luvdunhill said:
Hi, if you are referring to Thomas's v2 Dac, the interior measurements of my single box are 12" deep and 16.25" wide. I would not go under 12" on the depth...there's a little play on the width. My layout is...digital and analog side by side, with the three transformers in front of the boards in a row. If you are referring to the v3 dac you might try the other thread.
Hi,
I'm using the async reclocker circuit, and was wondering about the clock that Thomas supplied, which apparently is a <2ppm model which should be pretty good, and wether it would benefit from swapping it out with a known good clock, like the Tent labs clock.
Has anyone investigated the clock on this DAC?
I'm using the async reclocker circuit, and was wondering about the clock that Thomas supplied, which apparently is a <2ppm model which should be pretty good, and wether it would benefit from swapping it out with a known good clock, like the Tent labs clock.
Has anyone investigated the clock on this DAC?
My experience is that the XO supplied by Thomas is decent. I have tried another TCXO and one XO which were not as good.
The asynchronous reclock board based on 74hc74 does improve the sound quite a bit.
This is based on listening experience only.
However the experts here say that the ppm spec has nothing to do with quality of the XO.
There are also some discussions that asynchronous reclocker degrades the sound. To be honest I do not understand why. But listening experience is that it sounds good.
The asynchronous reclock board based on 74hc74 does improve the sound quite a bit.
This is based on listening experience only.
However the experts here say that the ppm spec has nothing to do with quality of the XO.
There are also some discussions that asynchronous reclocker degrades the sound. To be honest I do not understand why. But listening experience is that it sounds good.
Hi Qantran,
Thanks for the feedback. I agree, the clock and the async reclocker does sound good, so I was wondering what I could do to improve on it.
Interestingly when I send 48Khz signals to the DAC (for DVD or even resampled CD Audio) the sound quality with the reclocker which is set at 44.1Khz is still very good, arguably better than standard 44.1Khz. Not exactly sure why that is, even though resampling to 48Khz helps to fix the sinx/x problem with NOS for the higher frequences, I would have thought that the increased dropped samples by the mismatch in sample rates would make it sound bad???
The two things I'm looking at doing is:
- Send the clock back to the transport (PC pro-audio Soundcard, EMU 1820m) to make the clock synchronous between the transport & DAC. But I think as the async reclock seems to be working, I won't get much gain moving to sync.
- Try a different I/V and buffer. I'm just using the standard buffer board supplied with the kit, and it does sound decent, but I'm sure it can be upgraded (eg. remove the output coupling cap if there is minimal DC on the TDA output).
Any other ideas on tweaking this design would be appreciated!
I am still using
Thanks for the feedback. I agree, the clock and the async reclocker does sound good, so I was wondering what I could do to improve on it.
Interestingly when I send 48Khz signals to the DAC (for DVD or even resampled CD Audio) the sound quality with the reclocker which is set at 44.1Khz is still very good, arguably better than standard 44.1Khz. Not exactly sure why that is, even though resampling to 48Khz helps to fix the sinx/x problem with NOS for the higher frequences, I would have thought that the increased dropped samples by the mismatch in sample rates would make it sound bad???
The two things I'm looking at doing is:
- Send the clock back to the transport (PC pro-audio Soundcard, EMU 1820m) to make the clock synchronous between the transport & DAC. But I think as the async reclock seems to be working, I won't get much gain moving to sync.
- Try a different I/V and buffer. I'm just using the standard buffer board supplied with the kit, and it does sound decent, but I'm sure it can be upgraded (eg. remove the output coupling cap if there is minimal DC on the TDA output).
Any other ideas on tweaking this design would be appreciated!
I am still using
As you mentioned about removing output cap I guess you are using the opamp module based on opa2604.
I think the output caps are not a big issue. Blackgate cap can be used here. They are very transparent. However the opamp board design of this kit is not really up to the level of the DAC.
TDA1541a always produce DC offset at its output. One way to handle this is adjusting the supplyn voltage for the i/v opamp to compensate. I saw this done in some revox player. But again I think the output cap is not a big issue.
You may try either a discrete SS i/v stage like pass D1 or a passive i/v with tube buffer.
DEM recloking may be worth trying also. I have heard good comments about it.
I think the output caps are not a big issue. Blackgate cap can be used here. They are very transparent. However the opamp board design of this kit is not really up to the level of the DAC.
TDA1541a always produce DC offset at its output. One way to handle this is adjusting the supplyn voltage for the i/v opamp to compensate. I saw this done in some revox player. But again I think the output cap is not a big issue.
You may try either a discrete SS i/v stage like pass D1 or a passive i/v with tube buffer.
DEM recloking may be worth trying also. I have heard good comments about it.
deandob said:
Hi Quantran,
I'm sure the output board uses a blackgate at the output as standard. The opamp buffer also has a pretty decent power supply setup separate to the DAC, but I'm not sure about the filtering arrangements, I'd like to try something a little less filtered (although I use a Hypex Class D amp so the high frequency images on the DAC output may not be good for a Class D amp).
Do you mean adding a DC bias to one of the inputs of the opamp to compensate for the TDA DC offset?
I'll look into the pass D1 output stage & DEM reclocking, good tips, thanks.
I'm sure the output board uses a blackgate at the output as standard. The opamp buffer also has a pretty decent power supply setup separate to the DAC, but I'm not sure about the filtering arrangements, I'd like to try something a little less filtered (although I use a Hypex Class D amp so the high frequency images on the DAC output may not be good for a Class D amp).
Do you mean adding a DC bias to one of the inputs of the opamp to compensate for the TDA DC offset?
I'll look into the pass D1 output stage & DEM reclocking, good tips, thanks.
Hi Guys
Quan happy new year to you!
I'm still tinkering with Thomas's V2 and enjoying very much.
Thinking as well about building an SS IV stage as a complement to current tube stage. I'll toss out a few suggestions for deandob with links, one of which has been mentioned:
Pass D1
http://www.diyaudio.com/forums/showthread.php?s=&threadid=81739&highlight=
Rbroer Less Simple IV for TDA1541A
http://www.diyaudio.com/forums/showthread.php?s=&threadid=22542&highlight=
Swenson version of D1 using Pimm Self Bias CCS boards
for link look under John's name on AA Tube DIY board (AA is down right now)
And one tube stage:
Brian Beck and Eli Duttman brainstormed about how a Phono Stage functions similar to a gain stage for TDA1541 since like phono, the tda1541 is optimized with very low voltage at IV but then needs a quiet way to amplify to 2V. Circuit uses 6922. Posted somewhere on diyaudio.
I synchronously reclock using Tent XO powered by Tent Shunt Reg with nice results.
Quan happy new year to you!
I'm still tinkering with Thomas's V2 and enjoying very much.
Thinking as well about building an SS IV stage as a complement to current tube stage. I'll toss out a few suggestions for deandob with links, one of which has been mentioned:
Pass D1
http://www.diyaudio.com/forums/showthread.php?s=&threadid=81739&highlight=
Rbroer Less Simple IV for TDA1541A
http://www.diyaudio.com/forums/showthread.php?s=&threadid=22542&highlight=
Swenson version of D1 using Pimm Self Bias CCS boards
for link look under John's name on AA Tube DIY board (AA is down right now)
And one tube stage:
Brian Beck and Eli Duttman brainstormed about how a Phono Stage functions similar to a gain stage for TDA1541 since like phono, the tda1541 is optimized with very low voltage at IV but then needs a quiet way to amplify to 2V. Circuit uses 6922. Posted somewhere on diyaudio.
I synchronously reclock using Tent XO powered by Tent Shunt Reg with nice results.
Yes, this is a pretty good DAC, I'm quite pleased with the results but want to see how I can take it to the next level.
Riotubes, did you find a difference between async reclocking and sync reclocking? I have looked at taking the DAC clock back to my pro-audio soundcard source but am not sure if this will make a difference. I am losing samples with the async clock but in general the async reclocker improved the sound. I find it interesting that when I run the source at 48Khz the sound actually improves, and I must be dropping a whole lot more samples due to the DAC & source clock differences!
I'm not sure about tubes, I feel they "mellow" the sound too much. I have not built the tube board that comes with the DAC, how does it compare with the opamp board?
Riotubes, did you find a difference between async reclocking and sync reclocking? I have looked at taking the DAC clock back to my pro-audio soundcard source but am not sure if this will make a difference. I am losing samples with the async clock but in general the async reclocker improved the sound. I find it interesting that when I run the source at 48Khz the sound actually improves, and I must be dropping a whole lot more samples due to the DAC & source clock differences!
I'm not sure about tubes, I feel they "mellow" the sound too much. I have not built the tube board that comes with the DAC, how does it compare with the opamp board?
Hi Riotubes
I also still enjoy the TDA1541A V2 DAC.
I have also finished my PCM63 NOS DAC with D1 i/v stage. It also sound very good but has a very different sonic signature.
My transport has a different clock frequency (16.9344 and not 11.2896 like the DAC) so no chance to synchronize. I installed another XO in the transport.
Quan
I also still enjoy the TDA1541A V2 DAC.
I have also finished my PCM63 NOS DAC with D1 i/v stage. It also sound very good but has a very different sonic signature.
My transport has a different clock frequency (16.9344 and not 11.2896 like the DAC) so no chance to synchronize. I installed another XO in the transport.
Quan
riotubes said:
deandob
Having used both the tube buffer and the D1 i/v stage, I can only say that I prefer the sound of the tube buffer probably for the type of music I usually listen to (chamber music).
I have not tried the D1 i/v stage with the tda1541 dac. I only connect it to the pcm63 so far. I think the D1 i/v stage achieves lower distortion, more details but I think the tube buffer is more musical.
When I first assemble the tube buffer it did not sound as good as the opamp module but after some small tweaks I never go back to the opamp again.
Regarding the asynchronous reclocking if you are using the one with 74hc74 you are not losing any samples. I have study the design, Fsync is recovered from the source signal. it is then reclocked using the XO frequency. After reclocking Fsync is only delayed by 1/4 bck. If you feed the DAC with 48khz then Fsync output into tda1541a is still 48khz
Cheers
Quan
Having used both the tube buffer and the D1 i/v stage, I can only say that I prefer the sound of the tube buffer probably for the type of music I usually listen to (chamber music).
I have not tried the D1 i/v stage with the tda1541 dac. I only connect it to the pcm63 so far. I think the D1 i/v stage achieves lower distortion, more details but I think the tube buffer is more musical.
When I first assemble the tube buffer it did not sound as good as the opamp module but after some small tweaks I never go back to the opamp again.
Regarding the asynchronous reclocking if you are using the one with 74hc74 you are not losing any samples. I have study the design, Fsync is recovered from the source signal. it is then reclocked using the XO frequency. After reclocking Fsync is only delayed by 1/4 bck. If you feed the DAC with 48khz then Fsync output into tda1541a is still 48khz
Cheers
Quan
deandob said:
No, the revox B226 use +12V for positive supply rail and -15V for the negative rail. I think the designer was trying to compensate for DC offset this way.
You can take a look at the schematic on this site http://www.revox.name/
You can take a look at the schematic on this site http://www.revox.name/
deandob said:
Quan,
This is interesting, I'll need to look at the design again. My understanding with an async reclocker is that you ignore the incoming clock but buffer the input signal to the new clock, which would not completely match the source clock, hence periodically dropping samples. However you are probably right, as 48Khz source vs a async 44.1Khz clock should sound bad and it does not. When I setup my source for 88.2Khz I get bad distortion however (lack of sync, noise bursts).
I never did build the async board, but there is feedback by many that it does improve the sound. With the synchronous reclocking setup, the improvement was dramatic. My comments are in post #736 above. Of course you could do this with a traditional transport that has a 11.2896mHz clock, but compared to a transport costing 2-3X as much, the squeezebox is most likely better with the added bonus of freeing up your music collection imo. I encourage anyone with the tda1541 dac to try this.
In I2S format each data frame (1 fsync cycle) has 64 bck (32 bits for each chanel). the fist bit of each dataframe is 1 bck cycle behind Fsync.
CD stereo signal only occupied 16 bit. Therefore there are 15 empty spaces after each 16 bit of data.
In this design bck has been fixed at 2.8224mhz
When you increase to 48khz the impact is that some empty bits will disapear. Now there are only 58 bck per fsync intead of 64 in the case of 44.1khz signal.
When you increase to 88.2khz one fsync cycle is equal only 32 bck and remember the 1 bck delay in I2S format. The result is not enough space for 16 data bits. This leads to messing up data format --> distortion
I think the DAC will probably work with signal up to about 80khz.
Cheers
Quan
CD stereo signal only occupied 16 bit. Therefore there are 15 empty spaces after each 16 bit of data.
In this design bck has been fixed at 2.8224mhz
When you increase to 48khz the impact is that some empty bits will disapear. Now there are only 58 bck per fsync intead of 64 in the case of 44.1khz signal.
When you increase to 88.2khz one fsync cycle is equal only 32 bck and remember the 1 bck delay in I2S format. The result is not enough space for 16 data bits. This leads to messing up data format --> distortion
I think the DAC will probably work with signal up to about 80khz.
Cheers
Quan
deandob said:
Quan,
Thanks for the info on how this works.
Just so I understand it, because of the serial format of i2s there is no way that I can double the sample rate with this reclocker?
I would have thought the problem here with different sample rates when asynchronously reclocking is a timing problem, not a bit space/length problem? With the DAC expecting 44.1Khz input and the source clocked at 48Khz, then surely we are doing more than using up a couple of spare bits for each word (which would happen in a synchronous system) but instead we would be losing many samples because of the clock difference between 44.1Khz and 48Khz as the reclocker is async?
I'm sure you are right because 48Khz works well and 88.2Khz does not. Looks like if I want to use 88.2Khz I'd have to think of an alternative scheme (eg. sync reclocking). This DAC does sound better if the source is resampled at a higher rate (with a PC this is easy to do with high quality).
Thanks for the info on how this works.
Just so I understand it, because of the serial format of i2s there is no way that I can double the sample rate with this reclocker?
I would have thought the problem here with different sample rates when asynchronously reclocking is a timing problem, not a bit space/length problem? With the DAC expecting 44.1Khz input and the source clocked at 48Khz, then surely we are doing more than using up a couple of spare bits for each word (which would happen in a synchronous system) but instead we would be losing many samples because of the clock difference between 44.1Khz and 48Khz as the reclocker is async?
I'm sure you are right because 48Khz works well and 88.2Khz does not. Looks like if I want to use 88.2Khz I'd have to think of an alternative scheme (eg. sync reclocking). This DAC does sound better if the source is resampled at a higher rate (with a PC this is easy to do with high quality).
I think you can make it work with 88.2 and even 96khz by doubling the bit clock frequency.
You can see that U6 is only a frequency divider. Currently bck and inverted bck are taken from pin 8 and 9 of U6 (2.8224 mhz). As 8414 is is mode 11, bck is inverted for compatibility with tda1541a.
Now if you want to use 88.1 khz source you can use pin 5 and 6 of U6 which output 5.6448mhz
Modification steps are as follows:
1. cut pcb trace from U6's pin 8 to J2's pin 3 (this is original 2.8224mhz inverted bck feeding tda1541a)
2. cut pcb trace from U6's pin 9 to U7's pin 3 and J2's pin 4 (this is original 2.8224mhz bck feeding cs8414 and the reclocker (U7))
3. wire U6's pin 5 to U7's pin 3 and J2's pin 4 (this will provide bck to cs8414 and the reclocker chip at 5.6448mhz)
4. wire U6's pin 6 to J2's pin 3 (this will provide inverted bck to tda1541a at 5.6448mhz)
I have not done this mod before so cannot guarantee it. Just an idea for your consideration.
You can see that U6 is only a frequency divider. Currently bck and inverted bck are taken from pin 8 and 9 of U6 (2.8224 mhz). As 8414 is is mode 11, bck is inverted for compatibility with tda1541a.
Now if you want to use 88.1 khz source you can use pin 5 and 6 of U6 which output 5.6448mhz
Modification steps are as follows:
1. cut pcb trace from U6's pin 8 to J2's pin 3 (this is original 2.8224mhz inverted bck feeding tda1541a)
2. cut pcb trace from U6's pin 9 to U7's pin 3 and J2's pin 4 (this is original 2.8224mhz bck feeding cs8414 and the reclocker (U7))
3. wire U6's pin 5 to U7's pin 3 and J2's pin 4 (this will provide bck to cs8414 and the reclocker chip at 5.6448mhz)
4. wire U6's pin 6 to J2's pin 3 (this will provide inverted bck to tda1541a at 5.6448mhz)
I have not done this mod before so cannot guarantee it. Just an idea for your consideration.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.