Current output stage schematics
Hi tubee,
That's bad luck,
similar things happened to me after modifications or repairs. Have you already checked if perhaps you dropped some soldering tin on the circuit board that causes short circuit?, or accidently interrupted a track during component replacement. A thourough visual inspection might help locating the fault. Also check if all connectors are placed in the right locations, and also check for bad soldering joints, especially the connector pin's. These tend to interrupt after working on single-sided circuit boards. You can also remove the kwak 7 clock and restore the original setup, to see if it's causing problems. Try to elliminate errors step, by step. Another good tip with modding is to do one modification at a time and then recheck, so when something goes wrong you know exactly what modification caused it. It's also good practice to visually inspect the entire board after a modification has been made, before switching-on. When something doesn't work, switch-off immediately to avoid damage.
I also added a schematic diagram of the current octal D-I DAC op-amp / tube output stage, I am using right now, but it's still a concept.
This shows how both OP-amp and tube output can be switched in the mixed mode. The tube stage is optimized, it now uses ECC83 double triodes. Gain has been corrected, cascaded zener diodes reduce ripple current to very low values. I used a FET-based current source (J508), noise is very low now. Two cathode followers have been connected in parallell for lower output impedance, since I don't need a differential output. T1...T4 are used for power supply stabilization for each separate triode, they fully eliminate hum. Note that pin 9 of T1a is connected to ground, also to reduce hum. The I/V stage and differential OP-amp stage are added for illustration.
S1 closed: OP-amp mode, S2 closed: Tube mode, both S1 and S2 closed: mixed mode. output attenuators: R14,R15 and R24,R25
I also designed a different soft-start filament supply. It uses a 7805 voltage regulator together with a 33 Ohm resistor to form a 150mA current source (I=5/R). The filaments start-up beautifully without any current surge, because the current cannot exeed 150mA. The current sources also provide very low ripple current at the filaments.
I generated the +250V HV supply by using 2 standard transformers, When using 110V mains voltage, N3 has both 115V windings in parallel, N4 has both 115V windings in series. The 18V windings also provide the negative 50V supply.
I am using 2 cascaded schaffner mains filters and a 33nF capacitor for mains filtering. S3 switches-on the HV supply when the filaments have heated-up, and switches it off again before filament supply is switched-off.
Hi tubee,
That's bad luck,
similar things happened to me after modifications or repairs. Have you already checked if perhaps you dropped some soldering tin on the circuit board that causes short circuit?, or accidently interrupted a track during component replacement. A thourough visual inspection might help locating the fault. Also check if all connectors are placed in the right locations, and also check for bad soldering joints, especially the connector pin's. These tend to interrupt after working on single-sided circuit boards. You can also remove the kwak 7 clock and restore the original setup, to see if it's causing problems. Try to elliminate errors step, by step. Another good tip with modding is to do one modification at a time and then recheck, so when something goes wrong you know exactly what modification caused it. It's also good practice to visually inspect the entire board after a modification has been made, before switching-on. When something doesn't work, switch-off immediately to avoid damage.
I also added a schematic diagram of the current octal D-I DAC op-amp / tube output stage, I am using right now, but it's still a concept.
This shows how both OP-amp and tube output can be switched in the mixed mode. The tube stage is optimized, it now uses ECC83 double triodes. Gain has been corrected, cascaded zener diodes reduce ripple current to very low values. I used a FET-based current source (J508), noise is very low now. Two cathode followers have been connected in parallell for lower output impedance, since I don't need a differential output. T1...T4 are used for power supply stabilization for each separate triode, they fully eliminate hum. Note that pin 9 of T1a is connected to ground, also to reduce hum. The I/V stage and differential OP-amp stage are added for illustration.
S1 closed: OP-amp mode, S2 closed: Tube mode, both S1 and S2 closed: mixed mode. output attenuators: R14,R15 and R24,R25
I also designed a different soft-start filament supply. It uses a 7805 voltage regulator together with a 33 Ohm resistor to form a 150mA current source (I=5/R). The filaments start-up beautifully without any current surge, because the current cannot exeed 150mA. The current sources also provide very low ripple current at the filaments.
I generated the +250V HV supply by using 2 standard transformers, When using 110V mains voltage, N3 has both 115V windings in parallel, N4 has both 115V windings in series. The 18V windings also provide the negative 50V supply.
I am using 2 cascaded schaffner mains filters and a 33nF capacitor for mains filtering. S3 switches-on the HV supply when the filaments have heated-up, and switches it off again before filament supply is switched-off.
Checked it all
Ecdesign, thank you for all tips.
I did the mod as thourough as possible, did some other mods before with a CD640 to test a lot of parts, and on the 304 i visually checked everything 3 times before powering-on (also on polarity caps)
I think the uprocessor of display is faulty, strange, before modding it played fine. My first ESD-accident? Will lend a scope soon, and check some frequencies/waveforms here and there.
Now the CD 304 is dead, i have been looking for space in the CD660 CDP in have build to place my 4 dac project in. It will be very fiddly to get the D-I pcb's in it but when i move some pcb's maybe it could fit. Thinking of a acrylate top cover to show that very nice D-I dac V-8 engine!
Its to beautiful to cover under a hood.
Ecdesign, thank you for all tips.
I did the mod as thourough as possible, did some other mods before with a CD640 to test a lot of parts, and on the 304 i visually checked everything 3 times before powering-on (also on polarity caps)
I think the uprocessor of display is faulty, strange, before modding it played fine. My first ESD-accident? Will lend a scope soon, and check some frequencies/waveforms here and there.
Now the CD 304 is dead, i have been looking for space in the CD660 CDP in have build to place my 4 dac project in. It will be very fiddly to get the D-I pcb's in it but when i move some pcb's maybe it could fit. Thinking of a acrylate top cover to show that very nice D-I dac V-8 engine!
Its to beautiful to cover under a hood.Hi ecdesigns I have been following this thread since you started it.
I cannot design myself but I can copy others, and your design is fantastic . I was going to build a double da converter using two tda1541 s1 chips and a cs8412 but am holding off to see where yours goes.
Do you sleep?.
Its good to see someone who has great passion to build things to the best of there ability
I would be very keen to purchase some boards upon this circuits completion
I cannot design myself but I can copy others, and your design is fantastic . I was going to build a double da converter using two tda1541 s1 chips and a cs8412 but am holding off to see where yours goes.
Do you sleep?.
Its good to see someone who has great passion to build things to the best of there ability
I would be very keen to purchase some boards upon this circuits completion
Hi Ecdesigns,
about attaching schematics:
probably to large file. You can reduce file, if you put it in grayscale format. Then schematics will be better quality.
If you cant change format, i can do easy with Photoshop.
Just send me file and i will send you back.
It is my job- graphic designer.
I do agree that design is fantastic.
regards
my email: bkragl@yahoo.co.uk
about attaching schematics:
probably to large file. You can reduce file, if you put it in grayscale format. Then schematics will be better quality.
If you cant change format, i can do easy with Photoshop.
Just send me file and i will send you back.
It is my job- graphic designer.
I do agree that design is fantastic.
regards
my email: bkragl@yahoo.co.uk
output stage schematics, second attempt
Hi weegs,
Thanks for your compliments,
> The octal D-I DAC design is taking shape now, currently I am making last modifications to the tube output stage, then I will start with the circuit board for it. Again 2 separate compact modules for L and R channels will be used.
This design stage is quite complicated as these modules have to be "merged" with the housing I am planning to use.
Next on the agenda is the digital mainboard containing the already completed SPDIF interface, USB to I2S module, system control module and the digital input selection module. I will post a block diagram later.
There will also be a small attenuator module for Tube / OP-amp and 1 external analog source selection, so you will still be able to route a analog channel.
After all modules are ready and fuctioning, I will start designing the main power supply. Problem here is the high-quality DC converter for the tube output stage that is necessary for battery operation only. Yes, I will try to provide a connection for supplying this Octal D-I DAC, tube output included, on 4 X 12V batteries too. Power consumption is approximately 30 Watts at the moment, including the differential tube output stages, so battery operation must be possible.
> At the moment I don't sleep much, remember I also have to run my small electronics company and this project is a "hobby" project along the side. But as they say, "hobby" projects are most important.
> Yes, electronics is my passion, as I indicated before. When I build something I put my heart and soul into it, always striving for perfection, and that's about the best one can do.
> When the octal D-I DAC design is finished, I plan to get professional boards manufactured, I will start with a small batch, then re-order as desired.
Hi weegs,
Thanks for your compliments,
> The octal D-I DAC design is taking shape now, currently I am making last modifications to the tube output stage, then I will start with the circuit board for it. Again 2 separate compact modules for L and R channels will be used.
This design stage is quite complicated as these modules have to be "merged" with the housing I am planning to use.
Next on the agenda is the digital mainboard containing the already completed SPDIF interface, USB to I2S module, system control module and the digital input selection module. I will post a block diagram later.
There will also be a small attenuator module for Tube / OP-amp and 1 external analog source selection, so you will still be able to route a analog channel.
After all modules are ready and fuctioning, I will start designing the main power supply. Problem here is the high-quality DC converter for the tube output stage that is necessary for battery operation only. Yes, I will try to provide a connection for supplying this Octal D-I DAC, tube output included, on 4 X 12V batteries too. Power consumption is approximately 30 Watts at the moment, including the differential tube output stages, so battery operation must be possible.
> At the moment I don't sleep much, remember I also have to run my small electronics company and this project is a "hobby" project along the side. But as they say, "hobby" projects are most important
.> Yes, electronics is my passion, as I indicated before. When I build something I put my heart and soul into it, always striving for perfection, and that's about the best one can do.
> When the octal D-I DAC design is finished, I plan to get professional boards manufactured, I will start with a small batch, then re-order as desired.
Attachments
glowing tubes
Hi a333bt,
Thanks for your compliments [post#325],
Thanks for the tips, but the file was OK, I already used the GIMP to edit both photographs and scans when I started this thread. Seemed there was a problem with attachments, this was also indicated in the anouncements. But I am very pleased that you offered your help.
You indicated that you like the design, well I got a little surprise. I just couldn't wait any longer as I wanted to get an overall impression of how the octal D-I DAC could look like. So I mounted the sinewave shaped protection grid on a existing switchbox housing with a tube chassis plate I milled today, and mounted the tubes with status indicator LED's.
Enjoy!
Hi a333bt,
Thanks for your compliments [post#325],
Thanks for the tips, but the file was OK, I already used the GIMP to edit both photographs and scans when I started this thread. Seemed there was a problem with attachments, this was also indicated in the anouncements. But I am very pleased that you offered your help.
You indicated that you like the design, well I got a little surprise. I just couldn't wait any longer as I wanted to get an overall impression of how the octal D-I DAC could look like. So I mounted the sinewave shaped protection grid on a existing switchbox housing with a tube chassis plate I milled today, and mounted the tubes with status indicator LED's.
Enjoy!
Attachments
Hi EC,
This dac looks very nice, well done.... very impressive indeed
Have you thought about including Guido's XO DAC PLL so spdif IP
can be used and still get low jitter performance. It would be a
shame for such a nice design to have compromised jitter
performance.
The DEM reclocking is also a great idea.
However I think the OP stage maybe can be improved to use the
excellent design to it's fullest potential.
My thoughts are that if you have 4 DACs, each outputting +-2mA for
a total of +-8mA then this is a lot of OP drive to work with, so use it
to your advantage, don't just stuff it into an opamp to deal with.
If a zero FB current conveyor type is used then the load R needs
only be 350R for a full 2V rms / phase. This means a single stage
OP that has very high OP drive and no follower or opamps. Refer
to John Curls comments on the Blowtorch preamp thread WRT OP
stage without follower.
With the above circuit approach running into such low value load
(350R) a huge dynamic range can be had for the I-V. >120dB is
easily possible.
Also, the I-V should be ideally direct coupled without caps in
signal path.
I think this would consitiute the ultimate OP stage for what looks
like maybe an ultimate dac.
Any thoughts?
Cheers,
Terry
This dac looks very nice, well done.... very impressive indeed
Have you thought about including Guido's XO DAC PLL so spdif IP
can be used and still get low jitter performance. It would be a
shame for such a nice design to have compromised jitter
performance.
The DEM reclocking is also a great idea.
However I think the OP stage maybe can be improved to use the
excellent design to it's fullest potential.
My thoughts are that if you have 4 DACs, each outputting +-2mA for
a total of +-8mA then this is a lot of OP drive to work with, so use it
to your advantage, don't just stuff it into an opamp to deal with.
If a zero FB current conveyor type is used then the load R needs
only be 350R for a full 2V rms / phase. This means a single stage
OP that has very high OP drive and no follower or opamps. Refer
to John Curls comments on the Blowtorch preamp thread WRT OP
stage without follower.
With the above circuit approach running into such low value load
(350R) a huge dynamic range can be had for the I-V. >120dB is
easily possible.
Also, the I-V should be ideally direct coupled without caps in
signal path.
I think this would consitiute the ultimate OP stage for what looks
like maybe an ultimate dac.
Any thoughts?
Cheers,
Terry
Hi Ecdesign,
i am slowly collecting parts for DAC.
I know when some day is no post from you, then you are working.
Well i have question: i wrote wrong part number (Farnell) instead of 50 resistors, which is minimum, i got 50 ICs 74HC161 Philips, which is as you said good, but i need only one.
If you need them, maybe for offering with PCBs, i can send you right away.
Do you some part number of sockets for ICs? I have cheap sockets, but i guess, from previous posts, they are not good.
Design is very nice and special.
Myself i will use I2S connection from CDPRO2 transport, which i have many years.
regards
i am slowly collecting parts for DAC.
I know when some day is no post from you, then you are working.
Well i have question: i wrote wrong part number (Farnell) instead of 50 resistors, which is minimum, i got 50 ICs 74HC161 Philips, which is as you said good, but i need only one.
If you need them, maybe for offering with PCBs, i can send you right away.
Do you some part number of sockets for ICs? I have cheap sockets, but i guess, from previous posts, they are not good.
Design is very nice and special.
Myself i will use I2S connection from CDPRO2 transport, which i have many years.
regards
hi tubee,
thanks for your reply [post#328, #332]
[post#328]
> Yes, this thread has already passed 22,000 views, I am very gratefull for for the interest and support you all have shown for this project
. Therefore I will take this opportunity to thank you all .
[post#332]
> The transistors in the ECC83 annodes (above the ECC83) are no current sources, but stabilize the annode voltage at 200V, very close to the annodes, they also provide a isolated supply for each triode. This way hum is greatly reduced.
> The filament current sources, connected to 20V, provide a constant current of 150mA, this way the filament is current driven and automatically settles around 12.6V (2 X 6.3V filaments in series, pin 4 and 5 used). Of course this current has to match the tube's filament specifications. Major advantage is low hum and a perfect "cold start" , limiting surge current when the cold filament has a low resistance. I plan to put these current sources in the main power supply.
thanks for your reply [post#328, #332]
[post#328]
> Yes, this thread has already passed 22,000 views, I am very gratefull for for the interest and support you all have shown for this project
. Therefore I will take this opportunity to thank you all .[post#332]
> The transistors in the ECC83 annodes (above the ECC83) are no current sources, but stabilize the annode voltage at 200V, very close to the annodes, they also provide a isolated supply for each triode. This way hum is greatly reduced.
> The filament current sources, connected to 20V, provide a constant current of 150mA, this way the filament is current driven and automatically settles around 12.6V (2 X 6.3V filaments in series, pin 4 and 5 used). Of course this current has to match the tube's filament specifications. Major advantage is low hum and a perfect "cold start" , limiting surge current when the cold filament has a low resistance. I plan to put these current sources in the main power supply.
octal D-I DAC improvements
Hi Terry Demol,
Thanks for your compliments [post#329],
The octal D-I DAC design is constantly veryfied by measurements (shown on this thread), many listening sessions, and by comparing it with my NOS reference DAC. Circuits have been modified / optimized many times to get optimal verifyable results.
> I use a differential SPDIF input signal, and have modified my CD player by adding a differential RS422 output stage [post#268]. So jitter is very low, as the CS8412 PLL runs very stable due to the stable SPDIF input signal. Guido's XO DAC PLL seems to be used often to upgrade a standard oversampling CD player containing a digital filter. I can imagine that a jittery clock can affect digital filter operation, and adding a XO DAC PLL can significantly improve sound quality.
I could easily reserve some space on the digital mainboard for a jitter reduction module if desired.
However, the octal D-I DAC is quite special compared to other DAC concepts. It's completely filterless, no digital filter, no analog filter, and uses no oversampling. All 8 DAC chips are running on a relaxed 2.8224 MHz BCK. The timing chain seems to have an averaging effect on jitter as well, as the clone samples appear at the output after n*BCK delay . All these octal D-I DAC properties seem to make it relatively immune to jitter, this can easily be verified by listening sessions.
> Yes DEM clock synchronizing with BCK has made a big improvement, especially in improved clarity and detail.
> The OP-amp I/V stages are highly optimized, they have very low distortion as can be seen in [post#269] oscillograms. Output is a near perfect step signal without ringing or sagging, indicating feedback causes no appearent problems here. The unity-gain diff amp cancels out the last bits of interference and the large DC offsets from both non-inverted and inverted DAC chip groups. It's output signal high frequency rolloff is matched to the tube output stage. Output signal is now about 18 Vpp, (taking full advantage of the +/- 15V power supply). The signal can now be attenuated close to the RCA socket for optimal SN ratio. Basically if you hear the octal D-I DAC producing noise, it's on the CD recording, yes even studio recording equipment noise is reproduced accurately, microphones picking-up radio stations "hidden" in the noise, included. I have already tried a lot of different setups, passive I/V, several I/V stages with discrete components both FET's and transistors, with and without feedback.
Finally I got the precision, speed and natural sound reproduction using the current circuit. Same with the diff-amp. The circuit is very compact, needs no adjustments, and assembly is a piece of cake, making it ideal for this DIY project. Yes, I always have to take this into account, in every design phase. This DIY project must be relatively easy to build, and circuits have function rock-solid to avoid any problems after building it. And last but not least, the octal D-I DAC parts should be reasonably priced.
> DC coupling: have a look at schematiciv1.jpg [post#50] and read the text above the diagrams. Now have a look at diagram tube2sch.jpg [post#326]. Both my I/V and differental OP-amp stages are already fully DC coupled, only the tube output stage isn't. But when using the mixed mode, these two circuits "merge" and use the best of both worlds, precision and DC drive from the very low distorton OP-amp stage, and just that little extra transparency and warmer midrange from the tube output stage. So you get a typical semiconductortube sound .
When comparing the octal D-I DAC with my NOS reference DAC as I have been doing during octal D-I DAC development, it now makes my NOS reference DAC sound like a transistor radio. So both the I/V stage and the diff-amp aren't that bad after all.
> Sure improvements / modifications can be made, as this design is fully modular, and ideal for modding and optimizing, analog output stages included. One could design a passive I/V module and replace the existing module, using the existing output connector. Or build a I/V converter using discrete components. If you have any interesting diagrams, I could have a look at them. Personally I am starting to get a bit nervous, modifying this almost magic sounding DAC, as I don't want to loose even the slightest bit of sound quality it produces right now.
Hi Terry Demol,
Thanks for your compliments [post#329],
The octal D-I DAC design is constantly veryfied by measurements (shown on this thread), many listening sessions, and by comparing it with my NOS reference DAC. Circuits have been modified / optimized many times to get optimal verifyable results.
> I use a differential SPDIF input signal, and have modified my CD player by adding a differential RS422 output stage [post#268]. So jitter is very low, as the CS8412 PLL runs very stable due to the stable SPDIF input signal. Guido's XO DAC PLL seems to be used often to upgrade a standard oversampling CD player containing a digital filter. I can imagine that a jittery clock can affect digital filter operation, and adding a XO DAC PLL can significantly improve sound quality.
I could easily reserve some space on the digital mainboard for a jitter reduction module if desired.
However, the octal D-I DAC is quite special compared to other DAC concepts. It's completely filterless, no digital filter, no analog filter, and uses no oversampling. All 8 DAC chips are running on a relaxed 2.8224 MHz BCK. The timing chain seems to have an averaging effect on jitter as well, as the clone samples appear at the output after n*BCK delay . All these octal D-I DAC properties seem to make it relatively immune to jitter, this can easily be verified by listening sessions.
> Yes DEM clock synchronizing with BCK has made a big improvement, especially in improved clarity and detail.
> The OP-amp I/V stages are highly optimized, they have very low distortion as can be seen in [post#269] oscillograms. Output is a near perfect step signal without ringing or sagging, indicating feedback causes no appearent problems here. The unity-gain diff amp cancels out the last bits of interference and the large DC offsets from both non-inverted and inverted DAC chip groups. It's output signal high frequency rolloff is matched to the tube output stage. Output signal is now about 18 Vpp, (taking full advantage of the +/- 15V power supply). The signal can now be attenuated close to the RCA socket for optimal SN ratio. Basically if you hear the octal D-I DAC producing noise, it's on the CD recording, yes even studio recording equipment noise is reproduced accurately, microphones picking-up radio stations "hidden" in the noise, included. I have already tried a lot of different setups, passive I/V, several I/V stages with discrete components both FET's and transistors, with and without feedback.
Finally I got the precision, speed and natural sound reproduction using the current circuit. Same with the diff-amp. The circuit is very compact, needs no adjustments, and assembly is a piece of cake, making it ideal for this DIY project. Yes, I always have to take this into account, in every design phase. This DIY project must be relatively easy to build, and circuits have function rock-solid to avoid any problems after building it. And last but not least, the octal D-I DAC parts should be reasonably priced.
> DC coupling: have a look at schematiciv1.jpg [post#50] and read the text above the diagrams. Now have a look at diagram tube2sch.jpg [post#326]. Both my I/V and differental OP-amp stages are already fully DC coupled, only the tube output stage isn't. But when using the mixed mode, these two circuits "merge" and use the best of both worlds, precision and DC drive from the very low distorton OP-amp stage, and just that little extra transparency and warmer midrange from the tube output stage. So you get a typical semiconductortube sound
.When comparing the octal D-I DAC with my NOS reference DAC as I have been doing during octal D-I DAC development, it now makes my NOS reference DAC sound like a transistor radio. So both the I/V stage and the diff-amp aren't that bad after all.
> Sure improvements / modifications can be made, as this design is fully modular, and ideal for modding and optimizing, analog output stages included. One could design a passive I/V module and replace the existing module, using the existing output connector. Or build a I/V converter using discrete components. If you have any interesting diagrams, I could have a look at them. Personally I am starting to get a bit nervous, modifying this almost magic sounding DAC, as I don't want to loose even the slightest bit of sound quality it produces right now.
I2S interface
Hi a333bt,
thanks for the compliment [post#330],
Yes, I wanted to use a design for the housing that matches the octal D-I DAC concept. The status indication is functional and adds something very special to this DAC. Note that the photo only gives a rough impression, because the aluminum plates have not been annodized yet. The protection grid can also be made of high-gloss stainless steel.
> As you have guessed I have been very busy, despite the high temperatures over here. One result of being busy is that the tube output stage PCB is already routed, note all circuit boards are both routed and manufactured manually. I thought it would be interesting to show a photograph of how the analog mainboard was made, photo shows the analog mainboard being etched, bottom left the etched PCB, bottom right: PCB after tinning (using soldering paste). After tinning the PCB is thoroughly cleaned using terpentine. Then it's cleaned with liquid soap, to get a high-gloss finish. Finally the holes are drilled.
The differential tube output stage is a compact, easy to build module for one channel, so two of them are needed. It even includes the status LED's in the centre of the tube sockets. It has a power supply connector, a analog I/O connector and a status LED connector. Special problem was the location of the analog mainboard, so the tube output stage PCB needed to be very flat at both sides to create room for the OP-amp I/V output stage module heatsink. So the higher components (electrolytic capacitors and the audyn output capacitor) had to be located in the centre of the PCB. Also great care had to be taken routing the high voltage power supply lines. I am using High quality ceramic PCB sockets for the tubes. The tube output stage has one cermet trim potentiometer for adjusting the differential stage DC balance. The PCB is mounted flat against the aluminum chassis plate. Each tube socket is fixed firmly to this chassis plate as well.
I will post some pictures of the tube output stage when completed.
> Very good that you mentioned the I2S interface, I could provide a differential I2S interface, so the CD player and DAC don't have to be placed close together when using I2S directly from the player. When using a RS422 (DS8922 chip) all 3 signals (BCK, DATA and WS) can be differentially / buffered using just 2 chip's on each side.
Hi a333bt,
thanks for the compliment [post#330],
Yes, I wanted to use a design for the housing that matches the octal D-I DAC concept. The status indication is functional and adds something very special to this DAC. Note that the photo only gives a rough impression, because the aluminum plates have not been annodized yet. The protection grid can also be made of high-gloss stainless steel.
> As you have guessed I have been very busy, despite the high temperatures over here. One result of being busy is that the tube output stage PCB is already routed, note all circuit boards are both routed and manufactured manually. I thought it would be interesting to show a photograph of how the analog mainboard was made, photo shows the analog mainboard being etched, bottom left the etched PCB, bottom right: PCB after tinning (using soldering paste). After tinning the PCB is thoroughly cleaned using terpentine. Then it's cleaned with liquid soap, to get a high-gloss finish. Finally the holes are drilled.
The differential tube output stage is a compact, easy to build module for one channel, so two of them are needed. It even includes the status LED's in the centre of the tube sockets. It has a power supply connector, a analog I/O connector and a status LED connector. Special problem was the location of the analog mainboard, so the tube output stage PCB needed to be very flat at both sides to create room for the OP-amp I/V output stage module heatsink. So the higher components (electrolytic capacitors and the audyn output capacitor) had to be located in the centre of the PCB. Also great care had to be taken routing the high voltage power supply lines. I am using High quality ceramic PCB sockets for the tubes. The tube output stage has one cermet trim potentiometer for adjusting the differential stage DC balance. The PCB is mounted flat against the aluminum chassis plate. Each tube socket is fixed firmly to this chassis plate as well.
I will post some pictures of the tube output stage when completed.
> Very good that you mentioned the I2S interface, I could provide a differential I2S interface, so the CD player and DAC don't have to be placed close together when using I2S directly from the player. When using a RS422 (DS8922 chip) all 3 signals (BCK, DATA and WS) can be differentially / buffered using just 2 chip's on each side.
Attachments
Re: octal D-I DAC improvements
I basically think attenuation before the Interlink is the wrong way to go for ultimate designs. By doing this you introduce a high output impedance, possibly leading to more noise, less good transient behaviour and humm pick up. I think it is a shame of the well designed I/V stage to allow this to happen...
On your second comment I need to grins a bit. This can only mean 2 things:
1. Your reference NOS DAC indeed sounds like a transistor radio...
2. You are overvalueing the sound differences in too prozaic language....
Without kidding, I mean this. If you follow the post, you write from the beginning, the sound is great and supperb (which I believe from what I see, no issue with that), but every tweak in your design brings your sound experinece to unparalleled highs.... This may be the way you perceive it, which again is fine.... But everyone in the DIY, who has been building, tweaking and listening for a couple of years and went through several design(stage)s, knows that it is not realistic that such huge difference occur. This is blurring a bit your excelent story and is not in line with your otherwise so serious approach.
Have you been abe to compare your DAC already with high resolution media? This can be very revealing and helps put things in perspective.
Keep up the good work, I like very much what you are doing and sharing here !!!
doede
I basically think attenuation before the Interlink is the wrong way to go for ultimate designs. By doing this you introduce a high output impedance, possibly leading to more noise, less good transient behaviour and humm pick up. I think it is a shame of the well designed I/V stage to allow this to happen...
On your second comment I need to grins a bit. This can only mean 2 things:
1. Your reference NOS DAC indeed sounds like a transistor radio...
2. You are overvalueing the sound differences in too prozaic language....
Without kidding, I mean this. If you follow the post, you write from the beginning, the sound is great and supperb (which I believe from what I see, no issue with that), but every tweak in your design brings your sound experinece to unparalleled highs.... This may be the way you perceive it, which again is fine.... But everyone in the DIY, who has been building, tweaking and listening for a couple of years and went through several design(stage)s, knows that it is not realistic that such huge difference occur. This is blurring a bit your excelent story and is not in line with your otherwise so serious approach.
Have you been abe to compare your DAC already with high resolution media? This can be very revealing
and helps put things in perspective.Keep up the good work, I like very much what you are doing and sharing here !!!
doede
Re: Re: octal D-I DAC improvements
You have to go to a theater and listen to an orkestra or a live band to "know" how the huge difference is realistic or not . Sclaed out through your favourite listening parameters; Then , also how you are prepared to receive music is important Becouse the music "hosts" you if you are prepared , otherwise it will let you away .
dddac said:
You have to go to a theater and listen to an orkestra or a live band to "know" how the huge difference is realistic or not . Sclaed out through your favourite listening parameters; Then , also how you are prepared to receive music is important Becouse the music "hosts" you if you are prepared , otherwise it will let you away .
