What IS "Silent Computing"?

[clocker]

Well?
Seems like we have a golden opportunity here to define the PC version of a "silent" computer.
I would posit that, to date, SPCR has been the gold standard in the field but, by the very nature of their narrow focus, they may have developed standards too extreme* for a performance cooling oriented forum.

For instance, look in the Watercooling section and it is obvious that the Zalman Reserator is the favored waterloop.
Naturally, it would be an obvious choice for consideration, but probably not the subject of intense passion here.

So what qualifies as "silent" when noise output intersects with cooling performance?



So far there is a gaggle of SS Temjins...nice as they may be, are they ProCooling "silent" (I'm pretty sure they would fall just shy of "leaf blower" over at SPCR...)?


*Not measurement standards, rather the very concept of what a quiet PC is and what will be sacrificed in it's pursuit.

[Etacovda]

An interesting topic.

What defines 'silent'?

AskOxfords english dictionary -

• adjective 1 not making or accompanied by any sound.

But, silence is defined as such -

• noun 1 complete absence of sound.

Well, for something to be COMPLETELY absent of sound, it must be completely inanimate; so i say that no computers are 'silent'; if its got an hdd, its not a silent computer. Motherboard coils probably make high frequency squeals beyond what we can hear, i know some make them in the audiable range (my A7N8X-X's coils are very annoying...)

So, to me, no computers are 'silent', so we have go beyond the 'silent pc' idea, to 'quiet pc'. For a pc to be considered truely 'quiet' id say it would have to be below 21db or so.

For procoolers, I suggest that a silent pc is one that is quiet but perhaps audible (sub 25db), and is either at stock, or is overclocked. Anything underclocked here, for silence, just seems a bit too SPCR'ish for me

'Silent' in the pc world is 'inaudible', i guess; but thats a hugely subjective thing; whats quiet to another is deafening to me, and ive found that numerous times...

I guess my conclusion here should be, basically, what my ideal pc is.

A pro/quiet pc should be barely audible with no sacrifice to the computers usability or performance; ie, underclocking is OUT, instability is OUT. My ideal (realistic) pc at the moment would be;

Opteron 939 dualcore
2gb of ram
7800GT
Asus Premium motherboard (heatpiped mosfets + sb)
74gb raptor (the newer, quiet edition)
200gb seagate storage drive
DVD writer
Seasonic psu (or antec phantom, i believe they're fanless until above 30C)

All watercooled with a dynamically alterable mcp350, with 1 120mm fan in the psu, and one on a pa160 or similar (also dynamically altered), probably in a P180 Antec case, with suspended, enclosed drives.

If done right, i doubt you could hear it - and it would still overclock extremely well, and with dynamic water temp control, it would be EXTREMELY quiet when not being used for anything intensive. Best of all, its probably only around 2kUS to build...

[Marci]

Have been musing over Nexus' PSUs... ie: suspend PA160 by it's shroud from the base 120mm fan of the Nexus PSU - to cut that proposed 2x120 down to 1x120...

At the mo, CSP-MAG, PA160 & Nexus120mm, BarracudaIV HDDs (decoupled), and NeroDriveSpeed handling optical devices has resulted in a completely inaudible rig for me... (Tagan 480w PSU)

S'a bit lower spec'd than your list -

XP3000+
1Gb Ram
9800Pro All-in-Wonder
Abit NF7-S

CPU clocks to 2.4Ghz at 200FSB, 5-2-2-2 on the ram...


Would recommend mCubed T-Balancer to handle dynamic Pump and Fan control based on temp...

[pdf27]

Originally Posted by clocker

So far there is a gaggle of SS Temjins...nice as they may be, are they ProCooling "silent" (I'm pretty sure they would fall just shy of "leaf blower" over at SPCR...)?

Well, my TJ-06 is I think silent to SPCR standards. Total radiated noise is on a par with a Nexus 120mm fan at approximately 7v I would guess (unfortunately haven't been able to test that - I'm using AC fans to totally eliminate fan clicking noise which really irritates me). I would happily put it up against any reserator PC in the noise stakes, it is that quiet IMHO.

[clocker]

Fair enough, I was exaggerating for effect and didn't mean to impugn the serious approach to silence you took.

You do bring up another point I had considered...the subjective nature of noise and how to express it in an easily understood way.

Originally Posted by pdf27

Total radiated noise is on a par with a Nexus 120mm fan at approximately 7v

Assuming we all had said Nexus fan in hand this would certainly be one viable method of communicating an evaluation.
You also judged the noise output with the machine totally assembled (I am assuming the sidepanels were in place, etc) and in it's real environment- again, a valid method.

Until/unless someone wants to step and build an anechoic chamber big enough to hold say, um, a Stacker and invest in the sound gathering/analyzing hardware we're going to have to design a set of expressions that can be commonly understood and whose methodology is easily duplicated (i.e. "a Nexus <insert model # here>120mm fan at 12v held 12" from the ear (earwax recently removed and even layer of AS5 -not fully cured yet- applied as per website instructions).

If someone does build the sound testbench we can then get into giant debate about the testing validity/procedures which, with any luck will generate lots of pixelated O positive.

Anyway, for myself, I suspect the most interesting aspect of this subforum will be the pics anyway.
Seeing how folks approach a problem is more interesting to me than their analysis of the success. After all, a clever suspension that fails with a HDD might be just the ticket for a PSU or whatever.

[bigben2k]

Sound is very subjective.

"off the cuff" definition would be: if you can't tell wether or not it's on, then it's silent.

Technically, it would have to fall under a specific dBa level (on all frequency ranges). Probably in the low to mid 20's. Then there are music purists that would go lower, or require low dBa on various ranges.

[Joe]

I think "silent" in silent computing is more along the lines of an objective or a goal for the cooling more than the factual result.

I think everyone would want a machine that could OC to the moon, run cool, and silent, but few can get there.

[clocker]

Originally Posted by Joe

I think everyone would want a machine that could OC to the moon, run cool, and silent,...
.

Whoa, Nelly...them's some hifalutin goals you got there.
Personally I am primarily interested in reining in my quartet of 2GB Quantum BigFeet.

Wicked fast, dude, but a trifle noisy.

[Joe]

Well I think it was Bill then pH that said you can have any 2 of these:

High Performance Cooling.
Silent opteration.
Reasonable Costs.

With enough money you can get something that supercools with virtually no noise, but it wont be cheap by any measure

[Joe]

oh and I have a silent PC, no moving parts. A Webplayer Solid state HD, with just a HSF for the CPU. makes a nice thin client hehe.

[Rusty075]

The idea that computers must be noisey to be fast is outmoded. When PC's started getting faster they got noisier, so some people began to associate the noise with the speed. But in reality the noise increase has more to do with cheapness than the increased power output. Fans are cheap, and fast fans cost the same as slow fans, but quality heatsinks and design are expensive (relatively).

Now that there is an established backlash against pc noise the market is full of options. You don't even have to be a modder anymore. With just a bit of research and knowledge you can fill a Newegg shopping cart with all the parts needed to make a fast, powerful, but quiet machine. And it won't cost significantly more than a loud but equally fast machine would.

As for the semantics of "silent": all that really matters is the relationship between the PC's noise and the ambient noise of its relationship. A 50dBa PC is "silent" if it lives in a machine shop. The same pc would be a screamer in your bedroom. There a 20dBa machine would be "silent". Pursuing anything less than ambient is an academic challenge only. Not that that makes it a bad thing, but it just puts it in the category with OC'ing-as-a-hobby. Most of our articles we post at SPCR fall into the first category, while our forums tend to be of the "academic" variety.

[Joe]

Rusty I would have to disagree on a silent machine not costing that much more than a normal machine to get the same performance. I mean when you can buy $4.00 high speed fans that just SCREAM at like 50db or something, it makes buying $15.00 silent fans a bit of a money hit.

Also to get equal cooling for the slower fan speed, you need to drop money on more serious cooling solutions such as a more effective blocks, raditors, pumps, etc...

I think to get a reasonably quite machine with a limited performace cap, is able to be done for sorta cheap. (the middle of all worlds).

[pdf27]

Originally Posted by Joe

Well I think it was Bill then pH that said you can have any 2 of these:

High Performance Cooling.
Silent opteration.
Reasonable Costs.

With enough money you can get something that supercools with virtually no noise, but it wont be cheap by any measure

I'd add size/weight to that list - the bigger and heavier the PC, the easier it is to silence and the better whatever noise is generated will be suppressed.

[zoson]

86dba fans and very loud music = silent computing!!!

[pdf27]

Originally Posted by zoson

86dba fans and very loud music = silent computing!!!

Mainly because you'll soon be deaf

[Maestro]

I have to agree with what pdf said above: Size and weight are definitely also a BIG factor when you're looking for silence. I think that may be part of the reason that the Temjin-06's are prominent here so far--this is a big case with panels and supports that have some real heft to 'em. I know if I moved my current setup into a smaller, aluminum case, I would be forced to move to smaller and faster/louder fans to maintain the same cooling, plus the panels would transfer more noise, etc, etc...

Also though, I do not think that performance necessarily needs to be sacrificed. I think my PC posted here would pass muster over at SPCR, and as pdf says, his certainly would too, and neither of these is really sacrificing anything.

Really, I do not think we can define silent computing, and I think it is in itself a relative term. I would think that this section of the forum would be most useful as a place to discuss HOW to make a PC quieter, not necessarily as a showcasing of what components/modifications/whatevers will result in some pre-determined decibel level that defines "silent". I think that taking a set of hardware and reducing the decibel level from 36 to 24 without any loss in performance would qualify as a practice in silent computing, whether 24db is considered silent or not.

However, if you're looking for a number, I submit 20.31728db @ 2'3.8" as the gold standard.

[clocker]

Originally Posted by Maestro

Really, I do not think we can define silent computing, and I think it is in itself a relative term. I would think that this section of the forum would be most useful as a place to discuss HOW to make a PC quieter, not necessarily as a showcasing of what components/modifications/whatevers will result in some pre-determined decibel level that defines "silent". I think that taking a set of hardware and reducing the decibel level from 36 to 24 without any loss in performance would qualify as a practice in silent computing, whether 24db is considered silent or not.
I like this approach a lot.

However, if you're looking for a number, I submit 20.31728db @ 2'3.8" as the gold standard.

OK, I give up.
Where did your "gold standard" numbers come from?

[Etacovda]

im imagining from that wonderous thing called sarcasim

[clocker]

Originally Posted by Etacovda

im imagining from that wonderous thing called sarcasim

Ah.
Thought I detected the barest hint of eau du fecale but wasn't certain.

Have I just been "owned" as you kids say today?

[gmat]

Originally Posted by Joe

Well I think it was Bill then pH that said you can have any 2 of these:

High Performance Cooling.
Silent opteration.
Reasonable Costs.

With enough money you can get something that supercools with virtually no noise, but it wont be cheap by any measure

I second that.
Those 3 points form a triangle. Move towards the performance/silence side and you slide away from the 'reasonable cost' point...
For me reaching the 'silent computing' goal means you accidentaly turn off the computer, thinking it was off though it was actually running the screensaver. Or to put it another way, you cannot tell if it's running or not.
To reach this goal and keep a high performance system is very expensive. Overclocking makes it even more expensive. If i can i'll detail my current system (YY cube) and my next system (Lian Li V2000) to show the approach i took to this problem.

[tomservo]

I disagree, at least in one aspect - it's more of a square, or a tetrahedron or something... with technology level as one point. If you have old tech to cool, it can be cheap, silent and reasonably priced ;p

[tomservo]

I disagree, partly - there's a fourth point (think tetrahedron) and that's technology level (performance?). If you have an older processor, you can get high performance cooling, reasonable costs, and silent operation.. also depends on if you're buying everything, or scrounging/diy. We must make this more complicated!!! lol...

[msv]

Feels like I came in a bit late in this discussion, but to me silent is what I cannot hear.
I agree with Rusty in his comparsions with different environments. If I can´t hear my PC because of other sounds (that I consider normal in that environment) I regard my PC as "silent". If that means 20 or 30 dB is another question.

regards
Mikael S.

[bigben2k]

That's the problem; when you say "...to me...", it's not something that anyone else can relate to (i.e. subjective). Same for the environment.

It's hard to discuss sound levels. dB ratings do not account for the pitch of the noise either; some cannot tolerate a high pitch at a low level, others simply can't hear it, but the dB rating is exactly the same.

To make things worse, it's not even possible to attach a sound clip to this thread, and have everyone listen to it; you can't guarantee the actual sound level.

I've even given some thought on how people could make a "standard" sound using standard household items, and came up empty.

I'm down to a simple speaker and a frequency generator (at a specific wave / amplitude / frequencies), but few have one. Speakers respond differently...

[clocker]

This brings up the question of whether we compare sounds with the IHS removed from our ears or left in place.
I'm guessing that BillA has an opinion on this....