Making my old faithful PC shut up: from 48db(A) to 33db(A) !

Show off your quiet rig.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
drahthaar
Posts: 1
Joined: Thu Sep 11, 2014 10:47 am

Making my old faithful PC shut up: from 48db(A) to 33db(A) !

Post by drahthaar » Thu Sep 11, 2014 11:07 am

**Long story short:** Where I live it’s so quiet that at night you can literally hear sheep fart in the distance. My living room PC was too bloody loud for comfort even at my sitting distance of 2,5m. So with a 100 euro budget I set on a quest to make it as silent as possible while learning and remembering as much as I could about sound wave propagation and pc acoustics (I’m old guys, it’s been a long time since University). The PC is used for internet browsing, watching movies, streaming and gaming for older titles such as COD World at War (it still manages 30+ FPS @ 1920x1080 and ultra settings).

Anyway, for those interested, I went from this:
Image
to this:
Image

Build Log (kinda) is here:(http://imgur.com/a/3eQgM).

Sound graphs are here: (http://i.imgur.com/6tQIL5l.jpg) and here: (http://i.imgur.com/hQemyzX.jpg).

The results were downright impressive: The original rig was pumping out 48 db(A) spl @2.5m under load and 46 db(A) while idle. After a small tsunami of relatively cheap DIY modifications and a careful selection of a few components, the sound pressure level dropped to 33 db (A) under load and 32 db(A) idle. This baby is now hardly audible during the night and below ambient noise during the day so that you have to actually look at it to see if it’s on. Measurements do not claim to be scientific as I used my Galaxy S3 to record sound pressure levels, but they do give a comparative numerical view of the great noise difference my ears are recording!

Also keep in mind that DBs are measured on a logarithmic scale. This means that the actual spl reduction from 48 db to 33 db is not 15db. You don’t just subtract dbs like that. The actual reduction is 47.86 db. The actual perceived noise emitted after the modifications is less than 6.25% of the original setup (using psychoacoustics - less if calculated in other ways - any help out there?)

No performance related modifications where done. Old faithful hosts an [AMD Athlon 64 X2 @ 2.8GHz](http://www.cpu-world.com/CPUs/K8/AMD-At ... BOX\).html) with 4GB dual channel 800MHz DDR2 RAM, three HDDs (two of them JBODed, mix and match SATA and IDE), and the GPU is a two slot [GIGABYTE Nvidia 9600GT 1GB](http://www.gigabyte.com/products/produc ... id=2804#ov), passively cooled. Except for the HDDs everything else is the same.


**Modifications**:

1. **CPU Cooler**: Stock cooler was chucked and replaced by a [ZALMAN CNPS90F](http://www.zalman.com/global/product/Pr ... hp?Idx=785). This little baby contacts my CPU using my favourite Arctic Silver 5 paste and it is dead quiet even though I have seen reviews complaining for its noise. *Price: €10,90*

2. **PSU**: The OEM 425W PSU had a 92mm fan which was noisy and after years of continuous use had developed the habit of rattling besides getting hot. Also chucked and replaced by an active PFC 450W [Sharkoon SHA450-12EUP](http://www.sharkoon.com/?q=en/content/sha450-12eup). This one has a thermally controlled 120mm fan and it is also impressively quiet even under load. *Price: €44*

3. **HDDs**: The three older 3,5’’ drives were retired and replaced by two 2,5’’ SATA drives salvaged from older laptops. One is a Hitachi 320GB 7200 rpm for the OS and the games and the second one is a Seagate 500GB 5400 rpm for photos and stuff. The drives are elastically suspended from the chassis, placed right in front of the front intake fan. I also attached onto them (using Arctic Silver 5) a pair of chipset heatsinks salvaged from older mobos. You can never tell if they are busy - they are completely inaudible. *Price: free.*
3. **Case fans**: I bought two different sets mainly out of curiosity. A pair of [ANTEC TrueQuiet 120mm](http://uk.pcpartpicker.com/part/antec-c ... uequiet120) that feature a high-low rpm manual switch. One of them is set as front bottom intake and the other one as rear top exhaust. Both are set to the low rpm settings. Just as most reviews say, at low speed, airflow is not impressive but you have to look at them to actually notice that they are turning! The second set was a pair of [Arctic F12 TC 120mm](http://www.arctic.ac/eu_en/arctic-f12-tc.html). These have a thermistor switch, which controls the rpm. They turn at low rpm up to 38oC and then switch to high rpm at that threshold. Not as quiet as the ANTECs but they push more air. One of them was set at the bottom of the case as intake, blowing directly on the the CPU heatsink and the thermistor wire buried in the GPUs heatsink fins. The other one found its destiny jury-rigged in the 5,25’’ bay, blowing air towards the CPU and and PSU. Its thermistor switch was attached to the CPU cooler fins. All fans are controlled by a single MOBO fan header though the BIOS and they are all three pin.* Price for all 4 fans + splitter: €33*

4. **Case**: This is where most of the effort was funnelled. An old cheap OEM midi tower laying around with dead mobo was a perfect candidate. [This](http://i.imgur.com/NmftzdM.jpg) was before I started chopping, and then it begun:

* All fan bezels were chopped off and the openings were enlarged to host the 12cm fans. No obstacles for airflow and no whooshing sound of air through grills.

* All internal gimmicks such as the 3,5’’ bay were chopped off. The 5,25’’ stayed because it was essential for the structural integrity of the case.

* Bottom hole was cut out to host one of the Arctic F12TC fans.

* All cuts made by me in the case were lined with hard plastic edging to protect fingers and cables
* Old case stands were hard plastic, 0.5cm high. Replaced with 4 furniture rubber feet, 2,5 cm long. Thus no vibrations are transferred to and from the tile floor and also the case is raised so that the bottom intake fan sucks in less dust.

* The side panels and also top and bottom of the case were lined with heavy, high density thermoplastic material 2mm thick. This is normally used in roof construction for water insulation but it also raises the mass of the case and it reduces vibrations and case resonance while blocking low frequency sounds from escaping the case.

* On top of the thermoplastic, the side panels and case bottom were lined with 4cm thick acoustic absorbent material used in studios and clubs. It does a wonderful job but its egg carton design probably screws airflow - not that I really care though. More on temperatures later. The top panel of the case as well as the 5,25’’ bay were also lined with acoustic material but this was 1 cm thick and it is the type that they use to soundproof air ducts in buildings. Both soundproofing materials price: 30 euros but I have left over to build another 5 PCs :-)

* This case provided ZERO cable management capabilities. So I made an 1:1 model of my mobo, stuck it in the case and started drilling holes in the motherboard tray so that all cables (power, SATA etc) popped up exactly where needed. No cable clutter in the front, free case airflow and an incredibly big cable mess behind the mobo tray. Also a lot of small holes were drilled and an elastic rope was passed through them so that it acts as cable holder on the back of the case.

* DVD and floppy drives were removed since I never use them, and all front covers were drilled with holes to allow air intake.


Although I was prepared to take a 5%-7% premium in temperature increase due to messed airflow by the soundproofing material, I was pleasantly surprised to see my actual temperatures drop all over the system: (Ambient is 21C)

* Under full load (60 min of prime95) CPU went down from 60C to 52C and it idles at 39C (from 46C).
* My GPU went from 82C to 60C under full load with the furmark burn in test. It idles at 38C and when I game it rises up to 50C-60C.
* The three old HDDs averaged at 43C but the two new ones now average at 29C.

*Lessons learned:*

1. Going quiet or silent can get very expensive very quickly. You don’t have to go to fanless cooling and total silence to enjoy quiet computing. Careful planning and component selection can lead to impressive results without spending a fortune.

2. Contrary to common intuition, more fans of the right type/size and at the right spots in the case can lead to quieter computing than fewer fans. Credit for this observation must go to a fellow redditor whose original post I can’t find. Sorry stranger!

3. Early planning is always beneficial for sound contention when building a new PC, regardless of budget or scope of the build. I realise that most of the discussions in the PC building world revolve around power and performance but I believe that the overall computer experience whether for work or play is hugely enhanced when you can’t hear the bloody thing work!

4. There is a wealth of knowledge out there for anyone looking to build a quiet or silent PC. This project took 2 months in the planning (and studying) and less than 3 days in building it (had to let the glue cure). Don’t ignore older content from almost 10 years ago. Technology changes but the laws of physics don’t, they still apply.

5. 33db is still audible by htpc standards, and it is definitely not officially a “quiet” pc which requires sound levels not to exceed 30 db(A). But not many builders would try to build a pure quiet htpc using my pre-existing HW such as a 90W TDP CPU anyway (especially on this budget).

6. I am never EVER going to buy a non modular PSU again. If you see the pictures behind the mobo tray then you'll realise why ...

7. I overshot my budget by 9%. Up from 100 euros to 109, but hey, I’m Greek, I’m supposed to overspend right?

8. This was originally posted by me on reddit...

CA_Steve
Moderator
Posts: 7651
Joined: Thu Oct 06, 2005 4:36 am
Location: St. Louis, MO

Re: Making my old faithful PC shut up: from 48db(A) to 33db(

Post by CA_Steve » Thu Sep 11, 2014 4:13 pm

Welcome to SPCR.

Looks like a fun project. Next time, you might take an incremental approach and see what gains you get from each step. Adding mass is useful only if you have vibration related noise, etc..
Also keep in mind that DBs are measured on a logarithmic scale. This means that the actual spl reduction from 48 db to 33 db is not 15db. You don’t just subtract dbs like that. The actual reduction is 47.86 db. The actual perceived noise emitted after the modifications is less than 6.25% of the original setup (using psychoacoustics - less if calculated in other ways - any help out there?)
I think you may be conflating how we perceive sound and sound measurements. When it drops from 48 dBA to 33 dBA, it is a drop of 15 dB. You do just subtract dBs like that.

Phone noise meter apps are just a comparison tool. It's fairly safe to say you dropped the noise level 15dB, but the actual noise levels probably have an accuracy in the +/- 10dB range. The self noise of your SGS3's mic is probably > 30dBA. In any case, looks like you had fun.

Abula
Posts: 3662
Joined: Tue Nov 02, 2004 12:22 pm
Location: Guatemala

Re: Making my old faithful PC shut up: from 48db(A) to 33db(

Post by Abula » Thu Sep 11, 2014 5:12 pm

Welcome to SPCR.

I like what you did with $109, pretty nice work.
drahthaar wrote:1. Going quiet or silent can get very expensive very quickly. You don’t have to go to fanless cooling and total silence to enjoy quiet computing. Careful planning and component selection can lead to impressive results without spending a fortune.
That imo is the most important lesson in my experience buidling toward quiet pcs.

bobb
Posts: 20
Joined: Fri Aug 15, 2014 6:13 pm

Re: Making my old faithful PC shut up: from 48db(A) to 33db(

Post by bobb » Sat Sep 13, 2014 7:32 am

I counted 6 fans and only 33 db? Impressive!

Now you know you have to keep them all oiled up to maintain the silence right... KIDDING!

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Making my old faithful PC shut up: from 48db(A) to 33db(

Post by MikeC » Sat Sep 13, 2014 9:27 am

CA_Steve wrote:I think you may be conflating how we perceive sound and sound measurements. When it drops from 48 dBA to 33 dBA, it is a drop of 15 dB. You do just subtract dBs like that.
Yup. Subjectively, if the 15 dB measured drop is accurate, this means it should sound ~3x quieter. 10 dB SPL change = doubling or halving of perceived sound.

Telstar
Posts: 140
Joined: Fri Oct 05, 2007 2:20 pm
Location: Italy

Re: Making my old faithful PC shut up: from 48db(A) to 33db(

Post by Telstar » Sat Oct 04, 2014 2:35 am

Kudos to you for achieving quietness with old components and a tight budget :)
You are lucky of having a quiet environment...

Post Reply