Chameleon DAC listening and modifications
May 10, 2010 at 10:33 AM Post #931 of 1,158
Quote:
All SPDIF receivers need clock to recover the I2S signal either from external clock source or in-chip PLL-VCO clock source.   For the V2.5 DAC (Valab), the Dir9001 used the in-chip PLL-VCO to decode a biphase input signal with a sampling frequency from 28 kHz to 108 kHz.  For the wm8805, it uses the external clock and its internal PLL to decode the biphase input.   The quality of clock source and the internal PLL loop therefore determine the recovered digital audio quality.  ^-^
 
The pins 13 and 14  of Dir9001 are to select the clock source for system clock.  For pins 13 and 14 to 3.3v mod, it is to set the system clock to 512fs.  It may means the internal PLLed 512fs is better than the other frequency.


Teraguy, thank you for your recent descriptions and clarifications, especially since jkeny has been looking at Chameleon.  You seem to know the fine details very well.  It's made me very curious, if you're willing to say.  Who are you, where are you, and what's your experience concerning Teradak, Chameleon, and the Valab DAC?
 
May 10, 2010 at 10:55 AM Post #932 of 1,158


Teraguy : guys from teradak.    We only talk technique and don't talk business in this forum.  That will make us live longer in this forum.  ^-^
 
BTW, we don't know who will buy or are interested in this V4.5 (Chameleon) DAC.  We read the thread sometimes and shared our design experiences if necessary.
 
 
Quote:
Quote:

Teraguy, thank you for your recent descriptions and clarifications, especially since jkeny has been looking at Chameleon.  You seem to know the fine details very well.  It's made me very curious, if you're willing to say.  Who are you, where are you, and what's your experience concerning Teradak, Chameleon, and the Valab DAC?



 
May 10, 2010 at 11:33 AM Post #933 of 1,158
Quote:
Teraguy : guys from teradak.    We only talk technique and don't talk business in this forum.  That will make us live longer in this forum.  ^-^
 
BTW, we don't know who will buy or are interested in this V4.5 (Chameleon) DAC.  We read the thread sometimes and shared our design experiences if necessary.


Completely understand.  Thanks!
 
May 10, 2010 at 12:12 PM Post #934 of 1,158


Quote:
 

Hi Wood,
 
did you notice any problem with your ultravox at first? mine didn't work, wired it up with 12 volt, the lights lit up but no data signal to the DAC. My window 7 didn't recognise the DAC? weird.
 

Someone on Diyparadise had an initial problem but solved it,
 
http://diyparadise.com/forum/index.php?topic=1133.0
 
Its hard to guess what could be wrong, the best place to try to find a solution would be writting to Yeo directly, or bringing it up on the thread, They pretty sharp and very supportive of users of their products.
 

 
 
May 10, 2010 at 7:41 PM Post #935 of 1,158
Hi Rhodes,
I see that you have the Tentlabs setup on the 12MHz clock in your photo album. How is that working out? I would not expect quite as much difference as on the upsampling card but maybe I am wrong. The reason I hope that I am is that I will have to spend about $300 to do the mod.
Thanks in advance for your time. Tony
 
May 10, 2010 at 11:46 PM Post #936 of 1,158
 
Maybe someone is curious why the Chameleon doesn't use Dir9001 for SPDIF receiver, but uses the Wm8805 instead.  The reason is the system jitter is controlled by the internal VCO-PLL clock of Dir9001.  We can not use better clock to improve the digital recovery quality.   The Dir9001 however is proper for general consumer product because of  its good quality and easy usage.  To use the Dir9001 for the Teradak is also just a piece of cake based on the previous V2.5 experience. 
As for the Chameleon DAC, it is a continuous evolutive DAC.  It is necessary to improve its digital input and analog output continuously.  Therefore the Teradak uses the WM880x instead.  The WM880x allows us to upgrade its external crystal to increase the SPDIF decoded quality. To improve the decode quality in the Dir9001 system is also possible, but it shoud take a lot of efforts to FIFO, reclock or ASRC the I2S data.   The reclock used in the V2.5/V3.0 version DAC is to improve the quality.
 
 
The clock and the SPDIF receiver internal PLL determine the SPDIF decode quality.  In current Chameleon Wm8805 setting, the user can focus on the external clock itself.   Better clock should improve the decode quality.  Some day the WM880x can be also replaced because the Teradak find the other better solution.  ^-^
 
Quote:
Quote:

Completely understand.  Thanks!



 
May 11, 2010 at 12:08 AM Post #937 of 1,158
Quote:
Maybe someone is curious why the Chameleon doesn't use Dir9001 for SPDIF receiver, but uses the Wm8805 instead.  The reason is the system jitter is controlled by the internal VCO-PLL clock of Dir9001.  We can not use better clock to improve the digital recovery quality.   The Dir9001 however is proper for general consumer product because of  its good quality and easy usage.  To use the Dir9001 for the Teradak is also just a piece of cake based on the previous V2.5 experience. 
As for the Chameleon DAC, it is a continuous evolutive DAC.  It is necessary to improve its digital input and analog output continuously.  Therefore the Teradak uses the WM880x instead.  The WM880x allows us to upgrade its external crystal to increase the SPDIF decoded quality. To improve the decode quality in the Dir9001 system is also possible, but it shoud take a lot of efforts to FIFO, reclock or ASRC the I2S data.   The reclock used in the V2.5/V3.0 version DAC is to improve the quality.
 
The clock and the SPDIF receiver internal PLL determine the SPDIF decode quality.  In current Chameleon Wm8805 setting, the user can focus on the external clock itself.   Better clock should improve the decode quality.  Some day the WM880x can be also replaced because the Teradak find the other better solution.  ^-^


It is very reassuring and appreciated to hear you say things like this.
 
Since the Wolfson chip will benefit from better clocks, how do we approach it?  Do you have suggestions for replacements?
 
You can't talk about this for reasons we know, but as you briefly mentioned there needs to be something done about raising interest in Chameleon.  When Chameleon is more popular, more users will be interested in modding and getting optional daughter cards.  Making more people aware of Chameleon needs to be a priority.
 
May 11, 2010 at 10:49 AM Post #938 of 1,158
Quote:
Hi Rhodes,
I see that you have the Tentlabs setup on the 12MHz clock in your photo album. How is that working out? I would not expect quite as much difference as on the upsampling card but maybe I am wrong. The reason I hope that I am is that I will have to spend about $300 to do the mod.
Thanks in advance for your time. Tony


Hi Tony,
 
I've indeed implemented the second Tentlabs clock recently. Unfortunately I had only a few hours so far to listen to it as I had a bad XO at first and when I received my new one I was about to leave for a holiday to France. First impressions are: A big step up when used without upsampling board. Much more transparency, tighter bass, less grain: as expected from a better clock (and at least so important: a seperate and clean power supply). I also experienced it as more natural, musical and better timing.
Haven't listened yet with the upsampling board but I suspect less noticeble differences.
 
May 11, 2010 at 2:23 PM Post #939 of 1,158
hi rhodes54, I particulary like the fact you have found the mod made the Chameleon more musical, did you get a chance to compare using the tenlabs PSU with the existing clock? I am considering using a 3.3 LiFePo4 cell to power the clock.
 
May 12, 2010 at 5:37 AM Post #941 of 1,158

 
Quote:
Someone on Diyparadise had an initial problem but solved it,
 
http://diyparadise.com/forum/index.php?topic=1133.0
 
Its hard to guess what could be wrong, the best place to try to find a solution would be writting to Yeo directly, or bringing it up on the thread, They pretty sharp and very supportive of users of their products.
 
Thanks Wood,
 
sorted, it was a dodgy soldering job :)
 



 
May 13, 2010 at 2:16 PM Post #942 of 1,158


Quote:
hi jkeny,
I am not clear....are you saying that you have now been able to get the ADUM to work with the HiFace?
I am wondering about the compatibility of the Ultravox with the HiFace.
Thanks.


No, I haven't had the time to track down the problem - not sure if it's a Windows 7 problem or the ADUM board itself?
 
May 13, 2010 at 4:00 PM Post #943 of 1,158
Hi, taiphan
glad to see your up and running. Be great to know how it works with your setup.
 
May 14, 2010 at 1:01 AM Post #944 of 1,158
Hey, Wood.  Ultravox works on an Off Ramp without a hitch...
 ​

 
Characteristics should be of reduced jitter.  Transparency improved and the upper mids seem to have tightened up a bit.  For example, cymbal sizzle is clearer.  I want to say it's made an improvement but I need to spend some days listening to be sure it's not placebo.  The DAC19DSP is still burning in.  I probably should have waited another week on the Ultravox, but you know how that goes.  It was too easy to put it there.
 
May 14, 2010 at 4:20 AM Post #945 of 1,158
Similair to my findings, its subtle. More of a gentle dusting then a rigourous polish.
 

Users who are viewing this thread

Back
Top