telling me that it is normal to have 3% voltage oscillation and I was invited to perform another test with another power supply by providing a video with the results obtained.

Honestly, given the product category, I am disappointed to see already low starting voltages that go down to 11.78V and 3.26V under load. What would happen if I wanted to combine a second 1080Ti in Sli?

But what surprised me much is that the RMA Corsair department did not take caring that it is the second power supply that does not fully meet my expectations, limiting itself to having me run a test with another power supply rather than replacing the Corsair Rm1000X.

Anyway, I had to disassemble an Antec Eag 750W from my second system and test it on the main one.

First of all, the Antec power supply gives me much better voltages by default, that is:

12.172-12.076V and under load 11.980V; 3.31 at default and 3.296V under load.

I am surprised that this PSU, which costs much less than the Corsair, provides me with much better performance but I am convinced that my specimen does not provide the necessary performance for its category: 11.78V and 3.26V are as if they were over 100% load in spite of a maximum absorption of around 500-600W of my system ...

I sent all the results to Corsair and I sincerely hope that they replace the power supply because, I repeat, I am not at all satisfied and I consider it improper and unkind that the RMA Corsair department has not provided the right attention because it is the second unit that does not provide a perfect functioning and I believe that the quality control alone is not at the highest level.

At this point, in your opinion, why does my Corsair Rm1000x supply lower voltages than normal?

Thank you for your availability and attention and I apologize for the bad English language ;):

[LeDoyen]

well those voltages aren't bad. They are within the ATX spec

12V should be between 11.4v and 12.6v (+- 10%)

3.3V should be between 3.135v and 3.465v (+- 5%)

 

So it will be hard to justify replacing it since it works fine.

 

Now what can make your voltages drop quite a bit is if you use extension leads for the power cables. Some have thin wire gauges that will cause voltages to drop quite a bit under load.

[Spytek]

well those voltages aren't bad. They are within the ATX spec

12V should be between 11.4v and 12.6v (+- 10%)

3.3V should be between 3.135v and 3.465v (+- 5%)

 

So it will be hard to justify replacing it since it works fine.

 

Now what can make your voltages drop quite a bit is if you use extension leads for the power cables. Some have thin wire gauges that will cause voltages to drop quite a bit under load.

Hello and thanks for replying.;):

I know the Atx specifications but the voltages of the first Corsair Rm1000X (replaced in rma for internal noise) had much better voltages, in line with the category of the product rather thantis second supply.

Anyway,I can't explain how it is possible that a 750W Antec performs much better than the Corsair Rm1000X.

I hope they will replace it because if I decide to use Sli mode, I would not want to worry about possible system instability because the power supply may not hold the load. Then, once I have shown the Corsair technicians that the Antec 750W provides much better performance, it would be unfair of them to not replace the Corsair Rm1000X.

Consider that disassembling and reassembling the Antec 750w from the second system took several hours and I would not have wasted time unnecessarily as they asked me to test another PSU on the main system.

As concern the ATX specification,It was preferable for me to take a second Antec 750W, paying just over 100 euros, rather than spending 270 euros for the Corsair Rm1000X...

P.S. I do not use extension leads

Edited December 18, 2020 by Spytek

[LeDoyen]

I was mostly worrying about extensions. i used cheap ones when i started and lost 0.2V across the board, pretty terrible ^^'

 

Regarding the 12v rail, can you see in HWinfo what voltages you have on the GPU?

The latest versions show voltages coming from the PCIE socket and from the various PCIE cables you have plugged in (if the 1080Ti does it). Just to make sure it's not just the motherboard that reads low, and at least you'd know what the PCIE cables deliver since it's SLI that concerns you most.

[Spytek]

I was mostly worrying about extensions. i used cheap ones when i started and lost 0.2V across the board, pretty terrible ^^'

 

Regarding the 12v rail, can you see in HWinfo what voltages you have on the GPU?

The latest versions show voltages coming from the PCIE socket and from the various PCIE cables you have plugged in (if the 1080Ti does it). Just to make sure it's not just the motherboard that reads low, and at least you'd know what the PCIE cables deliver since it's SLI that concerns you most.

Sure ;):

Gpu Pci-Ex : 11.79 V

Gpu 6 pin : 11.94 V

Gpu 8 pin : 11.84 V

We are talking about a system that uses max 500-600W.

What about on 100% of load (Sly possibility)?

Edited December 18, 2020 by Spytek

[LeDoyen]

i doubt it would change at all. if you add a 1080Ti you'll still be well below the PSU's max rating.

voltages don't vary at all until the controller is at its maximum rating. and there's always some headroom. so even if you peak at 850 - 900w with a second card, you're fine.

 

Seems its only the motherboard that has some Vdrop. Anyway nothing uses 12v besides the PCIE sockets. All the rest uses VRM to step down so it's fine.

At full load, 11.84V on the PCIE is good.

 

I just loaded my GPU, powered by a HX1200i and that's almost the same voltages i got.

11.84

11.95

11.80

 

.. yet the motherboard reading is spot on at 12.00v.

 

So yea, the motherboard reading tells you nothing about what the GPU gets basically :)

[Spytek]

i doubt it would change at all. if you add a 1080Ti you'll still be well below the PSU's max rating.

voltages don't vary at all until the controller is at its maximum rating. and there's always some headroom. so even if you peak at 850 - 900w with a second card, you're fine.

 

Seems its only the motherboard that has some Vdrop. Anyway nothing uses 12v besides the PCIE sockets. All the rest uses VRM to step down so it's fine.

At full load, 11.84V on the PCIE is good.

 

I just loaded my GPU, powered by a HX1200i and that's almost the same voltages i got.

11.84

11.95

11.80

 

.. yet the motherboard reading is spot on at 12.00v.

 

So yea, the motherboard reading tells you nothing about what the GPU gets basically :)

Ok,perfect ;):

Under load,what is the voltage of your 12V and 3.3V?

I was thinking also a configuration with an Asus Rtx Strix 3080 and its bios of 380W worried me but I think it is possible to always stay on 11.78V with 100w of absorption of the Asus Rtx 3080 compared to my current 2080 which absorbs 280W.

Thanks a lot for your advice;):

[LeDoyen]

Running Valley benchmark with cinebench as i write that..

 

Honestly i wouldn't even look at the 3.3v rail. there's barely anything that uses that voltage, it's a relic of the 80's.

 

It stays at 3.26 under load.

 

the lowest reading i get on 12V are :

At PSU: 12.16

MB: 12.00

GPU PCIe: 11.8

GPU 6pin: 11.92

GPU 8 pin: 11.8

 

That's with sleeved extensions, beefy ones, but they still induce a little bit of drop. With my previous ones it was more like 11.6V across the board.

The rig is what i have on my profile's PC specs, everything overclocked. i peaked at 552W

 

If i bought Corsair sleeved cables to replace the default ones + extensions, i would probably gain a bit but it's not even close to causing troubles with overclocking.

 

I will also get a 3080 Strix whenever they appear from the aether, but with these it's not power usage you have to be wary of, but power spikes. Technically a 650W PSU would run one of these, but there's very short spikes of power usage that made it necesary for Nvidia to advise using 750W+ PSUs. Most corsair 850W can deal with it, so with a RMX1000 you also have a lot of headroom.

[Spytek]

Running Valley benchmark with cinebench as i write that..

 

Honestly i wouldn't even look at the 3.3v rail. there's barely anything that uses that voltage, it's a relic of the 80's.

 

It stays at 3.26 under load.

 

the lowest reading i get on 12V are :

At PSU: 12.16

MB: 12.00

GPU PCIe: 11.8

GPU 6pin: 11.92

GPU 8 pin: 11.8

12.16V under load, right? The argument is basic and if Corsair did not replace the power supply, it would not be fair to the customer. to say that a Ferrari does not reach 340 km / h but as long as it moves on the road ... Certainly I could have saved money by taking another 750w Antec since it costs less than half the Corsair 1000W. Consider that from the online reviews the value 11.81V was obtained at 100% full load; I have a system that as maximum load is around 50-55% and I get low voltages already at the start.I conclude by saying that if Corsair decides not to want to replace the power supply, it will be the last time I will trust this brand

[LeDoyen]

Remember we dont have the same PSU.

I don't know internally what the differences are.

 

What i know is the voltage shouldn't change significantly until you hit OCP. Most of the losses will be joule effect on the cables when you arrive at high currents (that's why high end PSUs measure voltage at the motherboard plug, to compensate for that). the Guru will be able to tell you if it's normal better than me talking out of my ***** :p

But to me, the PSU works just fine.

[jonnyguru]

Hi All.

 

By default the power supply delivers 11.980V and 3.296V; under load these voltages drop to 11.78V and 3.26V.

 

How are you measuring these voltages? An RM1000x does not support LINK or iCUE, so I'm not sure how you're getting accurate power measurements.

 

Are you actually experiencing PROBLEMS because of these voltages you're seeing or are you just assuming there is/may be a problem?

[jonnyguru]

12.16V under load, right? The argument is basic and if Corsair did not replace the power supply, it would not be fair to the customer. to say that a Ferrari does not reach 340 km / h but as long as it moves on the road ... Certainly I could have saved money by taking another 750w Antec since it costs less than half the Corsair 1000W. Consider that from the online reviews the value 11.81V was obtained at 100% full load; I have a system that as maximum load is around 50-55% and I get low voltages already at the start.I conclude by saying that if Corsair decides not to want to replace the power supply, it will be the last time I will trust this brand

 

Are you saying you cannot overclock with the Corsair, but can overclock with the Antec? Because if you're getting these voltage readings from the motherboard and are ASSUMING that they're preventing you from overclocking, you should know that motherboard voltage readings are completely bogus and are not reporting what is actually being provided to your components. For this, you should AT LEAST be using a digital multi meter.

[Spytek]

Are you saying you cannot overclock with the Corsair, but can overclock with the Antec? Because if you're getting these voltage readings from the motherboard and are ASSUMING that they're preventing you from overclocking, you should know that motherboard voltage readings are completely bogus and are not reporting what is actually being provided to your components. For this, you should AT LEAST be using a digital multi meter.

Hi. The voltages were not measured with a voltmeter but I got the same results on my second system. I'm not saying that I can't overclock with the Corsair but that the default voltages are below average. A power supply like the Corsair Rm1000X that at default delivers 11.98V and with load at 50% it drops to 11.78V means that it is not very efficient.The first Corsair Rm1000X that was replaced by internal noise, with the same system, provided me with much higher voltages. Excuse me Jonny, you are an expert: in your opinion, isn't it strange to see a power supply that by default delivers 11.98V and with 50% load drops to 11.78V? At 100% load, how long will it go down? Thank you for your availability ;):

[jonnyguru]

Hi. The voltages were not measured with a voltmeter but I got the same results on my second system. I'm not saying that I can't overclock with the Corsair but that the default voltages are below average. A power supply like the Corsair Rm1000X that at default delivers 11.98V and with load at 50% it drops to 11.78V means that it is not very efficient. The first Corsair Rm1000X that was replaced by internal noise, with the same system, provided me with much higher voltages. Excuse me Jonny, you are an expert: in your opinion, isn't it strange to see a power supply that by default delivers 11.98V and with 50% load drops to 11.78V? At 100% load, how long will it go down? Thank you for your availability ;):

 

No. Because you're looking at measurements made by an IC on the motherboard.

 

Nobody should be using this as a reference point. Even if they ARE experiencing problems.

 

But a multimeter if you're really concerned.

[Spytek]

No. Because you're looking at measurements made by an IC on the motherboard.

 

Nobody should be using this as a reference point. Even if they ARE experiencing problems.

 

But a multimeter if you're really concerned.

I have no problem in performing voltage readings with a voltmeter. But we are talking about readings on two motherboards with the same and identical values. I agree with you 100%, but when the readings are identical even on the second motherboard, do you have any doubts? Anyway, I repeat, I will rummage further doubts with the measurements of a voltmeter. Anyway,is it normal for you that a 1000W gold certified power supply has low voltages already at default and that at half load (500-550w max of my system) it supplies a voltage as if it were at 100% load (1000W)??

I think probably not :rolleyes:

With the online reviews of the Rm1000x you can note that by default it starts at over 12V and drops at 11.80-1178V ONLY at 100% Load (1000W).

That is the same value that I have with my system at half load. If this is normal, I am perplexed...

https://www.techpowerup.com/review/corsair-rm1000x/5.html

As I wrote several times above, my aim is to be able to feel comfortable in case I want to use a Sli configuration with two video cards. If I decide to combine a second 1080Ti that draws another 280W, are you sure that the system would be stable? If so, how can you say this certainty given that for me the basic values ​​are already low and not representative of the category to which they belong?

Edited December 21, 2020 by Spytek

[LeDoyen]

@Jonnyguru, it's a single rail PSU sooo.. i believe the motherboard and PCIE come from the very same regulator right?

 

like, fat common bus going to all 12v outputs, voltage reading on the 24 pin for regulation?

That's how i understood it, if you can enlighten me on that ^^

 

If that is so, the readouts are even more bogus, because the bulk of the load is on the PCIE cables but it's the motherboard reading that is the lowest.

 

in an ideal world, they should all be almost identical at idle.

 

Fun fact : at idle, the reading at the PSU from my HXi is 12.18v. at the motherboard it reads 12.192.

 

I achieved surgeneration :)

 

bottomline : We don't know what we have. only a calibrated DMM will tell you what the real voltages are.

[jonnyguru]

@Jonnyguru, it's a single rail PSU sooo.. i believe the motherboard and PCIE come from the very same regulator right?

 

Correct. ALL +12V POWER comes from the same source within the PSU. So, ALL +12V would be the same as it leaves the PSU. Variables such as cables, connectors, trace length on the PCB, etc. all are capable of making those numbers at the motherboard completely random. This is why I find BIOS voltage readings no better than using a random number generator.

 

I have no problem in performing voltage readings with a voltmeter.

 

Good. Do it.

 

But we are talking about readings on two motherboards with the same and identical values. I agree with you 100%, but when the readings are identical even on the second motherboard, do you have any doubts?

 

Yes. Of course I have my doubts.

 

As I wrote several times above, my aim is to be able to feel comfortable in case I want to use a Sli configuration with two video cards. If I decide to combine a second 1080Ti that draws another 280W, are you sure that the system would be stable? If so, how can you say this certainty given that for me the basic values ​​are already low and not representative of the category to which they belong?

 

Because that's a very typical behavior for motherboard voltage readings.

 

Here's somethings you can do if it makes you feel better:

 

Are you ONLY using the cables that came with the PSU? No aftermarket and no extensions?

 

Unplug and replug all cables to the motherboard. Make sure they are fully seated. While unplugged, check the pins for any kind of damage.

 

And keep in mind this:

 

Your GPU power comes from the PCIe cables. Your CPU power comes from the 8-pin EPS12V cable. The motherboard is getting it's "voltage readings" from the 24-pin. The 24-pin has only two +12V leads. So if the voltage readings are different between the PCIe, EPS12V and 24-pin, it's because of one thing and one thing only: There is resistance somewhere.

Edited December 26, 2020 by jonnyguru

[LeDoyen]

just to add to this, i removed my extensions last weekend and installed a set of Cablemod leads.

All voltages went slightly up. All the GPU readings are at or slightly above 12v, the PSU 12V supply stabilized dramatically now only showling very slight swings around 12.2v instead of the wide variations i had before because of added resistance... and the motherboard reading is still above what the PSU sends.

 

So, the Nvidia sensors seem to be closer to the truth (and there's 3 of them, roughly at the same voltage), and motherboard sensor is definitely out of whack :)

 

At idle :

PSU 12V supply : 12.188v

 

different readouts at the receiving end (with cable/pin contact losses etc):

 

GPU FBVDD input : 12.113

GPU PCIe :12.111

GPU 6 Pin : 12.112

GPU 8 pin : 12.089

GPU Misc2 input : 12.111

 

Aquacomputer Octo fan voltages (basically what the unit receives, they are PWM headers)

Fan 1 (3 fans): 12.100

Fan 2 (3 fans): 12.100

Fan 3 (2 fans): 12.110

Fan 4 (1 fan): 12.110

 

Motherboard reading : 12.192v

 

That mobo reading, although visually pleasing over 12v, is completely off.

 

That's for peace of mind :)