Optimal MoBo for Audio CAPS or Mac 2014

[tranz]

Logically the absolute best is a custom motherboard like Aurender makes for their units, however, are there any absolute top options for those building their own CAPS or Mac flavoured CAPS today?

 

Designed for low EMI. Ability to individually power every element with linear PSU, to be able to bypass the onboard smps of most boards.

 

Reading sites like hardsecrets, jplay.eu, etc. there are many electrical noise sources to be eliminated for optimal fatigueless playback.

 

Cheers

 

 

Everything You Need to Know About The Motherboard Voltage Regulator Circuit | Hardware Secrets

[tranz]

A few follow up findings:

 

- Intel S1200KPR, although EOL, seems to be a fav for Win. (Replacement is BBS1200KPR)

 

- Kyrill on the Jplay forum looked into a custom audio mobo a while ago without much luck, and it does explain why something like the Aurender would cost as much as it does.

 

Dedicated Audiophile Motherboard development | Computer Audio | Forum | JPLAY - hi-end audio player for Windows

 

A few thoughts from other threads and forums on the optimal mobo setup:

 

- keep it simple. No unnecessary chips and functionality, no COM port, mic, onboard audio, wifi, bluetooth, IR, graphics (or ability to turn of in BIOS)

 

- PCIe slot(s) for USB to DAC/SPDIF Converter through PCIe dedicated card (PPA,SoTM)

 

- direct multi-rail low noise linear PSUs to avoid DC-to-DC conversion circuits and high frequency pulses from PWM. This is a step further from the PicoPSU.

 

- CPU with no video circuitry inside, but instead add a video card for config and then remove and run headless

 

Cheers

[iago]

Even if you elliminate the switching power supplies, you will still be left with digital circuits running at several hundred MHz to a few GHz. The few hundred kHz of a switching converter are harmless compared to this. It might be more beneficial to completely isolate the computing section of a server from the rest of the layout, which is also a design element which you will find in high end music servers and also possibly the reason why some people prefer dual PC configurations.

 

Another limitation is that if you start using custom boards, you will seriously limit your choice of usable mainstream software. I doubt that you will be able to run something like Windows Server on a custom main board. Maybe Windows XP, because it is also used in embedded aplications. Fortunately there is Linux support for most hardware architectures.

 

A very interesting alternative is the embedded market. If I look at your wish list, I could immediately put check marks against all of them (the linear PSU would be an add on component). I would at least keep the COM port, because it is useful for debugging.

 

Due to smaller production runs and extended life cycles, the boards are more expensive than the consumer versions, but still affordable:

 

DigiKey | Single Board- Cmputers

[Sam Lord]

What about the Biostar audio-specific motherboards like the new Hi-Fi Z97WE?

 

"...The audio connectors are driven by improved audio hardware that is named "PURO Audio," and as part of the package users will find better capacitors along with a headphone amplifier, all on a separated part of the PCB..."

 

See:

Biostar's Flagship Z97 Motherboard Debuts

[iago]

What about the Biostar audio-specific motherboards like the new Hi-Fi Z97WE?

 

These boards are targeting the gaming market, it seems to become important to achive higher audio quality. There are more models on the market, and the use of separate headphone amplifiers seems to become a trend: Gigabyte.

 

Something like this might be an interesting solution for an entry level HTPC, but size and feature richness make it uninteresting for an audio-only setup.

 

Personally, I find the SoC (system on chip) based designs quite interesting. You will find them in entry level notebooks and tablets, but - more interesting - in industry solutions. These usually have better long term availability than consumer boards. Supermicro has some interesting designs; they even come with Windows drivers ;-)

[Sam Lord]

Ah, thanks iago.

 

I wondered what comprised their market. I'm looking to purchase a fairly powerful machine to enable use of a strong resampling engine like HQPlayer. I would use an ethernet streamer like the SoTM SMS100 to keep fan and/or disc noise and heat away from audio circuits. I expect to convert from 16/44 to DSD128 or 256 beforehand or on the fly. Supposedly that improves sound over redbook conversion in many cases.

[iago]

I would use an ethernet streamer like the SoTM SMS100 to keep fan and/or disc noise and heat away from audio circuits.

 

If you separate playback from processing hardware, you will definitely not benefit from enhancements of the audio output circuits. Probably the mainboard choice becomes secondary too. I would select a mainboard based on features (size, geometry {for passive cooling}, power supply options, necessary interfaces, BGA version) and convenience, possibly in that order. You could start with an existing machine as proof-of-concept. I usually have a medium priced laptop around as test environment.

[tranz]

Hi iago and Sam Lord.

 

Thanks for the suggestions.

 

The gigabyte for audio mobo I looked at, but I agree you do not want to use the mobo's audio circuit. The improved USB connection is interesting, but will not be close to a PPA or SoTM solution. A non-starter for this function.

 

A little side note on fans. They have no business in an optimal audio server solution, the least of the problem being the vibration, mechanical and wind noise. EMI and the constant varying electrical pull were in my experience detrimental.

[Superdad]

A little side note on fans. They have no business in an optimal audio server solution, the least of the problem being the vibration, mechanical and wind noise. EMI and the constant varying electrical pull were in my experience detrimental.

 

What seems to make fans so detrimental to SQ is that most of the ones used in computers these days are the 3 or 4-wire variety where the fan speed is controlled not by a clean and even DC voltage, but rather by PWM pulses. This is done so that the computer does not have to have a hot, power consuming variable linear regulator for the fan. And while of course most of the other DC-DC regulators our motherboards are filled with (which those of you trying to inject specific DC voltages are trying to bypass), some of the fans draw some serious current (and over wires that cross over other circuits), so their high-current PWM pulses are particularly unwelcome.

 

Just FYI because there is Mac discussion here and Tranz, myself, and others are doing mini mods: Apple is unique (at least the Mac mini is) in that they put the PWM controller chip somewhere inside the fan itself, whereas on most PC motherboards the PWM controller for the fan voltage is on the motherboard. John Swenson has designed for me a linear fan controller/filter circuit to be part of our Mac mini DC conversion board (it eliminates the PWM pulses but still allows the Mac to read the tachometer line and control the fan speed to suit thermal needs). When I asked him if the same circuit, which uses a servo controller op-amp, could be adopted for us to offer a linear controller/filter for general us with CAPS motherboards (ala SOtM's fan filter but with linear voltage control), that's when he explained how PC motherboards are different in that the PWM for the fan is on the motherboard and not possible to easily circumvent.

 

BTW, prototype boards to test John's linear fan control circuit have already been ordered (I got all the parts, micro connectors, and an OEM Mac mini fan sent to hims last week). His plan is to first measure the noise/radiation/pulses grunge from the stock PWM hook-up; then from the fan run on straight DC (at say 5V, 7V, 9V,12V--he'll just use the prototype of our JS-2 which has those voltages via switch setting); and finally with his controller circuit. For marketing purposes I asked him to try to capture some of those measurements (ideally in visual form, but if not, then even just in notes). I still wonder if there isn't some way to do this for non-Mac motherboards, but right now my focus is on completing the board for the mini so it will be ready in time to offer with the JS-2--which I start fully assembling late the first week of June (chassis, boards, and R-cores all ship to me from various corners of the globe on the 30th).

 

Ciao,

 

Alex C.

[Sam Lord]

If you separate playback from processing hardware, you will definitely not benefit from enhancements of the audio output circuits...

Thanks yes I knew that, the Biostar question was not for my needs but for others doing their own CAPS server. The later comments on typical mobo fan electrical noise are very useful, I'll pay attention to that aspect. A pre-built Zuma is my likely direction.

[JJJ]

I just wonder how many of you worried about little things such as PC motherboard EMI have examined, measured and sought to deal with the big issues such as room acoustics and it's influence on what you here. There are profound gains to be made there and so much more so than worrying about many of the topics discussed here. A few RPG Modex Plates, Bad Panels and Skylines coupled with implementation of Dirac Live will have considerably more impact on what you actually hear.

[Sam Lord]

...A few RPG Modex Plates, Bad Panels and Skylines coupled with implementation of Dirac Live will have considerably more impact on what you actually hear.

 

Sometimes yes, sometimes no. In my experience speaker alignment and positioning is far more important than room treatment or EQ for most rooms, and I have far more experience than most folks in that department. Would you rather have decent monitors in a perfect room or great monitors in an average, untreated room? Engineers would certainly choose the former by necessity, but most music lovers would choose the latter, even after comparing the two for a year. Some distortions are vastly more psychoacoustically important than others, IME. For the same reasons, I would choose a good vinyl system and a good record collection in an average room any day over an exclusively redbook audio system in an ideal room. My noise and distortion quantities from the usual culprits would be tens or hundreds of times greater, but I would be happier hearing music.

 

But you naturally wonder, why do I have an all-redbook system now (in my profile)? What hypocrisy! The reasons: 1) Vinyl is too hard for me to manage regularly because of my disability, so I donated my record collection to a music school; and 2) I am able and willing to create a digital playback system that sings. That said, I will be getting some basic 1st reflection attenuation and will* EQ my subs after I get the testing software up. I used to work in an average-sized but well-proportioned room with a raised double floor and 20 tons of sand between the double walls. I do miss that room...

 

You properly raise the oft-repeated issue: How can artifacts below -120dB (or lower?) possibly matter in an environment with ambient noise and distortion energy a *million* times higher? My answer is that our hearing is simply amazing, and the phenomenon is not only the result of expectation bias.

 

*Er, I might. The blending of subs is nearly ideal now, and the EQ needed would be very mild. The room actually has a pretty good response. It is too narrow but has a large opening that greatly mitigates the typical low-bass resonance.

[tranz]

Hi JJJ,

 

Thanks for the input. Indeed those are very important. I guess from my listening experience I split the impact into two loosely defined separate camps.

 

1. Musical impact.

E.g. Stereo imaging, sound stage, accuracy of reproduction of the recording, warmth, etc.

 

2. Reduction in digititus/listening fatigue.

E.g. that feeling of not wanting to stop listening. Anything sounds lush. As opposed to feeling irritated, tired, wanting to lower the volume or relief when turned off. Like listening live or to an extraordinarily expensive vinyl setup without hiss/pops/clicks. Your ears and brain are not distracted by digital hash and can focus on just music.

 

To my experience anything in the computer realm has had impact in #2, reduction of digititus. And it is very hard to eliminate.

 

Lowering as much electrical noise upstream helps, which includes all actions many feel are nuts: mobo selection, lpsu multi rail, shielded SATA, CAD or AO scripts to lower background processes, 3M sheets, underclocking, etc.

 

The #1 tweaking I find is more impacted by speaker placement, room acoustics etc. Very important! But if you have fatiguing electrical noise impact coming through the speakers, it really does not matter what degree of toe in you have; you will still get irritated.

 

By the way I did try Dirac and it did more harm than good to the sound in my setup.

 

Cheers

[JJJ]

Did you examine your system via REW (or similar) before and afterwards?