You are here

Archive: SPCR's Test / Sound Lab: A Short Tour

October 3, 2004 by Mike Chin

The first reviews for SPCR emerged from a rat's nest of equipment, computers, components, wires and other assorted geeky things that my home office had become during many months of experimentation prior to the launch of the site. Two and a half years have gone by since those early days, and now the lab has taken over the downstairs kitchen and half of the adjoining TV/den. A lot of equipment, instrumentation and plain old stuff has been acquired, some by happenstance, others as a result of careful planning.

The most recent additions to the lab are related to the sound equipment carefully acquired for the purpose of making recordings of computers and components. It's probably of interest to regular visitors of SPCR.

View of lab from rear corner: There are usually several complete systems on the go, a few open bench ones on the left running Folding @ Home for SPCR or heatsink tests; the desk on the right is usually reserved for the audio equipment, complete PCs under review or general assembly. Two 19" CRT monitors display output from any PC through multi-KVM switches. All the PCs are networked. There are two independent access lines to the Internet; one through cable and one through ADSL. This ensures constant connectivity.

View from the entrance, showing more... stuff. That's a very good camera stand with uni-pivot head in the center, one of the most valuable pieces of gear I own. The lab has been changed since the above photo was taken, as you will see below.

As the lab is set up today: Note the PSU test rig in the background, and the deliberately displayed sound recording gear in the center of the room.

Close up of the equipment needed to measure fans and record their noise: A Taylor Hohendahl Engineering (THE) KP- 6M Reference Microphone is in a suspension shock mount on a boom mic stand. It is a wonderfully sensitive and linear transducer, a generous donation to the SPCR cause by Taylor Johnson (as he is known); much thanks to Mr. Johnson! On the tall stool is a Nexus 120mm fan and a handheld digital display anemometer (to measure airflow) on soft foam pad atop wooden stool. On the lower stool are a B&K 2203 SLM (capable of reading below 10 dBA) with 1603 filter accessory attached and an ordinary digital multimeter. Not shown here are the power supply and voltage control devices to power the fan.

Audio recording system: Modified Shuttle Zen PC running a P4-2.53 and suspended Samsung 40G 2.5" notebook hard drive, with single channel M-Audio Tampa mic preamp and M-Audio Firewire 410 external digital sound interface feeding the signal from the microphone.

Mic Preamp & Sound Interface: The top item is a M-Audio FireWire 410, a FireWire-compatible audio/MIDI interface with 4-in/10-out configuration complete with preamps. However, for best results, I use the M-Audio Tampa professional microphone/instrument preamp which has lower noise and higher performance than the preamp built into the FireWire 410. The Tampa integrates a 96-kHz / 24-bit A/D converter, which means the signal remains in digital format from the Tampa onward. Both instruments were purchased directly from M-Audio.

Modded Shuttle Zen

The audio PC was modified for minimal noise so that it could be in the same room as the microphone if necessary. It measures about 16 dBA/1m.

The first mod is at the bottom of the case, where a spare piece of AcousticPack material was cut and affixed. The white rubber feet no longer make contact with the desk. This mod ensures that any vibration emanating from the internal suspended Samsung notebook drive does not conduct to the desk and cause any extraneous noise.

The second mod involved the removal of the back fan grill, which reduces impedance to fan airflow to an absolute minimum. The fins of the heatpipe cooler are visible in the photo above. Rubber edging to hide the sharp edges was provided by MNPCTech. The internal fan with its frame was also removed. A Zalman Fanmate1 was plugged into the motherboard header for the CPU cooling fan and run outside the case.
The third mod involved installing a Panaflo 80L on the outside of the case with rubber damping mounts from Verax. These easy-to-use devices work well with fans that have a cylindrical channel for mounting holes. They were not needed for the Panaflo (EAR fan mounts would have worked fine) but just happened to be handy. The external mounting without the shroud maximizes airflow, and keeps the fan from exciting any internal cavity resonances that can cause the aluminum case panels to resonate. The difference in noise between external and inernal fan mounting was noticeable even at a such low fan speed.
Here's the Panaflo decouple-mounted on the back panel, with the Fanmate1 turned to just about minimum: That's enough cooling for this PC at the hottest summer temps... admittedly, the temp in the lab rarely exceeds 26~27° even on the hottest summer days.

Numerous audio software was loaded in the Shuttle Zen and tried over the last few months. These include:

Sound Forge 7.0 - Digital audio editor that includes a powerful set of audio processes, tools, and effects for recording and manipulating audio. Features include DirectX® plug-in effects automation, automated time-based recording and audio threshold record triggering, VU/PPM meters for RMS playback and record monitoring, enhanced Spectrum Analysis tools, white, pink, and brown noise generators, clipped peak detection and marking, Vinyl Restoration™ plug-in, Media Explorer, Sound Forge project file creation, support for 24fps DV video files

Steinberg Wave Lab 5.0 - High resolution stereo and multi-channel audio editing, mastering, CD/DVD burning and for complete CD or DVD Audio production in outstanding audio quality. Not just a top level mastering tool, Wave Lab also offers capabilities for multimedia, sound design with samplers, radio broadcasts, and computer telephony applications.

True RTA - Audio analyzer for testing and evaluating audio systems using just a PC with basic sound capability. The instruments found in TrueRTA include a low distortion signal generator, a digital level meter, a crest factor meter, a dual trace oscilloscope and a high-resolution real time audio spectrum analyzer.

LameFE - Offers a simple GUI for LAME and other encoders, including Ogg/Vorbis, MP3, WAV, and Monkey Audio. LameFE features a fully automated batch mode (it can use all CD drives) and also supports CD backup, CDText, and CDPlayer.ini. A toggled ID3 tag editor allows quick edits to ID3 version 1 and 2 tags. Users can easily define filename format, create playlists, query FreeDB servers to access file and album information, and much more.

M-Audio WireWire 410 interface software, which comes with the hardware.

All of this software (and many others) were extremely useful in the learning process to make high fidelity recordings of PC noise sources. Currently, Steinberg WaveLab is the main tool of choice, for many reasons. It is used to capture the sound as a high resolution WAV file, and also to create 44.1kHz / 16-bit / 96kbps MP3s, the files actually posted for download to SPCR. They have high enough resolution for fine details of quiet fans at very low speed to be clearly discerned on a decent PC audio system or good headphones, yet are just 170kb file size for 15 second segments.

In case you have not already heard these files, this is the gist of the relevant portion of first SPCR review incorporating new sound files.

LISTEN TO THE FANS!

You can actually hear the noise of the PHT-3600 for yourself and compare it to the Panaflo 80L (mounted on a heatsink) with the MP3 files below. These recordings were made with a high resolution studio quality digital recording system assembled specifically for recordings of quiet computer gear. The recording system and the methodology used will be covered in detail in a forthcoming article.

For now, suffice it to say that the recording microphone is always 3" from the edge of the fan frame at a 45° angle, facing the intake side of the fan and just out of the high turbulence area to avoid direct wind noise. The background noise during all recordings is 17 dBA or lower. The input level settings of all the recordings are absolutely identical, so the differences between noise sources is retained with very high fidelity. The files themselves are mono, 44kHz/16-bit with 96kbps compression. Most recent vintage music playback software should have no trouble with these files.

The better your audio playback system, the more accurate the sound reproduction and comparisons of different noise sources will be. However, even with a modest PC sound system, it is possible to get a reasonably good idea of the differences between noises. The key is to calibrate your playback system to the correct volume level.

Audiophiles and those involved in music recording are aware that truly accurate reproduction requires playback at the same volume level as the original sound. There is a way to calibrate your playback system to the correct volume level for SPCR's noise recordings. To do this, you must have a Panaflo 80L, our reference fan, and a way to run it outside the PC at 12V.

Calibrate your Sound Playback System

  1. Hook up the Panaflo 80L fan to a 12VDC source and place it right next to a speaker. Preferably, the fan should be on a soft padded surface to minimize vibration couple effects.
  2. Turn the fan on and listen to it for a minute.
  3. Now download, save and play this reference file of a Panaflo 80L (Panaflo 80L 12V Reference MP) in free air, driven by 12VDC. Like all the other sound files, it is 15 seconds long. Set your playback software to repeat loop if you want to listen to it for longer than 15 seconds without hitting play over and over again.
  4. Adjust the playback level of your sound system until it matches (as closely as possible) the sound level of the actual Panaflo fan next to the speaker. (You must also turn off all special sound effects, and set equalizer / tone controls to neutral or flat.) You may find this easier to do while listening to each sound by itself; simply switch the fan off while listening to the sound file and stop or mute the sound recording when listening to the fan. Once the levels are matched, do not touch the playback level settings. You may wish to record these settings for future reference.
  5. Now, your sound playback system is calibrated for level. The Panaflo 80L sound file will be played back at the level it was when recorded, as will all other SPCR MP3 sound files. If you have a very high fidelity sound playback system, these MP3 files will be the next best thing to having the actual various noise-making PC components we review in your own room to listen and compare for yourself.

NOTE: The Panaflo 80L 12V Reference MP3 is not the same as the Panaflo 80L on HS @ 12V file. The reference Panaflo 80L file was recorded in free air, driven by 12VDC. The Panaflo 80L on HS files were recorded at the indicated drive voltage while mounted on an Alpha 8045 heatsink. This is to ensure a fair comparison between the Nexus HSF and a HS equipped with a Panaflo 80L. Any fan in free air is quieter than when it is mounted on a heatsink. The close proximity of the heatsink -- or any obstruction (impedance) -- to a fan's spinning blades causes air turbulence which adds to the overall noise. By mounting the Panaflo on the Alpha heatsink, both fans are operating against an impedance; hence it is a fair comparison.

If you don't have a Panaflo 80L fan, first turn off all special sound effects, and set equalizer / tone controls to neutral or flat. Listen to the quietest of the above sound files (the Panaflo 80L at 7V), then adjust the playback volume to the point where it is just barely audible a meter away, and don't touch the volume setting when you listen to any of the other files. That will be reasonably close to the actual recorded sound levels.

Finally, if your computer makes enough noise to drown out the Panaflo at 12V, you probably have little choice but to listen over the headphones, setting the level to where the quietest fan noise file can be heard comfortably, and then not resetting the volume. On the other hand, you probably should stop listening to these MP3s and just get cracking on making that PC quieter before your hearing suffers!

Finally, here is a view through a doorway of the adjoining TV/den that the SPCR lab is slowing annexing. It is far too much of a mess to show you more at this time. Suffice it to say that this room will be featured in many recordings. It is where most of the fan and hard drive noises will be recorded, as it has the benefit of a lower ambient noise floor (14 dBA after midnight), reduced reflections due to carpeting and heavy drapes, and larger dimensions (20' x 10' x 8') for reduced wall proximity effects in both recordings and SPL measurements.



Just a peek at the recording lab through the adjoining door. The visible table is temporary.

Much thanks to all the contributors and sponsors who gave valuable time, expertise, equipment, money and other forms of support for SPCR's lab resources. With such support and continued dedication, SPCR will keep growing our quiet niche of the PC world.

* * *

Comment on this article in our Forums.

Google

www SPCR