Hi AR2,
I had a look at the Tent link. The difference is that data source (CD player) and DACs operate at the same frequency at 44.1 kHz. In our case 44.1 kHz data from the CD is up-sampled and re-clocked by the SRC to the internal DCX frequency of 96 kHz. So SRC becomes the new data-source which is absolutely synchronous with the DACs and that's exactly what you are looking for.
Frank
I had a look at the Tent link. The difference is that data source (CD player) and DACs operate at the same frequency at 44.1 kHz. In our case 44.1 kHz data from the CD is up-sampled and re-clocked by the SRC to the internal DCX frequency of 96 kHz. So SRC becomes the new data-source which is absolutely synchronous with the DACs and that's exactly what you are looking for.
Frank
AR2 said:
Hi AR2,
What I mean is related to the last measured setup in your link, where they have a very good DAC clock, and slave the (cheap) transport to that clock. What my point is, when you don't slave the transport to the DAC clock, but let the transport run on its own, you would have the same result. In other words, because the data is re-clocked in the DAC by the excellent DAC clock, it doesn't matter what the jitter was before, from the transport.
An analogy: suppose you want to cut some wood panels for a speaker. You ask the lumber yard to cut them as accurate as they can. Then you take the panels home and you finally cut them exact to size with the best tools and measuring equipment available. Does the final accuracy depend on the accuracy of the initial cut at the lumber yard?
Jan Didden
To be more precise jitter of the data source as well as cabling is important because the PLL must be able to lock stable on a certain frequency (in your case 44.1 kHz), otherwise the SRC which has to calculate the up-sampling factor can't work properly. But data source need not be synchronous at all. So your CD should have at least a good crystal or oscillator and you should use a good and not too long SPDIF cable. That's the disadvantage of up-sampling. The advantage is that the DSP algorithms are able to work better at the higher frequency of 96 kHz. I wouldn’t worry too much about this problem. It’s a minor one.
Re: New to DCX2496
Hi There.
I have a modded DCX and the difference in the sound quality is staggering - even on my cheap amp and speakers.
I believe the best order is this: first mod is for the output stages, next would be the clock and then the power supply.
The December meeting of the Melbourne Audio Club will be the DIY night, and I should have the modded DCX to audition at that meeting.
cheers,
Simon
ttan98 said:
Hi There.
I have a modded DCX and the difference in the sound quality is staggering - even on my cheap amp and speakers.
I believe the best order is this: first mod is for the output stages, next would be the clock and then the power supply.
The December meeting of the Melbourne Audio Club will be the DIY night, and I should have the modded DCX to audition at that meeting.
cheers,
Simon
I'm with you, although I was really surprised about the improvement caused by the SRC/clock mod. I hadn't expected this.
I think it's also worth to improve the power supplies of the DACs and (if there) op amps but I'm not happy with the existing mods. I made Spice simulation and couldn't see a real benefit. Inserting some simple coils between DCX-supply and DSP-board would have a similar result.
I think it's also worth to improve the power supplies of the DACs and (if there) op amps but I'm not happy with the existing mods. I made Spice simulation and couldn't see a real benefit. Inserting some simple coils between DCX-supply and DSP-board would have a similar result.
Re: Re: New to DCX2496
I agree. I have done two output mods to mine, first a passive mod and after that an active output stage because I was not happy with the limited level control options of the stock DCX and the lack of remote control. The active mod was best for me, both in sound quality and ease of use. The active mod also removed almost all the influence of interlinks on the sound. Details are on my website.
Jan Didden
Pulse-R said:
I agree. I have done two output mods to mine, first a passive mod and after that an active output stage because I was not happy with the limited level control options of the stock DCX and the lack of remote control. The active mod was best for me, both in sound quality and ease of use. The active mod also removed almost all the influence of interlinks on the sound. Details are on my website.
Jan Didden
oettle said:
If I'm not wrong the SRC2496 is doing this. When you connect a CD player to it, the signal is upsampled and "reclocked". if you connect the SRC to a DCX using AES/EBU, the signal follow with the clock and the final result will be close to the oettle's mod.
As well, the small oettle's board is better than to stack Behringer products... ;-)
For the mods's order, I personally prefer this: input/output mod (caps and op-amp), power supply, others caps, clock. PSU is really important as a large part of the low noise quality of the device comes from it (specially for the +-15v if used in analog way).
.
I agree with you as well.
For the PSU, I posted this months ago.
# beween the switch mode PSU and a linear PSU on +15v (inside a DEQ).
http://www.diyaudio.com/forums/showthread.php?postid=712050#post712050
.
For the PSU, I posted this months ago.
# beween the switch mode PSU and a linear PSU on +15v (inside a DEQ).
http://www.diyaudio.com/forums/showthread.php?postid=712050#post712050
.
Hi,
The standard DCX already has a PLL, SRC and clock as well and so it’s already able to up-sample and re-clock any incoming data. So my mod isn't a new function at all (it’s even not a new idea except the combination and improved supplies), but an improvement of the SRC and clock.
Because sound quality is always the sum of errors in the chain an additional SEC2496 won't improve something but would worsen it.
By sure I’m not aware of all PSU mods. For those I have seen I would recommend to make some SPICE simulation regarding PSRR and noise for the whole frequency band not only HF noise. Insert 3 simple 100µH coils (about 0.5 ohms) in the +/-15V and +9V rail between the PSU and DSP board and make the same measurements. Surprise, surprise you will end up in a similar result like the PSU mods. More relevant than HF noise is the noise in the audible frequency band (10 Hz to let say 40 kHz). The typical used voltage regulators have a PSRR of about 60 dB and get much worse above 1 kHz. Also output noise of these standard regulators isn’t the very best. So there is plenty of room for new mods. I hope I wasn’t too frankly.
Frank
The standard DCX already has a PLL, SRC and clock as well and so it’s already able to up-sample and re-clock any incoming data. So my mod isn't a new function at all (it’s even not a new idea except the combination and improved supplies), but an improvement of the SRC and clock.
Because sound quality is always the sum of errors in the chain an additional SEC2496 won't improve something but would worsen it.
By sure I’m not aware of all PSU mods. For those I have seen I would recommend to make some SPICE simulation regarding PSRR and noise for the whole frequency band not only HF noise. Insert 3 simple 100µH coils (about 0.5 ohms) in the +/-15V and +9V rail between the PSU and DSP board and make the same measurements. Surprise, surprise you will end up in a similar result like the PSU mods. More relevant than HF noise is the noise in the audible frequency band (10 Hz to let say 40 kHz). The typical used voltage regulators have a PSRR of about 60 dB and get much worse above 1 kHz. Also output noise of these standard regulators isn’t the very best. So there is plenty of room for new mods. I hope I wasn’t too frankly.
Frank
There is a lot of discussion here lately about if and how much any mod improves the DCX. In one regard the DCX is a very very good platform to learn about sound of different analog stages for example. As there is basicly ability to make the output into 3 times stereo output one can easily do a different analog stages for each output pair (or leave one pair for stock version) and then use the DCX as stereo DAC.
This is exactly how I started to do it - I figured that before I dive into an active crossover world at all I would try to make the DCX to be as good stereo DAC as possible.... so I experimented with passive analog stages, JFET followers, transformers, opamps etc etc. And at all times there was 3 different stages active on 3 output pairs and DCX configured FLAT. This way one can switch back and forth between these outputs that have a signal chain EXACTLY the same upto DAC chip output pins.
For all the sceptics here - try to mod one pair of outputs and leave the other stock - comepensate the levels - and A-B listen to those and say you will not hear the difference....... I'm very sure you will.
Ergo
This is exactly how I started to do it - I figured that before I dive into an active crossover world at all I would try to make the DCX to be as good stereo DAC as possible.... so I experimented with passive analog stages, JFET followers, transformers, opamps etc etc. And at all times there was 3 different stages active on 3 output pairs and DCX configured FLAT. This way one can switch back and forth between these outputs that have a signal chain EXACTLY the same upto DAC chip output pins.
For all the sceptics here - try to mod one pair of outputs and leave the other stock - comepensate the levels - and A-B listen to those and say you will not hear the difference....... I'm very sure you will.
Ergo
What did you think?Wich sounded best to you?
Another question:Is there any sonical difference of tentlabs,Selectronic or http://diykits.com.hk/ oscillators,except for the price..
(I mean just the oscillator not the whole clock)
Ryssen,
Well my feelings after all trials are following
* Passive mod is an upgrade compared to original output stage. The datasheet suggests it is able to drive even the passive preamp after it. In reality the one to one comparison to active stages gives afeeling in the end that there is a bottleneck there. The sound is clean and neutral but seems to lack some punch and energy that make music sound alive.
* Transformers (I tried Lundahl) are very good sounding, but as I'm also a kind of nerd and also look at objective measurements I never got the spectrum to look "clean" with transformers. Being a digital / analog mixed design device there is still a lot of RF energy in the DCX case and I was not able to shield the transformers well enough to keep all that out.
* Same comments as for previous apply for passive RLC filtering.
* Simple one stage opamp design is the one I have been using in the end. With shunt regulators on each opamp and simple RC before and after opamp it seems that the plusses are all there and I do not feel the "lack of punch" I mentioned before.
*****
So where I ended up was to have a RC filter, then a 6 channel ALPS pot and then a single opamp stage. All this inside the DCX case.
For current versions I have moved the 6ch APLS pot + opamp stage to separate box as I sometimes want to emulate the crossover with PC and LspCad and soundcard and there was the same problem with the lack of analog volume control....
So I can imagine that Jan's solution for analog stage is a very good one as it is very similar topology wise where I ended up.
*****
As for clock - I have only tried the Tent XO - so unfortunately can not comment on Tent versus Selectronics issue.
I'm lucky in one more point - I have unmodified DCX2496 at work - so I can do comparisons between where I'm at with my mods compared to original one. Plus I have modified one DCX for my friend (where we modded the analog in and out but not the clock) so I have even a third midway modded reference.
Having done this since spring 2004 I'm pretty sure I'm not imagining the positive progress compared to stock unit - having listened to the unit in different homes and moods etc. But each has to travel their own way as not all mods are quaranteed sucess. Some trial and error is always good and educates a lot
Ergo
Well my feelings after all trials are following
* Passive mod is an upgrade compared to original output stage. The datasheet suggests it is able to drive even the passive preamp after it. In reality the one to one comparison to active stages gives afeeling in the end that there is a bottleneck there. The sound is clean and neutral but seems to lack some punch and energy that make music sound alive.
* Transformers (I tried Lundahl) are very good sounding, but as I'm also a kind of nerd and also look at objective measurements I never got the spectrum to look "clean" with transformers. Being a digital / analog mixed design device there is still a lot of RF energy in the DCX case and I was not able to shield the transformers well enough to keep all that out.
* Same comments as for previous apply for passive RLC filtering.
* Simple one stage opamp design is the one I have been using in the end. With shunt regulators on each opamp and simple RC before and after opamp it seems that the plusses are all there and I do not feel the "lack of punch" I mentioned before.
*****
So where I ended up was to have a RC filter, then a 6 channel ALPS pot and then a single opamp stage. All this inside the DCX case.
For current versions I have moved the 6ch APLS pot + opamp stage to separate box as I sometimes want to emulate the crossover with PC and LspCad and soundcard and there was the same problem with the lack of analog volume control....
So I can imagine that Jan's solution for analog stage is a very good one as it is very similar topology wise where I ended up.
*****
As for clock - I have only tried the Tent XO - so unfortunately can not comment on Tent versus Selectronics issue.
I'm lucky in one more point - I have unmodified DCX2496 at work - so I can do comparisons between where I'm at with my mods compared to original one. Plus I have modified one DCX for my friend (where we modded the analog in and out but not the clock) so I have even a third midway modded reference.
Having done this since spring 2004 I'm pretty sure I'm not imagining the positive progress compared to stock unit - having listened to the unit in different homes and moods etc. But each has to travel their own way as not all mods are quaranteed sucess. Some trial and error is always good and educates a lot
Ergo
oettle said:
Thank you Jan and Oettle.
No I am not worried at all. It is just that I didn't do any clock improvements, and since I was getting ready I was trying to find out which one will be the best one to do.
I am very excited about Oettle's mod since I have ordered one. I have Behringer SRC and DEQ and DCX. I do not think it is good to stack more units one after each other since every unit in the chain and S/PDIF cable will bring more jitter.
To answer Oettle's question - my CD transport is pretty good one, it is Sony XA20ES. Very solid mechanism and tray with CD clamp, but unfortunately in the Sony style they have provided only optical out. Because of that I have to go through SRC. Regarding that I am researching how to mod my CD as well. I am thinking of inserting transmitter CS 8406 and appropriate digital transformer, so I could have AES/EBU output. That could go straight to my DCX with short cable which certainly will be much better than having several units. I just ordered my Sony schematic in order to see how to access the signal. Any advice here is welcomed.
Another mod that I will be doing is as I mentioned earlier, exchange of AK4393 DACs to AK4396s. I have two DCX units and on one one DAC died. So I will exchange that one first and see if there is any improvement. There is a lot written about these chips, but I do not expect any big improvement. We will see. I will document in pictures my surgery. I just got my hot air soldering station, so I feel no IC soldering is a problem:
(Wait until I start screaming for the help)What is interesting CS8416 in Oetl'e mod is 192 capable, as well as AK4396. I do not know if DSP is 192 capable or not, or if that matters. Question is: Is it possible to put throgh 24/192 signal after all this modifications. I will see, since my sound cards are capable of producing 192 digital signal, as well as there are modifications that are done to the SACD that allow that kind of digital signal to be sent out.
Anyway, I have a lot planned to do. I will be doing Power supply mod as well after I find the right transformer for 110v.
AR2
AR2,
The modding of the XA20ES should not be difficult at all - I have XB930 which is in many ways similar unit and I assume in you player the SPDIF signal comes out from the transport PCB also with a separate 3 wire connector that then goes to DA and analog PCB somewhere near the Optical out socket.
Basicly all you have to do is to use this signal and make a circuit after it to feed the 75 ohm coaxial digital cable. There are many circuits in this forum for that purpose just run a search and pick what seems best to you.
Ergo
The modding of the XA20ES should not be difficult at all - I have XB930 which is in many ways similar unit and I assume in you player the SPDIF signal comes out from the transport PCB also with a separate 3 wire connector that then goes to DA and analog PCB somewhere near the Optical out socket.
Basicly all you have to do is to use this signal and make a circuit after it to feed the 75 ohm coaxial digital cable. There are many circuits in this forum for that purpose just run a search and pick what seems best to you.
Ergo