Simple DIY: SPDIF output from Mobo

[Santy]

I had this RCA socket lying with me for a long time, decided to try the SPDIF experiment which I had in my mind.

I guess this can be done on any mother board with SPDIF header pins. Useful for attaching entry level DACs without USB or to save a USB port of PC.

I opened the case and chopped the 3 pin header wire off a cabinet fan (it still runs through 4 pin molex connector). This is also available from computer shops.

RCA female plug was bought from DIYaudiocart

Inserted the wire through the rubber sleeve of the RCA female connector (MX) and soldered it.

Made a hole through the PC back plate (closest to the SPDIF pins and inserted the socket through it.

Connected the back plate and plugged it on the header pins on mobo.

Paired the DAC to it through a coaxial cable.

Fired up the PC, DAC and everything that's needed to play music. Set the audio output device on PC to "Onboard SPDIF". Chose Coaxial input on DAC and bingo, it worked. I reversed the polarity on the header (grnd & line) just to check and it didn't work.

Now after lot of quick A/B switching, between USB and DAC (need to change inputs on DAC and output device on PC/ player), I am unable to find any difference in sound. They were sounding absolutely same to my ears. Be it freq response or resolution or layering or soundstage, all were same. May be the DAC was not capable enough of revealing the differences. Or may be I didn't listen critically or long enough. Or may be there was no difference. But this digital output just worked fine.

Total cost: Rs. 250
Total time: 25 min

 

[baijuxavior]

Good guide for those who are trying to get digital out from mobo without buying additional sound card or spdif board.

 

[Santy]

I realize that there are a few more situations wherein this could be useful:

To connect a non-USB external DAC to a HTPC (already mentioned)
To connect a HTPC to a old AVR without HDMI
To connect a AVR to a PC without HDMI
To connect a HTPC to a CDP with digital input (without USB) and use it as DAC
Ofcourse only if you don't want to spend a sound card.

 

[ranjeetrain]

Now, that's what they call DIY. Very indigenous

 

[Thad E Ginathom]

Nice idea and nice execution. It probably would not be much more work to fit an actual RCA socket to the bracket?

Another idea that occurs to me now is to take an ordinary RCA/RCA cable, cut off one RCA, and terminate the connection with a header plug. How are those flat header plugs made? I suppose the ones we see would be factory terminated moulded?

Ofcourse only if you don't want to spend a sound card.

I've always thought it rather a waste to buy a good sound card and use only the digital out. However, Clifford of Genelec, who is a very down-to-earth engineer, told me that he certainly prefers not to use the motherboard SPDIF. Further, more recent reading shows that, across a range of tested devices, digital output is not all equal. I'll add the link for that if I find it. EDIT: It was one of Archimago's Musings

 

[Santy]

Nice idea and nice execution. It probably would not be much more work to fit an actual RCA socket to the bracket?

I did consider but didn't get a good quality RCA connector; the ones meant for cabinets.

Another idea that occurs to me now is to take an ordinary RCA/RCA cable, cut off one RCA, and terminate the connection with a header plug. How are those flat header plugs made? I suppose the ones we see would be factory terminated moulded?

Yes but the challenge is to insert the wire into that header plug and to secure it in its position. I guess you can pull out the metal pin from the female header socket, insert the wire through the pin, crimp it tight and put it back into the plastic head. The RCA wire that you have cut could be too big to fit I believe.

I've always thought it rather a waste to buy a good sound card and use only the digital out. However, Clifford of Genelec, who is a very down-to-earth engineer, told me that he certainly prefers not to use the motherboard SPDIF. Further, more recent reading shows that, across a range of tested devices, digital output is not all equal. I'll add the link for that if I find it. EDIT: It was one of Archimago's Musings

Of course there WILL be difference but whether it is perceivable or not is the question. For instance, 0.4db tilt beyond 20k+ khz as measured in that experiment is something too microscopic for the 'normal' ear to differentiate according to me . All articles and advises are digestible only till you test it or hear it yourself. There are plenty of articles which claim that SPDIF output even from sound card cannot better a USB output.

 

[Thad E Ginathom]

Yes, a lot of these things are academic. We like to get them "right" if we can, and each one that we can tick that box for is, at least, one less thing on the trouble-shooting list if there needs to be one.

Ah, those header plugs are crimped. Forgive me for being to lazy to look! In that case, the special little tool for insertion/removal may be needed, and crimping is probably not reliable unless done with the proper tool for that particular pin. I used to do computer serial cables.

There are plenty of articles which claim that SPDIF output even from sound card cannot better a USB output.

I "claim" () that the specific DAC implementation and circuitry is more important than the technology. I make this claim only of the basis of reading, though ( )

 

[reignofchaos]

^^Very true. Implementation matters more than anything. That unfortunately is also what makes motherboard spdif headers outrageously bad - they are best not used as they sound like crap. Spdif out of a realtek chip that is the most common chip on motherboards is not worth using unless you have very low end equipment.

 

[Santy]

I have to admit that USB audio is coming straight from the heavens and not from the mother board, and completely devoid of noise or jitter.
The music is streamed straight from the hard disk magically onto the USB cable, it appears. :lol:

Low end or mid-end or high end, as long as both the inputs are tested and found to be same to the one to whom it matters, why should one disregard a particular connection with a 'prejudiced' opinion in mind is not clear to me.

I "claim" () that the specific DAC implementation and circuitry is more important than the technology. I make this claim only of the basis of reading, though ( )

Of course yes! The implementation is more important than the technology as well as the D/A chip that is used (which generally is given lot of weightage). That is why different DACs using same chip sound different. I have seen AVRs and dedicated DACs using same chip but sounding quite different. I guess its proven. I fully agree to it.

At the same time, implementation of an SPDIF input on a DAC can be as good as its USB input as well. May be that is one reason why I could not make out any difference. Why to rule out the possibility that USB and SPDIF can sound same on a particular configuration?

Whatever, I dont think it (SPDIF) sounds like "crap" on my Genelecs compared to the USB counterpart. If it did, I would not bother to post this thread. Moreover, I have seen such terms are used more often only by 'audiophiles' who rely on armchair research. Otherwise there are a thousand different ways one can differentiate between the sound heard from two different systems/ configurations.

 

[Thad E Ginathom]

You both have valid points of view, although I wonder about taking it as far as "absolute crap." I wonder, because even the analogue out from the MB sounds so much better than the equivalent of a few years ago. I wonder --- but I haven't tried, because I have nothing with an SPDIF input, So nothing dogmatic from me on this topic :lol: --- only armchair research

Listening through Genelec monitors is probably as tough a test as any, I guess.

 

[Santy]

If difference in sound quality in analog section (DAC to our ears) can be measured in miles, then that in digital section (source to DAC) could be measured in feet, I believe. IMO, there are a thousand factors that can mess with the analog sound and a dozen parameters affecting the signals in digital domain. The implementation on DAC side is more important since that is where clocking and clean-up happens. I do not deny the fact that SPDIF quality matters, when the DAC can appreciate it to that extent. For instance, buying a quality sound card (like Essence STX) to connect the PC to the non-USB DAC should be justified by the cost and performance of the DAC.

This is not a suggestion for everyone. Onboard SPDIF is a solution for those who have a decent non-USB DAC costing X and do not want to spend X or even 0.5X on a sound card just for its 'digital out'. I hope its loud and clear. A sound card's primary purpose is the D/A conversion. So I guess more than half the money we pay for it is unused if only the digital connection is used. This is just a cost effective alternative. May be it cannot tap your system to its fullest potential but I promise its nowhere close to 'crap'.

 

[aramosfet]

I have done the same thing, However, the SPDIF output on the motherboard is a 5V TTL signal. But most inputs on our AVRs are Co-axial. Hence they may not be "impedance matched". There's a simple circuit here --> epanorama.net/S/PDIF Interface to convert TTL to Co-axial signal.

Look at "Simplest TTL to S/PDIF coax interface" Doing this will be safer for the Receivers, Not all Co-ax SPDIF inputs may be able to handle 5V. And also less likely to be affected by noise if you are using a long SPDIF cable.

Regards,

 

[Santy]

I will measure the output voltage on these pins and revert. May be the SPDIF brackets sold online does this job of conversion? The article says compatibility could be an issue but it doesn't say its implications on sound quality. Did you meant to say a high current signal (as in this case) is less susceptible to external noise? BTW I wonder why the mobo manufacturer will call it as SPDIF if it doesn't meet the standards. The manual unfortunately dosen't give the ratings of this output.

EDIT: Its no where close to 5V.

 

[aramosfet]

Yes, A good quality SPDIF bracket will do this conversion. The Susceptibility to external noise is due to the Impedance mismatch. The Co-ax inputs will be 75ohms on the receiver. The output impedance on the transmit should also be 75ohms. If you have a long SPDIF cable, It will be less susceptible to external noise when impedance is matched. I don't think the 5V would damage the SPDIF input. Most recievers will have some form of protection, But its better to be on the safer side and conform to the standard.

How will you measure the Voltage? you will need a CRO/Oscilloscope since a multimeter will only show the RMS value. But i'm pretty certain that the Motherboard output will be 5V TTL. This is done to support having Toslink Interface which would require the input to be 5V.

Regards,

 

[reignofchaos]

Whatever, I dont think it (SPDIF) sounds like "crap" on my Genelecs compared to the USB counterpart. If it did, I would not bother to post this thread. Moreover, I have seen such terms are used more often only by 'audiophiles' who rely on armchair research. Otherwise there are a thousand different ways one can differentiate between the sound heard from two different systems/ configurations.

Apologies if it sounded harsh but have been through the whole wheel of upgrades from basic motherboard audio all the way to an m2tech evo now. In the middle I've been through a dozen different spdif outputs and my personal experience says that realtek sounds the lousiest of the lot. A few older intel oem boards came with analog devices chips which sounded very good but these are rare to come by.

Also I think realtek's spdif out is definitely not bit perfect. There's some major resampling happening in the middle. Also let me not talk about jitter due to poor voltage regulation on a motherboard. If the board has optical out, it is better to use optical over coax just for this reason even though on paper coax is better. Also as aramosfet mentioned, the motherboard spdif out is 5V p-p not the usual 2v p-p.

 

[aramosfet]

Apologies if it sounded harsh but have been through the whole wheel of upgrades from basic motherboard audio all the way to an m2tech evo now. In the middle I've been through a dozen different spdif outputs and my personal experience says that realtek sounds the lousiest of the lot. A few older intel oem boards came with analog devices chips which sounded very good but these are rare to come by.

Also I think realtek's spdif out is definitely not bit perfect. There's some major resampling happening in the middle. Also let me not talk about jitter due to poor voltage regulation on a motherboard. If the board has optical out, it is better to use optical over coax just for this reason even though on paper coax is better. Also as aramosfet mentioned, the motherboard spdif out is 5V p-p not the usual 2v p-p.

I don't think optical would be any improvement over coax with respect to jitter. If your source has high jitter it will eventually be seen at the receiver irrespective of Coax or Optical. However almost all SPDIF receivers use a PLL to recover to clock and eliminate jitter. So the output will largely depend on the receiver PLL design and PLL clock jitter. The transmitter SPDIF clock is only needed to track the sampling rate clock which may vary in every system to some extent.

I agree that some resampling happens in realtek or the audio drivers. It may not be bit perfect.

Regards,

 

[aramosfet]

I will measure the output voltage on these pins and revert. May be the SPDIF brackets sold online does this job of conversion? The article says compatibility could be an issue but it doesn't say its implications on sound quality. Did you meant to say a high current signal (as in this case) is less susceptible to external noise? BTW I wonder why the mobo manufacturer will call it as SPDIF if it doesn't meet the standards. The manual unfortunately dosen't give the ratings of this output.

EDIT: Its no where close to 5V.

Yes, the multimeter will only show the RMS value. Mine shows 1.7V and its 3.3Vp-p . So yours also should be 3.3Vp-p.

 

[Santy]

Yes right, it will give only RMS value and not the pulse voltage. Thanks for correcting me. For a moment, you took me to my college days.

I understand that the SPDIF header is usually meant for the older graphic cards which do not process audio and not actually meant for external DAC use. It was an experiment for fun that actually worked. It can be a cheaper alternative to a sound card. That is why I highlighted the cost of this project (Rs.250). There are chances that cheap sound cards may not be bit perfect either.

If you read the post, I did not rate or define the sound quality. I just stated that I found no difference between USB and SPDIF. It could be because SPDIF implementation of DAC is better or USB implementation is poor or the coarse PSU or the overall transport being messy or may be wrong USB streaming settings. It could also because USB cable is picking up stray noise from several power lines behind the rack more than the fairly insulated SPDIF cable. Well, I am not sure.

The DAC (Caiman Gatorized) however, is not really low end; everyone who listened it have appreciated it and I am yet to come across a better DAC at 2x of its price.

Since recently I am trying to figure out ways of improving sound quality from USB of my PC apart from DAC upgrade, which is also on the cards. A good DAC can do the cleanup only to a limit. I think I am aware it cannot purify all the dirt generated by the source.

I hooked up the CA 640C V2 which I acquired yesterday as a transport with the DAC and was blown away by the clarity and definition. The noise floor that I am used to simply vanished and the dynamic range improved a lot. I fell in love with my DAC once again.

So this suggestion of onboard SPDIF is just a relative solution, a pitstop for those who need SPDIF output but need to save up for an upgrade. Its acceptance depends on which orbit we are in (tolerance and sharpness of ears), what we have and what we want. Check out my signature, you will know what I mean.

 

[Thad E Ginathom]

Gosh, you've got an oscilloscope! I'm impressed

You can probably tell much more about the output of your SPDIFF than by simply listening. I wish I could read such things and understand them!

Browsing, yesterday, i ended up watching a whole hour of video about an oscilloscope. Most of the stuff meant nothing to me at all, but sometimes, people who really know what they are talking about, and have a great enthusiasm for it, have a charisma that is hard to ignore. Also, it was a half-million dollar 'scope . He was demonstrating stuff with gigahertz waveforms.

If you like tooldrooling, you might enjoy: Playing with high end oscilloscope. And it'll probably mean more to you than it did to me

 

[silking]

Santy sir it is useful to me, I have dac w/o usb.Can it be used in my pc? .I'm in idea of a soundcard to improve sq.