Hello XG:
I know that Arcam used those high resistor-values (390 ohm) also in their 1541-players, but not that high. It could have problably the same result as that 10pF cap parallel of the I2S level, slowing down the highest frequencies? Not sure about that.
Hello Dragonmaster: i don't want anybody to test the schematic for me, with time i could do that myself, and figure out the right R-values.
If one of us has build a schematic to put a dc-voltage on the I2S signal, i would be curious for the results of it: Does it give better sound, lower jitter or whatever.
In short: Is it worth time and effort to build it?
(i want 4 dacs parallel, so have to build 12 of them then.)
I know that Arcam used those high resistor-values (390 ohm) also in their 1541-players, but not that high. It could have problably the same result as that 10pF cap parallel of the I2S level, slowing down the highest frequencies? Not sure about that.
Hello Dragonmaster: i don't want anybody to test the schematic for me, with time i could do that myself, and figure out the right R-values.
If one of us has build a schematic to put a dc-voltage on the I2S signal, i would be curious for the results of it: Does it give better sound, lower jitter or whatever.
In short: Is it worth time and effort to build it?
(i want 4 dacs parallel, so have to build 12 of them then.)
hi,tubee
IMHO,the layout of TDA1541 NOS DAC is much much more important.here is my experience in the TDA1541 NOS DAC which designed by my friend....
http://www.hifidiy.net/dispbbs.asp?BoardID=2&id=4499&replyID=4499&star=38&skin=0
edit: correction of URL ( Jean-Paul )
IMHO,the layout of TDA1541 NOS DAC is much much more important.here is my experience in the TDA1541 NOS DAC which designed by my friend....
http://www.hifidiy.net/dispbbs.asp?BoardID=2&id=4499&replyID=4499&star=38&skin=0
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
edit: correction of URL ( Jean-Paul )
Hi XG
I have just mounted four 1541's on perfboard with control logic, i use smaller MKT caps for current divider, so they sit just beside dacs. Groundplane is C-shape around each dac. All 4 ground ''C''s will be connected in the middle of 4 dacs by a star ground, on which all other grounds are connected (tube I/V and Gnd of SAA7210 board) Each dac gets time-shifted signals from shift registers, so the output signal should be much smoother on the output with 4 dacs. I'm curious for the results of it, more then a year busy with it.
And will show some pictures in future ofcourse.
I have just mounted four 1541's on perfboard with control logic, i use smaller MKT caps for current divider, so they sit just beside dacs. Groundplane is C-shape around each dac. All 4 ground ''C''s will be connected in the middle of 4 dacs by a star ground, on which all other grounds are connected (tube I/V and Gnd of SAA7210 board) Each dac gets time-shifted signals from shift registers, so the output signal should be much smoother on the output with 4 dacs. I'm curious for the results of it, more then a year busy with it.
And will show some pictures in future ofcourse.
Forgot to say: the PS will be double regulated (LM317/337 & TL431) but cannot be mounted beside dacs(space) That's why i make a sandwich of 2 pcb's, 1 with logic and dacs, underneath it the PS with 12 TL431's to all different power connections. The connections from "down to above" will be with bead-on-wire's.
XG: i saw you used tantalium for + and - 5V, and Elna lytic for the analog -15V PS, is that right?
XG: i saw you used tantalium for + and - 5V, and Elna lytic for the analog -15V PS, is that right?
sorry,off topic
hi,tubee
I don't think a star ground is suit for digital circuit layout.In my mind,some keys of layout is
1,reduce the return current loop area as soon as possible
2,control the return current loop path.
3,lower the indutance of ground
4,.....
tubee said:
hi,tubee
I don't think a star ground is suit for digital circuit layout.In my mind,some keys of layout is
1,reduce the return current loop area as soon as possible
2,control the return current loop path.
3,lower the indutance of ground
4,.....
yes,but the lytic cap is sanyotubee said:
http://www.diyaudio.com/forums/showthread.php?postid=376132#post376132
Did anybody try this schematic in practice?
It only works for me with up to about 4K as opposed to 6K2.
I have not checked how it performs on a scope, but I get clean audio trough up to about that point.
Edit: Im feeding it NOS so im using at most 2.822Mhz
Did anybody try this schematic in practice?
It only works for me with up to about 4K as opposed to 6K2.
I have not checked how it performs on a scope, but I get clean audio trough up to about that point.
Edit: Im feeding it NOS so im using at most 2.822Mhz
Nice work hifi! Especially in smd it's possible to get such good results.
I problably drop the use of the I2S leveler, because i have to build 12 of them (4 dacs) so space will be limited. Every time i get a step forward with pcb, the space gets smaller.
I work allready on two levels now, and i see with only 1 dac you worked with an extra level also.
I am curious if it is possible to fiddle a dem-reclock on the pcb too.
regards:
I problably drop the use of the I2S leveler, because i have to build 12 of them (4 dacs) so space will be limited. Every time i get a step forward with pcb, the space gets smaller.
I work allready on two levels now, and i see with only 1 dac you worked with an extra level also.
I am curious if it is possible to fiddle a dem-reclock on the pcb too.
regards:
I settled on 3K1 in series with the datalines and I do belive the sound improved by it. So sure its worthwhile but not earthshattering.
But I also implemented the DEM-reclocker att about the same time and had a little difficulty with using 4FS, the lower bits would on occasion distort. But with 8FS everything works well and I can pass lowlevel sinewaves without trouble.
So, Im not quite sure what to attribute to what. In any case you can always implement both of these circuits on separate boards without making a total mess, after the dac is completed that is.
To set things in perspective, using a separate PLL in the dac and making it controll the oscillator in the CD & Using Jocko:s Spdif interfaces on both ends. Made a much larger difference to me then the Dem/Level adjustment did.
But I also implemented the DEM-reclocker att about the same time and had a little difficulty with using 4FS, the lower bits would on occasion distort. But with 8FS everything works well and I can pass lowlevel sinewaves without trouble.
So, Im not quite sure what to attribute to what. In any case you can always implement both of these circuits on separate boards without making a total mess, after the dac is completed that is.
To set things in perspective, using a separate PLL in the dac and making it controll the oscillator in the CD & Using Jocko:s Spdif interfaces on both ends. Made a much larger difference to me then the Dem/Level adjustment did.
I use one of Tent:s PLL:s in the dac. And a homebrew digital out in the transport which I made "tentlink" capable.
http://www.diyaudio.com/forums/showthread.php?postid=794910#post794910
http://www.diyaudio.com/forums/showthread.php?postid=794910#post794910
Hifi, Jocko's interface is not of any use in my setup, i will simply take I2S of the Saa7210 in a cd-player, and feed this to logic and four 1541's.
Have tried Kwak 7 and this clock will deliver the 11.2896 clock to SAA7210. But there has allways been thoughts in my mind try a PLL in the clock. First will use kwak's, draw a SMD pcb for the Kwak with french TCI-software.
Have build a solid sandwiched MDF/lead/acryllic housing allready.
Why I2s? To keep it as simple as possible. When there is a need of S/PDF in future, gonna use the CS8402.
Have tried Kwak 7 and this clock will deliver the 11.2896 clock to SAA7210. But there has allways been thoughts in my mind try a PLL in the clock. First will use kwak's, draw a SMD pcb for the Kwak with french TCI-software.
Have build a solid sandwiched MDF/lead/acryllic housing allready.
Why I2s? To keep it as simple as possible. When there is a need of S/PDF in future, gonna use the CS8402.
I was just trying to put things in perspective, not marking out the way for you to go..
Btw I did some measurements on the dac. Using The -60db dithered sine provided by Pedja.
The meassurens are quite repeatable, and averaged ten times, they are almost identical between channels. (With exception of a little more K5 in one channel)
K2 -49db
K3 -58-60db (it hides in the noisefloor between measurements)
Quite good for a Tda1541 but Im not sure if its attributed to the S2 grade or the i2s level adjustment.
Btw I did some measurements on the dac. Using The -60db dithered sine provided by Pedja.
The meassurens are quite repeatable, and averaged ten times, they are almost identical between channels. (With exception of a little more K5 in one channel)
K2 -49db
K3 -58-60db (it hides in the noisefloor between measurements)
Quite good for a Tda1541 but Im not sure if its attributed to the S2 grade or the i2s level adjustment.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.