There is a general app note on the Tent website, but from reading earlier in the thread it looks like there was a specific one for the DDDAC boards
Guido provides a "+8.0 Doede" version of the shunt reg based on the 15V reg, these are the ones I have
That explains why one of the red boards has a small board hacked onto it with two more of the 74VHC04MX on board
Thanks for the advice, I will get one of the blue main boards
Good things take a while
Finally I had some time to do the almost final tests with the Fifopi and the accusilicon clocks.
Remember, in earlier tests I reported back, that the PI3 with the fifopi AND the accu clocks made a clear difference for the positive. Very nice analog natural sound. Compared to this the WaveIO connected directly to the DDDAC sounded less spatial, less natural. Still great, but hey, easy to hear and you do not want to go back from the fifopi...
I could run WaveIO from the Aurender as well from the USB output on the raspberry.... compared between the two, the Aurender clearly is the better streamer. But it could not match with the raspberry with the fifopi and accu clocks...
So what I did today, was make a new Frankenstein set up (see images ) where the fifopi was disconnected from the PI and connected directly to the DDDAC I2S input. Power supply as earlier. 3,3 Volt LDO to the clean clock section. Raw and before isolation side fed from a nothing special 5volt supply which also feeds the raspberry side of things
The INPUT of the fifopi is connected to the output of a selector board (see pics ... with the 4 blue relays) - the two inputs from the selectorboard are the raw I2S from the raspberry pi and from the WaveIO . The selector board is controlled by the Arduino in my DDDAC, so both can run in parallel from the same track and with the remote control I can switch back and forward between the two sources and do a 100% A to B comparison. No issues with getting up, selecting another source and start the track again from start and try to remember before and after...
Well this was VERY necessary, as the result was extremely simple: I could NOT hear any difference between sources, regardless of the WaveIO was fed by the Aurender or the PI or if I2S was coming from the dirty PI I2S output.... I kept switching back and forward and simply could not notice ANY difference. It might be my old ears are not so good anymore, but hey, the other difference in the previous setup was easy to hear. So IF there is a difference, that is up for people with better ears than mine, which is absolutely possible of course
But for now, my only conclusion can be, whatever you throw at the FIFOPI, it comes out at the same high quality, IF you use the super clocks....
This makes things VERY easy... everyone who wants a nice upgrade for his DDDAC, just put a Fifopi with the Accusilicon clocks at the DDDAC input and feed it with whatever I2S you have at hand....
So my next (probably long) step is to rebuild my DDDAC to accommodate the boards a little bit better than what you see at the pictures ( understatement). Meaning new front panel and backplane... and I will build the FIFOPI very close to the input and connect ithe fifopi with U/FL connectors to the DDDAC input. Dick checked this and was positive this put some dots on the i.... no reason to ignore a simple “tweak” like that, right ? And I VERY much trust his ears, which I know are better than mine
From there it maybe makes sense to check if any other power supply for the clocks makes a difference... but as said, first bring my DDDAC to the next construction phase....
I will keep the WaveIO input, to keep the DAC as compatible as possible to other sources, but that is just me of course trying to keep my DAC ready for all kind of testing etc....
Finally I had some time to do the almost final tests with the Fifopi and the accusilicon clocks.
Remember, in earlier tests I reported back, that the PI3 with the fifopi AND the accu clocks made a clear difference for the positive. Very nice analog natural sound. Compared to this the WaveIO connected directly to the DDDAC sounded less spatial, less natural. Still great, but hey, easy to hear and you do not want to go back from the fifopi...
I could run WaveIO from the Aurender as well from the USB output on the raspberry.... compared between the two, the Aurender clearly is the better streamer. But it could not match with the raspberry with the fifopi and accu clocks...
So what I did today, was make a new Frankenstein set up (see images
) where the fifopi was disconnected from the PI and connected directly to the DDDAC I2S input. Power supply as earlier. 3,3 Volt LDO to the clean clock section. Raw and before isolation side fed from a nothing special 5volt supply which also feeds the raspberry side of thingsThe INPUT of the fifopi is connected to the output of a selector board (see pics ... with the 4 blue relays) - the two inputs from the selectorboard are the raw I2S from the raspberry pi and from the WaveIO . The selector board is controlled by the Arduino in my DDDAC, so both can run in parallel from the same track and with the remote control I can switch back and forward between the two sources and do a 100% A to B comparison. No issues with getting up, selecting another source and start the track again from start and try to remember before and after...
Well this was VERY necessary, as the result was extremely simple: I could NOT hear any difference between sources, regardless of the WaveIO was fed by the Aurender or the PI or if I2S was coming from the dirty PI I2S output.... I kept switching back and forward and simply could not notice ANY difference. It might be my old ears are not so good anymore, but hey, the other difference in the previous setup was easy to hear. So IF there is a difference, that is up for people with better ears than mine, which is absolutely possible of course
But for now, my only conclusion can be, whatever you throw at the FIFOPI, it comes out at the same high quality, IF you use the super clocks....
This makes things VERY easy... everyone who wants a nice upgrade for his DDDAC, just put a Fifopi with the Accusilicon clocks at the DDDAC input and feed it with whatever I2S you have at hand....
So my next (probably long) step is to rebuild my DDDAC to accommodate the boards a little bit better than what you see at the pictures (
understatement). Meaning new front panel and backplane... and I will build the FIFOPI very close to the input and connect ithe fifopi with U/FL connectors to the DDDAC input. Dick checked this and was positive this put some dots on the i.... no reason to ignore a simple “tweak” like that, right ? And I VERY much trust his ears, which I know are better than mine From there it maybe makes sense to check if any other power supply for the clocks makes a difference... but as said, first bring my DDDAC to the next construction phase....
I will keep the WaveIO input, to keep the DAC as compatible as possible to other sources, but that is just me of course trying to keep my DAC ready for all kind of testing etc....
Attachments
Last edited:
Good things take a while
Finally I had some time to do the almost final tests with the Fifopi and the accusilicon clocks.
Hallo Doede,
Nice to see you are discovering the fifopi. This is indeed a very nice addition to your dddac.
I have done a test with different clocks a while ago, see post #1975 : https://www.diyaudio.com/forums/dig...itter-crystal-oscillator-198.html#post6067562
Using the fifopi with a clean battery power of 3.3v will improve the sound even more; with 3.3v you can bypass the on board ldo’s, all ldo’s make some degree of noise and this is audible through the fifopi.
I have changed the rpi for an Allo USBridge Sig. This is also an improvement.
Nice to read that the Aurender is on par with the rpi in this configuration. That means my USBridge sig solution is even better than the Aurender
Attachments
Last edited:
Great work Doede!
I love my LUMIN U1 and waveio and was worried a cheap PI would be better
Why wouldn’t it?
The cheap raspberry pi is in general comparable with the processor technology in all these expensive streamers. The thing good streamers add is a good quality power supply and clock system (and a nice box). This can also be done with a raspberry pi. A very good example is the USBridge where the raspberry pi compute module is used in a mother board with extensive power supply and a good usb implementation. Another advantige for a separate board is that you can combine it with the dddac with very short i2s cables, this is always a problem with separate units. The shorter the i2s cables the lower the jitter will be.
My opinion is that separate streamer-dac combinations are always one step behind of intergrated solutions due to the needed cables and connectors.
Excellent comments I agree and I have been experimenting with the Pi myself and it is VERY impressive.
I use Tidal a lot which I have not seen a way to use with a Pi, any ideas?
Commercial streamers help in this regard but happy for suggestions to use Tidal as well as a Pi
Excellent comments I agree and I have been experimenting with the Pi myself and it is VERY impressive.
I use Tidal a lot which I have not seen a way to use with a Pi, any ideas?
Commercial streamers help in this regard but happy for suggestions to use Tidal as well as a Pi
When using Roon, RopieeXL on the Raspberry, it is easy to use Tidal. It is then integrated in Roon. I use it that way, with my Raspberry Pi, FiFoPi DDDAC streamer.
Excellent comments I agree and I have been experimenting with the Pi myself and it is VERY impressive.
I use Tidal a lot which I have not seen a way to use with a Pi, any ideas?
Commercial streamers help in this regard but happy for suggestions to use Tidal as well as a Pi
Moode audio can stream Tidal through UPnP: moodeaudio.org
So Doede conclusion regarding the source was...
....so after reading other comments on here it seems source gains can happen when....
1. You chose an optimized solution like the USBridge sig solution
2. keep the I2S very short
3. Use the best reclocker and clocks like fifopi and the accu clocks
4. Use great power supplies like the battery supply
....if you can do all that then the source can sound better than a commercial streamer.
Is my summary correct, anything to add?
....so after reading other comments on here it seems source gains can happen when....
1. You chose an optimized solution like the USBridge sig solution
2. keep the I2S very short
3. Use the best reclocker and clocks like fifopi and the accu clocks
4. Use great power supplies like the battery supply
....if you can do all that then the source can sound better than a commercial streamer.
Is my summary correct, anything to add?
I know the story was pretty long, so let’s try to summarize what my real conclusions are
Reclocking with good clocks, directly connected in front of the DDDAC gives a very noticeable sound quality improvement
Whatever I2S signal, regardless if coming from low end or high end source, was connected to the fifopi input did not make any difference to my ears and if there was it was no way near the improvement the reclocker did compared to the best unclocked source I have available, remember REAL A to B comparison remotely controlled from the listening position with 100% time synchronized tracks playing
For me this means. If you do not have a reclocker as yet, start with this and (using good clocks) you will make a nice jump, regardless of all the other aspects. Just stay with whatever I2S source you might have right now. Pi or WaveIO or whatever you have. Just enjoy this step for now
Now from here, there is always a path to improve with things like cable length, power supplies, other clocks, etc etc
If any one believes that changing things before it goes to the reclocker, fine of course, but nothing I could seriously hear a difference from. So my focus will be at the reclocker and after the reclocker side of things for now
Haha, again long story, but I hope it helps putting in perspective
Reclocking with good clocks, directly connected in front of the DDDAC gives a very noticeable sound quality improvement
Whatever I2S signal, regardless if coming from low end or high end source, was connected to the fifopi input did not make any difference to my ears and if there was it was no way near the improvement the reclocker did compared to the best unclocked source I have available, remember REAL A to B comparison remotely controlled from the listening position with 100% time synchronized tracks playing
For me this means. If you do not have a reclocker as yet, start with this and (using good clocks) you will make a nice jump, regardless of all the other aspects. Just stay with whatever I2S source you might have right now. Pi or WaveIO or whatever you have. Just enjoy this step for now
Now from here, there is always a path to improve with things like cable length, power supplies, other clocks, etc etc
If any one believes that changing things before it goes to the reclocker, fine of course, but nothing I could seriously hear a difference from. So my focus will be at the reclocker and after the reclocker side of things for now
Haha, again long story, but I hope it helps putting in perspective
I know the story was pretty long, so let’s try to summarize what my real conclusions are
Reclocking with good clocks, directly connected in front of the DDDAC gives a very noticeable sound quality improvement
Whatever I2S signal, regardless if coming from low end or high end source, was connected to the fifopi input did not make any difference to my ears and if there was it was no way near the improvement the reclocker did compared to the best unclocked source I have available, remember REAL A to B comparison remotely controlled from the listening position with 100% time synchronized tracks playing
For me this means. If you do not have a reclocker as yet, start with this and (using good clocks) you will make a nice jump, regardless of all the other aspects. Just stay with whatever I2S source you might have right now. Pi or WaveIO or whatever you have. Just enjoy this step for now
Now from here, there is always a path to improve with things like cable length, power supplies, other clocks, etc etc
If any one believes that changing things before it goes to the reclocker, fine of course, but nothing I could seriously hear a difference from. So my focus will be at the reclocker and after the reclocker side of things for now
Haha, again long story, but I hope it helps putting in perspective
Thanks Doede it makes sense and helps alot.
Now to get the time to put all the good stuff I bought into my 2nd DDDAC and once happy put it all in my main DDDAC
Thank You Doede for taking a very logical and organized approach to the clock upgrade/I2S source issue. I have been reading here about all the various upgrades and combinations of PS, source and clocks and there seemed to be an almost infinite variety of methods and opinions on the results.
Far from being a long story, you have successfully summarized your approach and results. You have also supplied a logical and practical way forward for all of us who are curious about improving sound quality.
My own dac is very basic. 1 board and the waveIO mounted very close so that the cable to DDDac is less than 2 inches long. I also have my regulators mounted at the main board connectors. I have no idea if these methods improve the sound.
What did improve the sound by a huge margin was going to a balanced preamp and running dddac balanced with coupling capacitors. Yes, the dreaded capacitors!!
I think that much of the gains made when going to transformers in a single ended system are from using the full balanced signal output of DDDAC. I would like to try more boards and transformers in the near future. Perhaps even clock upgrades. But for anyone with a balanced preamp, I suggest going balanced with capacitors as a first step.
Far from being a long story, you have successfully summarized your approach and results. You have also supplied a logical and practical way forward for all of us who are curious about improving sound quality.
My own dac is very basic. 1 board and the waveIO mounted very close so that the cable to DDDac is less than 2 inches long. I also have my regulators mounted at the main board connectors. I have no idea if these methods improve the sound.
What did improve the sound by a huge margin was going to a balanced preamp and running dddac balanced with coupling capacitors. Yes, the dreaded capacitors!!
I think that much of the gains made when going to transformers in a single ended system are from using the full balanced signal output of DDDAC. I would like to try more boards and transformers in the near future. Perhaps even clock upgrades. But for anyone with a balanced preamp, I suggest going balanced with capacitors as a first step.