Xonar ST/STX mods...

Member
Joined 2009
Paid Member
1. Which caps should be replaced if I only plan to use the digital coax out?
2. Has anyone replaced the digital out with a WBT 0210Ag?

It is not necessary any modification if you intend to use only the digital out on this card. The mods about Xonar (ST/STX) are mainly meant about the analogue stage of this card, when is to use the analogue out (RCA outputs)...
But buying a such card to use only its digital in/out do not make much sense either...
Any type of sound card is very good if one only want to use its digital interface. The only problem is to find that card which can output a digital max samplings frequency (192Khz). If so, then are not so many which can do it. Xonar series can very well this... The case of mods is about the analogue stage and its digital to analogue converter used for have the best out of a such card.
 
It is not necessary any modification if you intend to use only the digital out on this card. The mods about Xonar (ST/STX) are mainly meant about the analogue stage of this card, when is to use the analogue out (RCA outputs)...
But buying a such card to use only its digital in/out do not make much sense either...
Any type of sound card is very good if one only want to use its digital interface. The only problem is to find that card which can output a digital max samplings frequency (192Khz). If so, then are not so many which can do it. Xonar series can very well this... The case of mods is about the analogue stage and its digital to analogue converter used for have the best out of a such card.

I have not tried any lower end (i.e. not so expensive) cards in the Xonar lineup, but from all that I have read, people can even hear a difference in the sound quality of the digital only between the ST and STX, with the ST having a better quality. Looking at it from outside of the computer, lets take cheap DVD players vs. very expensive standalone transports. How many would ever say the digital out is no different on a free to good home DVD player vs. some $50K transport known to be the very best transport out there. The primary thing that gets argued is jitter, and the ST with the CS2000 is supposed to take the jitter down over say, the STX, that probably has lower jitter somehow than the cheaper cards...same deal goes with transports...dvd and cheaper or even expensive cd players/transports can have terrible jitter specs while those raved about have the very lowest jitter specs.

I have no answer since I have limited experience with sources.
 
Member
Joined 2009
Paid Member
In the case of a digital output from ST/STX or other cards, is right that the jitter is or could be important. The jitter of one type card or another make the difference in the sound quality. This is right too.
The question(s) answered in that post refer to some capacitors in this digital stage which if changed, (supposable) will improve that digital out.
Changing the capacitors will not do so much improvement referring to a jitter problem/level. To improve the jitter level (lowered it) in such cases is about working with the oscillator on that card (changing it out or better regulator for it).
Personally, I just ignored the digital stage of my STX, and I`ve been focused in the improvements concerning the analogue stage (oscillator improvement it was included too in those mods, as to be seen here).
 
In the case of a digital output from ST/STX or other cards, is right that the jitter is or could be important. The jitter of one type card or another make the difference in the sound quality. This is right too.
The question(s) answered in that post refer to some capacitors in this digital stage which if changed, (supposable) will improve that digital out.
Changing the capacitors will not do so much improvement referring to a jitter problem/level. To improve the jitter level (lowered it) in such cases is about working with the oscillator on that card (changing it out or better regulator for it).
Personally, I just ignored the digital stage of my STX, and I`ve been focused in the improvements concerning the analogue stage (oscillator improvement it was included too in those mods, as to be seen here).

Thanks Coris. So changing the caps to something better or even larger values may help with the digital out, though that's basically up in the air as you put it. Makes me wonder just how these various sound cards sound with the digital out only and if there are real/true differences that can easily be heard or if it's more or less just a matter of preference. I have seen too much of the arguments going both ways about sound cards, usb converters, and so on as the source to an external dac, but I have used a number of USB converters, stock Xonar ST, and frankly, I honestly cannot put a finger on the differences. I don't consider my dac to be some great device that eliminates the jitter, but it's obviously something about the way the dac is designed that does the majority of what I am hearing. I use preamp+dac in one unit, so a 2" wire from analog line stage output to my digital dac chip's input inside one box. I REALLY question whether what people are hearing with the differences of the USB converters or a sound card (since some do prefer the sound card's digital out to even very expensive USB converters, much the same as my own experience with the ST vs. say, the Audiophilleo), but I don't hear such a difference where it's like, ahhh, name your adjectives, etc. etc. etc. So I'm thinking along the same lines as you are regarding the caps, and just thinking that the "majority" of this transport vs. that transport really is system dependent, where in my system, they are so mild in differences that I can essentially choose whichever one and stick to it. I just have never used a very cheap/inexpensive sound card. Just ST, and the more expensive USB converters shy of the most expensive ones of course like Empirical for $1300:)).

Maybe it also has a lot to do with the fact that a computer is not exactly a great source, and the primary thing that makes it a great source is all the software...in other words, hardware side of things is noise noise noise, but software side of things is something other devices just cannot do...well servers, but those are computers too...standalone, I can think of say a nice up converting standalone cdp, and tapping in my preamp section in place of the standalone cdp's analog section...then that could be competitive, perhaps even better? But the computer software side is so powerful that it really seems to eliminate all of these "noise" issues and make probably a super cheap sound card sound just fine as a digital source so long as you have a nice front end including the dac in the equation. Others will obviously disagree saying you need the fanciest transport device, but I'm still not convinced having heard up to $600 price range USB converters and Xonar ST should be darn good enough shy of the uber expensive "pro" sound cards.
 
Member
Joined 2009
Paid Member
The clue to using digital out of a sound card (or another source) is the posibility to carry the signal over quite long distances (meters...). But using the digital out of a source, mean too that the original signal have to suffer of many conversions, adding more jitter from the involved stages, and so on. Is not more reasonable to just take out the analogue signal from the output DAC on that source?
What mean using an external DAC? The I2S output from processor is converted to SPDIF standard and outputted so via Toslink or coax... Then SPDIF is converted again in the external DAC to get analogue out... What about jitter in all this process?
The audio card have already its local DAC, which produce an analogue out. Using an external DAC with SPDIF from an audio card is not a very big clue in my opinion. The best is using I2S output from the audio processor. But this is not possible over long distances (but only few cm). I just preffered to use the analogue out and the existing DAC (PCM179, which is a very good one) on the ST/STX.
The noises in the computer come essencially from the switching power and lot of high frequencies digital signals inside. What about to isolate the audio card power system from the computer switching power? And shield accordingly this sound card inside the computer? I just did like this on my STX. My sound card is now feed it from a completely different power system. I can now even switch to a battery to supply my STX with power, but is not big improvement in the sound doing so...
If one care a little bit more to improve the clock for the sound processor/DAC, then the jitter in the audio system is lowered quite much.
I can not say that a computer is a bead audio source. IS NOT! The only problem is (were for me) that the DAC on the sound card is not at last as performant as I wanted. Now is the ESS9018 which is the best! That because I intend to use the computer for tapping the I2S from the processor, and use this interface to a more performant DAC system. Another good solution is ExaU2I through USB and an external DAC. I work on this now... Else I can say that a computer is best as source and versatility (except the transport for CD/Blu Ray, etc)
The main improvements for the sound card (ST/STX) is the power system, and the oscillator/clock. There is not only to change the decoupling capacitors on it... This is only a part of the mods, but isolate the card from the computer power system is a must for a big improvement, and change the oscillator with a lowest as possible jitter one is a must too...
Is also essential to have a good amplifier/speaker system to hear the difference in between sound cards or improvements on an ST/STX card. I do not agree that the software in a computer lower the hardware jitter, but are some techniques to improve the clock system by software. Very important is a good motherboard, with a performant power system on it (mine have 48 parallel PSUs only dedicated to the CPU), and not at last, a performant processor (I have an i7 quad on mine).
 
Last edited:
Member
Joined 2009
Paid Member
Is always positive to experiment... So just do it and her the result.
In my opinion, the less capacitors in the signal path is better, as the large as possible and good quality (no ESR/ESI) for decoupling/bypass to each of the opamps power pins.
I can see in your design quite much capacitors and quite high capacity values. I used in my I/V converter few pF and few pF in the final opamp stage. No any filter at all. If you will place a such 200pF or more in the I/V stage, you may have in mind that such cap will cut quite much of the high frequencies. Just use first that schematic as it is, try afterwords to lower the caps values, and hear the difference. Solder the resistors (SMD best) and caps right on the opamp pins and make the connections as short as possible, or just solder the opamps together using only theirs pins. This bring the best result. My I/V converter is not more than 20 mm long and almost the same bread... Take care of the GND connections and maybe an accordingly shielding may be taken in to consideration....
On the original ST/STX are many small caps on the signal path and some ferrite beads on the output lines. I will recommend to remove all those and output the signal directly as possible in to the amplifier. And keep also in mind: so long one keep the power in the right limits for the active components, is nothing to be wrong or to be burned out. The worst thing if is something wrong when experimenting in the signal path is only: no sound... This can easily be fixed.
I will say at last that personally, I did not find the right way from the beginning, or reading papers only. I`ve experimented a lot to find the best way. So, DIY!
 
Last edited:
I can see in your design quite much capacitors and quite high capacity values. I used in my I/V converter few pF and few pF in the final opamp stage.

I'm amazed at your final mod of the card, but I've done smth as well. I use external supplies for all consumers, even +3,3V.
As for analog stage-Do you really think my values of caps are that much? Original caps in final stage were 2700 pf. Could you give the actual circuit of your output.
 
Member
Joined 2009
Paid Member
You are doing right about power supplies.
About the analogue stage I will want to use like this (picture) your design (it depends a little about what opamp you will use). You will find a picture with my I/V - final stage on the 1 page of this thread.
But anyway, just experiment first as you thought and then try to use my proposal. You will hear very well the difference...
If you will build those stages compact (module) on a very little board and only connect to the rest the in, out and supplies, will be easy to take all out an modify/experiment afterwords...
They who original designed this stage, and used 2700p caps and filters, like more to hear the math not the sounds coming out of such stage...:)
 
Well, I just don't know what to say, your output stage doesn't seem to filter anything, but on the other hand, it provides linear phase and good impulse response.It must sound perfect but only with a tube amp, where the transformer is like low-pass filter. And I would not connect it to a transistor amp with deep negative feedback as it might oscilate and damage or even burn down tweeters.
By the way, just noticed u are from Norway. I was there lust year. Wish I listened to your tweaked card then.
 
Member
Joined 2009
Paid Member
Yes, it is not any filtering in my proposal. Do you need filtering? For what? The low pass filter after I/V convertor is meant to filter out quite high frequencies which come from the DAC chip as residual for the converting process. But a quite simple filtering will alter very much those frequencies involved in sound/music. Will also alter the phase and the response of the system... By the way, the phase accuracy is crucial for the sound stage and detailed sound of a sound system. Such DAC residual frequencies are so high and so low power at your amplifier will never succeed to amplify it. In opposite, such frequencies will be strong attenuated by the further power amplifier circuits, and first by the cables it self between source/computer and power amplifier... By the way, a such middle to high quality cable, few meters long can rich a capacity of 100 - 200pF... There are solutions to get 20 - 50pF on a similar lengh cable, but this is another story...
A deep negative feedback meaning is also to prevent oscillations... Oscillations occur when is about a positive feedback in a system...
Here is the clue I pointed in my last post about short connections, and very carefully placement of the components in making the I/V - final stages, and not at last local shielding...
I use not any filter in my modded STX after I/V stage. I have a 26 -28vpp signal swing on RCA output of the sound card. I have this signal on a 100k potmeter, and then directly in an 3886 based amplifier. By the way, my cable between computer and amplifier (a little over one meter is not shielded at all...) I have not any oscillations in my system, no any distortions or noises, and I never destroyed my tweeters, even though I`ve experimented a lot... I have for now only very clean, detailed, high dynamic, and beautiful sound out of my (computer) system... So beautiful sound as an PCM1792 can give...

You do not have to adopt my proposal, if you do not trust it... But trying it will (very much) surprise you...

You are welcome to listen my card if you will come back to Norway...:)
 
Last edited:
Hello Coris

I have some questions about your external PSU.
Can you explain a little bit about yor 5V and 3.3V supply ? How much is the power consumption on the 5 and 3.3 rails ?
What traces must be cut(or desoldered) and where are the supply rails soldered on the board(3,3V,5V and ground) ?

I'am very happy about some more Pictures...
(sorry for my bad english...)

Thanks,

Frog...

P.S. A last question, when using a external PSU, the molex connector must be used too ?
 
Member
Joined 2009
Paid Member
Hello Coris

I have some questions about your external PSU.
Can you explain a little bit about yor 5V and 3.3V supply ? How much is the power consumption on the 5 and 3.3 rails ?
What traces must be cut(or desoldered) and where are the supply rails soldered on the board(3,3V,5V and ground) ?

I'am very happy about some more Pictures...
(sorry for my bad english...)

Thanks,

Frog...

P.S. A last question, when using a external PSU, the molex connector must be used too ?

Hi Frog
I`ve already published here some pictures of my dedicated PSU for the STX board. If you will go back in this thread, I think you can find those pictures and some text explanations. But I can attach here again some... One of those pictures show an 3,2W toroid transformer, so one can easy conclude that is not about important power here...
Anyway, I want to underline again that when I refer to an external PSU, I mean an PSU which will deliver power (+/-15vDC) exclusively to the final opamp stage and the I/V converter. I`ve used for this PSU LM2941/2991 regulators.
I`ve used 12vDC from the computer through Molex connector. So, yes the Molex connector have to be in place. Cut it out the 5v connection from that connector. The bords circuits need also this tension as it is, because it is controlled and used for the logic inside the board (Cmedia processor). If this tension is not seen by the processor, then error messages occur. I think that I`ve explained here (earlier posts) how is to be cut it the power on this rail so to not generate error messages from the board main processor. It is about to cut the power only for the I/V converter and the final stage on the board, to make possible to supply those stages from a better (external) PSU.
I`ve used computer`s 12v rail to create on the board 5v rail and 3,3v too. Please go back an read earlier posts where is described what about 3,3v rail and the components to be removed to isolate compleatly the board from the rest of the computer power system. I can come back to this step if you will feel the need for more explanations.
I`ve used local regulators to create both 5v and 3,3v. Keep in mind that 3,3v rail need a little more power, because the CM processor is to be feed it from this too. I did not measure the current need it by the involved stages, but I`ve used such regulators which can much more than the need of the involved chips. Here is about not more than 200-300mA (3,3v). The 5v is used only for the DAC chip (analogue part of it) and the quartz/clock stage.
Also please keep in mind that the ground plane/connections do not have to be touched or cut it ever. Keep it as it is from the Molex connector too
The aluminium cover of the original board is very good as a heat sink for the small regulators used for both 5 and 3,3v...
The 12v rail from the computer have to be used too to action an relay which will connect the main to the +/-15v PSU in the same time when the computer starts. This is very important! And please be carefully with the 220vAC inside the computer... Much care and good isolations will make everything safe here...
The picture of my final PSU published in an earlier post looks like quite complicated. Is because I`ve used the same board to build up some additional power and switching/relay circuits to supply the STX board from a battery instead for main 220vAC. I have to say that is no any difference between using a battery and an serial/analogue PSU for the resulting sound. So, I do not use now the battery anymore to supply with power my STX. So, let`s forget about battery...
I call this 15v PSU as an external PSU, but in fact is mounted inside the computer, just over the STX board (as to be seen in some pictures here). The STX board it self is not designed to support a quite heavy toroid transformer, so I had to make an supplementary support plate attached to the board, and find a way to support the whole construction quite well to avoid any mechanical tension over the STX board and the PCI Express connector in the mother board of the computer. You can also see how the whole construction look like in an earlier picture here.
I do hope those explanations help now better.
 
Last edited:
Frog and all ...
Coris' (and similar) PSU is/are WAY more trouble than it's/they're worth (time/effort wise)... unless you like to experiment just for the "FUN" (but not necess for the PERFORMANCE) of it!
I have tried many solutions for the Xonar and am most pleased with the better COMMERCIAL "higher-end" PC PSUs, like from Antec, Seasonic, etc.
HardwareSecrets.com conducts very comprehensive OBJECTIVE tests on commercial PSU, so you can choose one with, e.g., the best ripple rejection, and linear current delivery, etc.:
Power Supples: Tests + Articles | Hardware Secrets
 
Last edited:
So help us understand!

Sorry for you, but you do not understand what this is all about... So, be happy in your world!
So help us understand in context, manner, methodology as comprehensive is this:
Hardware Secrets Power Supply Test Methodology | Hardware Secrets
Do you conduct truly insightful and OBJECTIVE PSU load tests using scientific instruments such as:
Dongguan Sun Moon Electronics Co.,Ltd-www.sunmoonate.com

I already commented on the potential dangers of floating (or otherwise unreliable) grounds in DIY jobs -- what may have led to my HDD failure.

It's PARAMOUNT for "DIYers" to know some of the POTENTIAL hazards of pure-DIY approach. I also feel that the pure-DIY approach to PC PSU is not all that much worth the effort in ANY case (so not a whole lotta gain WRT sound compared to better comm. PSU, as noted earlier).
As far as having fun, learning, testing, building ... well, get a oscope out and test the comm PSU, read the informative articles on sites like HardwareSecrets (lotta auto didactism there)....this is "DIY" from now on as far as I'm concerned....everything else is UNSKILLED "labor".

[PS: I HATE soldering burns and fumes. LOL!!]
Oh, and ... uh ... one more thing...
 
Last edited:
Hi Coris

Thanks for your explaining.

Ok, i have re-read and will try to explain what you mean.

For seperating the analog outputstage:

The +-15V comes from the external PSU with his own transformer and rectifiers etc.
For this, the output pin of the 7812 and one side from the diode between L20 will be cut/desoldered.

For seperating the 5V for the DAC and Clock:
You cut the 5V cable from the molex connector or cut the outputpin from the 7805(that a little bit unclear) ??
Then insert 5V from your own shuntregulator thats powered from the 12V molex connector, right ? Where is good point to soldering the 5V rail to the board(on the last photo it shows that it goes to small ceramic cap left from the DAC) ? My intension where on the 7805 outputpad...

For seperating the 3.3V Rail:
Desoldering the small ferritebead on the back of the card.
Then soldering the cable from 3.3V shuntregulator on one (positive)solderpad from the cap right from the DAC ?
Sorry, the final picture from yor card is a little bit crazy...

The Essence Soundcard is for my AudioPC (see picture below). At the moment i have installed a RME 9632. The 9632 comes with one (little bit old)AD1842 DAC so i will give the Asus a try because of its more modern DAC and better chances to optimate.

For +-15V rails i think the LM317/337 is a good choice(someone like this one that works for my preamp, but a little bit smaller)

Edit: What is with the extensionport for the H6 modul ? Normaly there must be 3.3V,5V and ground etc. Is this not a good place to put on the cables for 3,3 and 5V rails ?

@Hollowman

A opamp like LM49720 only works at 80% of what he really can when its driving with crappy +-12V...The -12V will create with a switching psuchip from the +12V rail. That is not the best idea...
Same is for the DAC and clock...

Greets,

Frog
 
Last edited: