My Corsair HX520W is my loudest component...

Alex · Post by **Alex** » Sat Jun 02, 2007 4:38 am

Well my HX620 is the noisiest component in my non completed rig.
When it is really silent at night (1 o'clock) you can clearly hear that my rig is on anywhere in the room. At the same time when lying on my bed I hear my own breathing as loud.
In a very silent environment you seem to be able to hear almost any moving parts.

How well you hear noise seems to be a question about how silent your environment can become. During daytime it is harder to notice if the rig is on (more distant traffic outside and other noises).

I barely made the HDD WD5000YS less noisy than my PSU with the help of a softmounted scythe quiet drive in the lower bay of my P180B. At seeks the HDD is possibly louder than the PSU.
I have written a post about my silencing experience here
viewtopic.php?p=335645#335645

The rig is not completed but I have 2 Scythe S-Flex D undervolted by my ASUS Commando board in it and they are less noisy than the PSU too.

When I put my ear against the S-Flex D motor I can clearly hear a "ticking" sound. I guess this is because of PWM switching on the mobo? Since I can not hear the ticking under any normal conditions I don't care much about it but still as a SPCR fan you get somewhat annoyed by this.

Aren't we picky bastards here.

I understand why some people buy fanless PSU. During daytime use in a suburb or town it might be less useful though.

Khrono Devil · Post by **Khrono Devil** » Sun Jun 03, 2007 8:53 pm

Either you guys have a faulty PSU or a very quiet rig. In early morning I can hear it but certainly isn't loud. Compared to all the other PSUs I have used it is by far the quietest even when the load is on.

Erssa · Post by **Erssa** » Sun Jun 03, 2007 10:44 pm

Khrono Devil wrote:Either you guys have a faulty PSU or a very quiet rig.

If the computer fills "spcr standards" psu will most likely be the loudest component.

Every self-respecting spcr enthusiast has his fan(s) running at less then 1000rpm, most likely in the range of 600-800rpms. S12, Corsair etc have adda fans. Adda's are "ok", but not as good as Nexus, Scythe or Noctua, even if they were running at the same speeds, which they won't doing, since S12 and Corsair both have their fans running very close to 1000rpm even at the lowest speed.

It's very easy for the psu to be the loudest component, it doens't necessarily mean, that the computer itself is loud. I think I will try passive psu next time I'm building a computer.

nutball · Post by **nutball** » Sun Jun 03, 2007 10:56 pm

Khrono Devil wrote:Either you guys have a faulty PSU or a very quiet rig. In early morning I can hear it but certainly isn't loud. Compared to all the other PSUs I have used it is by far the quietest even when the load is on.

My main grumble is that the fan growls at the lowest revs (ie. when the machine is idle). Once I stick some load on it and the fan speeds up the sound becomes much more ignorable... but I don't really want to have to run my PC loaded all the time to stop it being annoying!

jaganath · Post by **jaganath** » Sun Jun 03, 2007 11:53 pm

My main grumble is that the fan growls at the lowest revs

could swap the fan but for such an expensive PSU I can understand if you feel you shouldn't have to.

nutball · Post by **nutball** » Mon Jun 04, 2007 12:21 am

jaganath wrote:
My main grumble is that the fan growls at the lowest revs
could swap the fan but for such an expensive PSU I can understand if you feel you shouldn't have to.

Fortunately I didn't have to pay for it so it's just the hassle, and the "here we go again" factor. But yeah if I had paid the Â£75 that Dabs are asking for it I'd be livid. Maybe I'm getting too fussy in my old age

Seems that even today if you want guaranteed quiet then it's best to buy a cheapo Seasonic and mod it with a Yate Loon or similar. Depressing.

jaganath · Post by **jaganath** » Mon Jun 04, 2007 1:26 am

Fortunately I didn't have to pay for it

wow! free corsair psus! where can I get one?!

Seems that even today if you want guaranteed quiet then it's best to buy a cheapo Seasonic and mod it with a Yate Loon or similar. Depressing.

I can see why this is how it is; the quietest fans are sleeve-bearing, but manufacturers do not want to put these in PSUs because of very reduced fan lifetime compared to ball-bearing; so we have to make that trade-off, reduced lifetime vs quiet. Although the fact that SPCRers tend to run their fans at reduced RPMs should slightly offset the reduced lifetime from using a sleeve fan in a PSU.

SnooP · Post by **SnooP** » Mon Jun 04, 2007 4:05 am

Also current trend is 120mm fans on the bottom - sleave bearings run noisier and at a much reduce life when mounted horizontally - ie like in a psu. Hence adda ball bearings being the popular quiet choice, though be interesting to see how 'hydrowave' or other more exotic bearings go.

Alex · Post by **Alex** » Mon Jun 04, 2007 4:19 pm

When my Corsair guarantee goes out (in 4.5 years) i was thinking of removing the Adda fan and put a Scythe S-Flex D or something in the lower chamber of the P180 (=Vertical fan).
EDIT: Sorry, Before editing I accidentally wrote =Horizontal Fan. I meant vertical mounting (instead of the originally mounted Antec Tri Cool in the P180B lower chamber).

My misstake.

With taped holes around the PSU on the back of the P180B case this should be OK shouldn't it? I will connect the new fan to the PSU controller of course (=some modding and soldering needed because I have heard the controller is a 2-pin header?).

I guess the original Adda is a 1600 rpm fan but in the lower bay there is no mobo (CPU, GPU) heat to take care of so it should be OK to swap for a 1200 rpm fan (Such as S-Flex D) instead. Noise should be lowered considerably then right?

What do you think of this? Any links to similar projects?

My PSU is not loud at all but if there is no other noise around (well withing reasonable practical limits) there is no problem to hear if my computer is on if you listen for it.

My fans are probably running ~700 rpm so as Erssa says

If the computer fills "spcr standards" psu will most likely be the loudest component.

Every self-respecting spcr enthusiast has his fan(s) running at less then 1000rpm, most likely in the range of 600-800rpms. S12, Corsair etc have adda fans. Adda's are "ok", but not as good as Nexus, Scythe or Noctua, even if they were running at the same speeds, which they won't doing, since S12 and Corsair both have their fans running very close to 1000rpm even at the lowest speed.

It's very easy for the psu to be the loudest component, it doens't necessarily mean, that the computer itself is loud. I think I will try passive psu next time I'm building a computer.

it is normal that my HX620 is the noisiest component of my computer. Glad to get it confirmed.

Thanks Erssa because I was beginning to wonder if my relatively quiet PSU was faulty after all. But now I am quite sure this is not the case.

And yes Khrono Devil I second your statement.

In early morning I can hear it but certainly isn't loud. Compared to all the other PSUs I have used it is by far the quietest even when the load is on.

SnooP · Post by **SnooP** » Mon Jun 04, 2007 10:33 pm

I guess the original Adda is a 1600 rpm fan but in the lower bay there is no mobo (CPU, GPU) heat to take care of so it should be OK to swap for a 1200 rpm fan (Such as S-Flex D) instead. Noise should be lowered considerably then right?

A s-flex may work well because its a fluid dynamic bearing fan. But not a sleave bearing like nexus as they aren't designed to do horizontal mounting and my guess will be louder than a good ball bearing mounted this way, though would be good if anyone has/knows of tests of this. See silentpcreview article:

http://www.silentpcreview.com/article690-page1.html

Erssa · Post by **Erssa** » Mon Jun 04, 2007 10:47 pm

SnooP wrote:But not a sleave bearing like nexus as they aren't designed to do horizontal mounting and my guess will be louder than a good ball bearing mounted this way, though would be good if anyone has/knows of tests of this.

I think this whole -sleeve bearing can't be mounted horizontally- has gotten out of hands. Maybe people don't remember, but early S12 psus had sleeve bearing fans and they were very quiet, even more quiet then new revs with Adda's.

I also think my Nexus NX-4090 is also very quiet at low loads. It was the only exhaust in my Breeze case, so all the hot air exhausted through it. The noise signature is still the same it was in mint condition, after 2,5 years of hot temps. YMMW, but I would have no problem using sleeve bearing fan in a psu.

SPCR tests have changed a bit since NX-4090 was reviewed, but it measured 21dba at low load compared to 22dba from Corsair. Early S12-430 models with Yate Loons measured 18dba at low load compared to 20dba with Adda. I think the biggest reason why they changed to ball bearing was reliability. It's worth while for the company, even if the ball bearing were only 0,5% more reliable. I'd gladly take my chances with sleeve bearing, if it meant less noise...

Khrono Devil · Post by **Khrono Devil** » Tue Jun 05, 2007 1:52 am

A 1200rpm fan won't work if you plan on using loads over 500W. The adda fan spins up to around 2000rpm at 600W loads. As heat rises inside the PSU effiency and max output drop. This is why some generic PSUs have great specs but fail because of heat.

Corsiar did a great job. The adda fan is not the quietest at low loads but using the quietest fan wouldn't be enough to cool this PSU at full load. I think this is what alot of people are forgeting, it is a 620W PSU. You can always change the fan easily if you don't like the noise this one makes at low loads/rpm but be careful choosing another fan, it may not be good enough to cool the PSU at high loads.

Steve_Y · Post by **Steve_Y** » Tue Jun 05, 2007 4:33 am

I can see why people recommend against using sleeve bearing fans in PSUs, especially when quiet fans are around using other bearings. Having a reliable PSU fan is probably more important than avoiding CPU fan failure. I doubt that a PSU would run for very long if the single 120mm fan failed while a power hungry system was under load.

Having said that, I used a horizontally mounted 92mm Nexus sleeve bearing fan for over three years in a hot running Enermax PSU. It worked perfectly until it was retired and fan noise never audibly increased. Back when I did the fan swap people weren't voicing the current concerns about sleeve bearing fans...

A176 · Post by **A176** » Thu Jun 07, 2007 5:02 am

I made another thread before I saw this one ... maybe I should've posted here.

but i'm assuming the general consensus now is that the PSU really is the last noisy component in computers.

does replacing the adda fans in most of these PSUs quiet them down?

nutball · Post by **nutball** » Fri Jun 08, 2007 1:58 am

A176 wrote:does replacing the adda fans in most of these PSUs quiet them down?

That's been my experience.

Just to complete my whinging about my free HX520W, it's exhibiting some very noticeable coil whine too when running RTHDRIBL.

Alex · Post by **Alex** » Fri Jun 08, 2007 4:01 am

I take it my fanswap idea is feasable.

I used a horizontally mounted 92mm Nexus sleeve bearing fan for over three years in a hot running Enermax PSU. It worked perfectly until it was retired and fan noise never audibly increased.

This is interesting information about horizontal fan mounting you have there Steve_Y.

Although I will not use horizontal fan (I edited my post to vertical fan now).

There shouldn't be any high heat or "horizontal fan" issues for the fan if I mount the fan vertically in the middle of the lower chamber of the P180B.

I really love my P180B and my Corsair PSU (feels like a very high quality PSU). Only negative thing that can be said about this is that the P180B is a little too big and heavy.

Khrono Devil · Post by **Khrono Devil** » Fri Jun 08, 2007 5:11 am

Alex wrote:I take it my fanswap idea is feasable.

I would have to say no, replacing a 1200rpm fan with a 2000rpm it just wouldn't cut it at the higher loads. The adda fan speeds up to 1200rpm at about a 60% load @ 22 ambient on a workbench ( aka not in a case ).

Erssa · Post by **Erssa** » Fri Jun 08, 2007 8:09 am

Khrono Devil wrote:
Alex wrote:I take it my fanswap idea is feasable.
I would have to say no, replacing a 1200rpm fan with a 2000rpm it just wouldn't cut it at the higher loads.

I think it all depends on the usage. Fanswap in general might not be worth it, because it void warranty, but if you have a low power computer, you just might use the fan with an external fan controller (fan mate?) set to 5-7v.

Alex · Post by **Alex** » Fri Jun 08, 2007 3:58 pm

Khrono Devil wrote:
Alex wrote:I take it my fanswap idea is feasable.
I would have to say no, replacing a 1200rpm fan with a 2000rpm it just wouldn't cut it at the higher loads. The adda fan speeds up to 1200rpm at about a 60% load @ 22 ambient on a workbench ( aka not in a case ).

Yes but does the fan need to be spinning that fast when there's almost only the PSU heat to take care of. The PSU must have been designed for a normal case (take care of more heat) and not a 2 chamber case or is this wrong?

But this is irrelevant according to you. I just want to now why?
Please tell me what I have missed or is it simply so that the PSU will be to hot at >60% load despite it's in a separate chamber? Will a 1600 rpm fan be enough to handle say 150W or should it be a 2000 rpm one?

As an example say that the PSU has 80% efficiency. Then 600 W/0.8 = 750W will be fed to it. => 150 will be lost in PSU.
If the fan had to take care of all the case heat (in a normal 1 chamber case not my P180B) it would have to take care of all 750W right?
So considering the 2 chamber it should be enough if the fan takes care of 150W of 750 W (=20%) of what the PSU is designed for.

But still this seems to be wrong according to you or will a 1200 rpm fan not have 20% of 2000 rpm fan airflow?
What is wrong with this? Please help me and probably others to understand? I will gladly listen and learn from your experience/thoughts about this.

I will not use the PSU for more than about 200 W draw (I should have bought a smaller one) but still I would like it to be OK with up to 400 W Load (=margin).

jaganath · Post by **jaganath** » Fri Jun 08, 2007 4:52 pm

If the fan had to take care of all the case heat (in a normal 1 chamber case not my P180B) it would have to take care of all 750W right?
So considering the 2 chamber it should be enough if the fan takes care of 150W of 750 W (=20%) of what the PSU is designed for.

It's not clear what you're saying here, but the PSU is not designed to get rid of all the case heat, the designers expect the case fans to do that.

will a 1200 rpm fan not have 20% of 2000 rpm fan airflow?

two identical fans, one at 1.2krpm the other at 2krpm, the 1.2krpm will delivery roughly 60% of the airflow of the 2krpm in free air. In a psu the air resistance means the airflow will be less in both examples.

I guess the original Adda is a 1600 rpm fan but in the lower bay there is no mobo (CPU, GPU) heat to take care of so it should be OK to swap for a 1200 rpm fan (Such as S-Flex D) instead.

The original Adda is a 2200rpm fan rated for 85cfm. The problem with swapping for a 1200rpm fan (and using the psu's fan controller) is that until >300W fan control circuit only provides <3.9V; it is very likely the 1200rpm fan will not start at this voltage.

you should just put a 1200rpm fan on a fanmate and set it where you cannot hear it but still producing a decent amount of airflow.

Alex · Post by **Alex** » Fri Jun 08, 2007 5:32 pm

Aha.

Now I'm with you 100%.
I must use external (not PSU) fan control if I want to use a "slower" fan.

My fanswap idea was clearly not feasible.

I am very satisfied with your superb answers Jaganath.

Khrono Devil · Post by **Khrono Devil** » Fri Jun 08, 2007 5:54 pm

1200rpm @ around 60% load @ 22C ambient on a test bench. Its getting more fresh air than what it can handle, putting it in a case would restrict its air flow. Also the ambient inside the case would be higher than open air.

As the heat builds up inside the PSU its total output and efficiency decrease, up to 60% load it would most likely be perfectly fine with a 1200rpm fan.

Like I said in my earlier post - "replacing a 2000rpm fan with a 1200rpm it just wouldn't cut it at the higher loads". (edit typo)

I never said anything about it not handling loads below 60%, I was just trying to warn you so that you don't end up destroying a perfectly good PSU by putting to much load on it than what it can cool.

I don't think you realize just how much 150W of heat is, a typical CPU uses under 80W and that can easily load up to 60C with stock heat sinks.

As an example say that the PSU has 80% efficiency. Then 600 W/0.8 = 750W will be fed to it. => 150 will be lost in PSU.
If the fan had to take care of all the case heat (in a normal 1 chamber case not my P180B) it would have to take care of all 750W right?
So considering the 2 chamber it should be enough if the fan takes care of 150W of 750 W (=20%) of what the PSU is designed for.

But still this seems to be wrong according to you or will a 1200 rpm fan not have 20% of 2000 rpm fan airflow?
What is wrong with this? Please help me and probably others to understand? I will gladly listen and learn from your experience/thoughts about this

The PSU doesn't deal with 750W of heat when placed in a normal ATX case. If a PC outputted 750W of heat you could use it as a heater, just because a piece of hardware uses so much power it doesn't mean that all that power is then lost as heat. Most of the power is lost/used in the working of the transistors and heat is caused by the friction and current leakage inside the transistors/CPU.

Ok so that doesn't sound very good but basically not all of the Watts that your computer uses is not converted to heat.

jaganath · Post by **jaganath** » Sat Jun 09, 2007 3:16 am

Like I said in my earlier post - "replacing a 1200rpm fan with a 2000rpm it just wouldn't cut it at the higher loads".

this is the wrong way round, you mean "replacing a 2krpm fan w/ a 1.2krpm fan wouldn't cut it at the higher loads".

Alex · Post by **Alex** » Sat Jun 09, 2007 4:50 am

jaganath wrote:
Like I said in my earlier post - "replacing a 1200rpm fan with a 2000rpm it just wouldn't cut it at the higher loads".
this is the wrong way round, you mean "replacing a 2krpm fan w/ a 1.2krpm fan wouldn't cut it at the higher loads".

Yes that is what I think he meant too.

Alex · Post by **Alex** » Sat Jun 09, 2007 5:53 am

Khrono Devil wrote:As the heat builds up inside the PSU its total output and efficiency decrease, up to 60% load it would most likely be perfectly fine with a 1200rpm fan.

OK. Yes as you say this might work. As I understand it from your comments there is probably no margin then though.

Khrono Devil wrote:I never said anything about it not handling loads below 60%, I was just trying to warn you so that you don't end up destroying a perfectly good PSU by putting to much load on it than what it can cool.

Yes you are right. Thank you for warning me. I got the message 100%.

Khrono Devil wrote:I don't think you realize just how much 150W of heat is, a typical CPU uses under 80W and that can easily load up to 60C with stock heat sinks.

Actually I reacted to this myself. Wanted to go through with my initial thoughts though. I think 2200 rpm fan might handle 150 W (with very hot components) but certainly not 750 W for sure.

Khrono Devil wrote:The PSU doesn't deal with 750W of heat when placed in a normal ATX case. If a PC outputted 750W of heat you could use it as a heater, just because a piece of hardware uses so much power it doesn't mean that all that power is then lost as heat. Most of the power is lost/used in the working of the transistors and heat is caused by the friction and current leakage inside the transistors/CPU.

Ok so that doesn't sound very good but basically not all of the Watts that your computer uses is not converted to heat.

Well I agree that for example a very small amount of energy is converted to light in light diodes and some is transfered to air flow by the fans but most of the energy is normally finally converted to heat like through resistances in circuits or motor friction.

In digital circuits (transistors) the energy is temporarily stored in "capacitors" (also MOS transistors and electrical lines have "parasitic capacitances"). When these "capacitors" are uncharged all their energy is normally converted to heat. This happens when short circuiting (switching) to ground.

As transistors and capacitors do have resistance when switching and ground leads are not superconductive the energy is converted to heat through these resistances. Of course there are some resistive heat losses during capacitor (up) charge switching also.

It is possible to store a part of the capacitively stored energy for reuse in other energy storage capacitors or batteries but the discharging time gets severely decreased then, so this is normally only done in slower low power circuits that rarely is used in desktop computers yet.

Also the low power circuits have to have a more complex design (=more expensive and longer time to market). This will come in the future with the help of more advanced design tools and cheaper chip area manufacturing (denser chips / More's Law). I don't think we are there yet for our desktop CPUs and surrounding circuits but this is possibly true for laptops already?

Moogles · Post by **Moogles** » Tue Jun 12, 2007 7:43 am

This might be interesting: http://silentpcreview.com/forums/viewto ... ht=#334430

My Corsair HX520 was virtually inaudible when I first got it (late January), but has recently gotten quite a bit louder. It also seems to have developed a MINOR case of coil whine. Compared to the first HX520 I got (defective unit with ridiculous coil whine) this is nothing. Still enough to cause annoyance, unfortunately. I'm looking to do either a fanswap, or get a passive PSU.