Re: Octal D-I DAC sketches
Absolutely beautiful, you seem to have the rare balance of technology + art.
-ecdesigns- said:
Absolutely beautiful, you seem to have the rare balance of technology + art.
Ok Ecdesigns, i am very curious about the sound of it when you write this. I allmost cannot wait to get some pcb's and get started with it.
Btw, what do you think, should i build it as a dac, or build it in an existing player with I2S.
I have a solid Philips CD304mk2 (has CDM-1) with Spdif, dig out signal has to be modified then.
The player i wanted to put my own 4 dac project in is a newer modified CD660 with CDM-2, this CDM is very reliable (listened 20 years to aCD640 with CDM-2) The housing have i completely rebuild from wood/acrylic/lead, and is very solid. The CDM-2 is dampened with lead, and double isolated from case with self made rubber suspensions. When i listen with an stethoscope on the CD-clamp, i hear allmost nothing, even with hard tapping on case. From all pcb's i scavenged not used parts, wanted to tap I2S from SAA7210 and inject a separate (good) clock signal.
This is the CDP where i wanted my 4 dac project in:
[img=http://img217.imageshack.us/img217/4580/cdp1ml.th.jpg]
CDM 2 with self made rubber suspensions under the base (from a Toyota cooling system tube!)
[img=http://img135.imageshack.us/img135/9444/cdm25iv.th.jpg]
An externally hosted image should be here but it was not working when we last tested it.
The copper pins turned myself and have a very tight fit in base, the drawer slides with practically no play now.
Hi ecdesigns,
Thanks for explaining your reasons for your design choices. I'm impressed that your design choices are based on solid practical engineering and listening tests rather than voodoo that's rampant in high end.
I really like the look of the housing, has a clean classic look to it (eg, McIntosh), and the glow of tubes is a big plus. Keep up the great work.
BTW, where did you get your TDA1541A?
Do you have an estimate on when your kit will be finished for our use?
Thanks for explaining your reasons for your design choices. I'm impressed that your design choices are based on solid practical engineering and listening tests rather than voodoo that's rampant in high end.
I really like the look of the housing, has a clean classic look to it (eg, McIntosh), and the glow of tubes is a big plus. Keep up the great work.
BTW, where did you get your TDA1541A?
Do you have an estimate on when your kit will be finished for our use?
Forgot something: Ecdesigns: i like simple "box" design of gear, like Musical Fidelity, T&A and Marantz does. I don't like extravagant gear. But the idea of visibly showing the working principle in the cabinet design: great!
This thread is quiet...Looks like ecdesigns has finally decided to get some well deserved sleep or may be he finally decided to step out into the sun.
Re: Differential hybrid output stage
And that with a whole lot of room to improve the tube circuit...
dave
And that with a whole lot of room to improve the tube circuit...
dave
Re: Re: Differential hybrid output stage
I think so too. But I do like the idea of a grounded cathode amp into a cathode follower. Very Aikido like. See
http://www.tubecad.com/articles_2001/Inv_Dist_Cancellation/CompInvDisAmp.pdf
planet10 said:
I think so too. But I do like the idea of a grounded cathode amp into a cathode follower. Very Aikido like. See
http://www.tubecad.com/articles_2001/Inv_Dist_Cancellation/CompInvDisAmp.pdf
Re: Re: Re: Differential hybrid output stage
Nothing wrong with LTP>CFs. I'd put the CSS in the tail of the diff pair -- you could still have something on top to isolate it from the power supply or just build a power supply sufficient to make it unnecessary, The CF also need CCSs in their tails, and the ECC82 would be really low on my list of candidate tubes.
If i didn't need gain i'd consider losing the 1st stage completely, and if gain was needed something like a 12B4 CC diff amp, losing the CF stage.
dave
PS: i haven't built or listened to an Aikido yet (i do have boards), but i have heard some dissappointing comments from people i respect that have.
agent.5 said:
Nothing wrong with LTP>CFs. I'd put the CSS in the tail of the diff pair -- you could still have something on top to isolate it from the power supply or just build a power supply sufficient to make it unnecessary, The CF also need CCSs in their tails, and the ECC82 would be really low on my list of candidate tubes.
If i didn't need gain i'd consider losing the 1st stage completely, and if gain was needed something like a 12B4 CC diff amp, losing the CF stage.
dave
PS: i haven't built or listened to an Aikido yet (i do have boards), but i have heard some dissappointing comments from people i respect that have.
Re: Re: Re: Re: Differential hybrid output stage
agreed on all points. I think that Ecdesigns already has a CCS on the diff pair. We can tube roll later. I am sure ECC82 is fine because Ecdesigns has it and has designed for it. Having said that, I have some Tungsram E80CC that will be used in this project.
maybe DHTs for the Dac?
planet10 said:
agreed on all points. I think that Ecdesigns already has a CCS on the diff pair. We can tube roll later. I am sure ECC82 is fine because Ecdesigns has it and has designed for it. Having said that, I have some Tungsram E80CC that will be used in this project.
maybe DHTs for the Dac?
Differential tube output stage
Hi, planet10
Thanks for your reply,
Yes I agree, there is much to improve at the tube output stage, the diagram I posted is only a concept, I am already redesigning it. I have increased the power supply voltage to 210V, reduced input amplitude, lowered distortion. The only thing that is still bugging me is a weak hum on the output, that has to be solved too. I am also thinking of a cross feedback between the cathode followers and the differential stage. Then there will be a microcontroller in the octal D-I DAC that among a lot of other things, controlls sequential soft start-up and shut-down procedure of both, filament power supply and 210V main power supply to increase tube life.
Tips for improvement are very welcome.
Hi, planet10
Thanks for your reply,
Yes I agree, there is much to improve at the tube output stage, the diagram I posted is only a concept, I am already redesigning it. I have increased the power supply voltage to 210V, reduced input amplitude, lowered distortion. The only thing that is still bugging me is a weak hum on the output, that has to be solved too. I am also thinking of a cross feedback between the cathode followers and the differential stage. Then there will be a microcontroller in the octal D-I DAC that among a lot of other things, controlls sequential soft start-up and shut-down procedure of both, filament power supply and 210V main power supply to increase tube life.
Tips for improvement are very welcome.
Hi, MGH
Thanks for your reply
Yes, it has been "quit" on this thread, and that usually means I am working very hard. Coverage of new developments on the octal D-I DAC is comming soon. By the way, I bought the TDA1541A years ago here in Holland, but now the supplier ran out of stock. You should be able to find some TDA1541A's by searching the internet. I don't want to rush things, I have to stay critical to obtain the best performance for the D-I DAC, as you noticed this involves redesigns and optimalization, therefore it is very difficult to make an estimate of when the octal D-I DAC kit will be ready.
Thanks for your reply
Yes, it has been "quit" on this thread, and that usually means I am working very hard. Coverage of new developments on the octal D-I DAC is comming soon. By the way, I bought the TDA1541A years ago here in Holland, but now the supplier ran out of stock. You should be able to find some TDA1541A's by searching the internet. I don't want to rush things, I have to stay critical to obtain the best performance for the D-I DAC, as you noticed this involves redesigns and optimalization, therefore it is very difficult to make an estimate of when the octal D-I DAC kit will be ready.
OPA627 I/V output voltage
Hi, agent.5
Thanks for your reply,
The maximum output voltage of the OPA627 I/V stage with 4 TDA1541A's connected in parallell and 470 Ohm resistor is approx. 7.5Vpp, there are 2 I/V stages, so this means around 15Vpp going into the differential triode amplifier. This was a bit too high for the differential triode amplifier running on 120V. With 210V power supply and 100K annode resistors, I can have a low distortion signal at the cathode follower of approx. 30Vpp. The cathode resistor is now approx. 150K
The differential amplifier gain is approx. 3. So I had to attenuate both 7.5pp signals from ihe I/V converter to keep the output amplitude below 30Vpp. I use 4,7K from grid to ground and 4,7K from grid to OPA627 now, but I still have to optimize this. I Want the output signal to be as high as possible, so I can attenuate it, reducing noise and hum levels. According to measurements this seems to work fine. The output attenuator also limits "peak" currents caused by load capacitance.
Filaments are running on stabilized DC voltage.
These modifications significantly improved sound quality. As I menioned earlier, the only thing that is still bugging me now is a weak hum in the output signal that can be heared with very high volume settings.
I will post modified schematics soon.
Hi, agent.5
Thanks for your reply,
The maximum output voltage of the OPA627 I/V stage with 4 TDA1541A's connected in parallell and 470 Ohm resistor is approx. 7.5Vpp, there are 2 I/V stages, so this means around 15Vpp going into the differential triode amplifier. This was a bit too high for the differential triode amplifier running on 120V. With 210V power supply and 100K annode resistors, I can have a low distortion signal at the cathode follower of approx. 30Vpp. The cathode resistor is now approx. 150K
The differential amplifier gain is approx. 3. So I had to attenuate both 7.5pp signals from ihe I/V converter to keep the output amplitude below 30Vpp. I use 4,7K from grid to ground and 4,7K from grid to OPA627 now, but I still have to optimize this. I Want the output signal to be as high as possible, so I can attenuate it, reducing noise and hum levels. According to measurements this seems to work fine. The output attenuator also limits "peak" currents caused by load capacitance.
Filaments are running on stabilized DC voltage.
These modifications significantly improved sound quality. As I menioned earlier, the only thing that is still bugging me now is a weak hum in the output signal that can be heared with very high volume settings.
I will post modified schematics soon.
Re: OPA627 I/V output voltage
This also increases dramatically the output impedance of the stage as well (what need for a CF than?) and does not allow for interconnect cables with high(er) capacitance.
With an output impedance of lets say 15k, you will have a -3dB roll of at 40kHz with a cable with 265pF. Not unusual for interconnect cables
Question, I believe this tube stage has little to do with your DAC concept as such. Why not open a thread at the tube group? Sure you will get some more feedback there
What you publish here as Tube Stage might be kind of OK, but has imo nothing to do with "Ultimate". Your multiple DAC idea comes much closer to that statement I believe. Even though I don't like opams, but lets leave that discussion for a while... 😉
doede
-ecdesigns- said:
This also increases dramatically the output impedance of the stage as well (what need for a CF than?) and does not allow for interconnect cables with high(er) capacitance.
With an output impedance of lets say 15k, you will have a -3dB roll of at 40kHz with a cable with 265pF. Not unusual for interconnect cables
Question, I believe this tube stage has little to do with your DAC concept as such. Why not open a thread at the tube group? Sure you will get some more feedback there
What you publish here as Tube Stage might be kind of OK, but has imo nothing to do with "Ultimate". Your multiple DAC idea comes much closer to that statement I believe. Even though I don't like opams, but lets leave that discussion for a while... 😉
doede
Hi ecdesigns,
If you are still using ecc82 tubes on your dac output, please note that these require much more current than provided in your schematics. For ecc82 the current should be between 5-10 mA per triode. That would lower the gain also. And there really is no sense to have active anode loads together with a ccs on the differential tail. Simple resistor anode loads are sufficient (also on the CF). And as some others have already stated, ecc82 is probably not the optimal tube for this purpose.
If you are still using ecc82 tubes on your dac output, please note that these require much more current than provided in your schematics. For ecc82 the current should be between 5-10 mA per triode. That would lower the gain also. And there really is no sense to have active anode loads together with a ccs on the differential tail. Simple resistor anode loads are sufficient (also on the CF). And as some others have already stated, ecc82 is probably not the optimal tube for this purpose.
Hi ecdesigns,
I have a friend who designs and sells his own tube equipment. He's been designing tube amps/preamps since the '70s. I will ask his advice on improving the tube output section.
I have a friend who designs and sells his own tube equipment. He's been designing tube amps/preamps since the '70s. I will ask his advice on improving the tube output section.
Re: Direct Interpolation DAC circuit boards
tubes are good at amplifying voltage. I am just not sure about going up to 15Vpk-pk and then drop the voltage with resistors and then re-amplify the just reduced signal to 2VRMS. Maybe you can try to combine both the op624iv and the tube linestage with one tube stage?
http://www.tubecad.com/march2001/2001_03.pdf
It is just a suggestion for something to try, as I think your parallel dac idea is just incredible.
Another option (especially for tube guys that already have a lot of mercury and deadly voltages in the system) is to maybe forget the tube stage altogether. Take a look at the Pass D-1 and use mosfet (or power jfet) to do the iv. You will get very clean current gain, and low ouput impedance.
tubes are good at amplifying voltage. I am just not sure about going up to 15Vpk-pk and then drop the voltage with resistors and then re-amplify the just reduced signal to 2VRMS. Maybe you can try to combine both the op624iv and the tube linestage with one tube stage?
http://www.tubecad.com/march2001/2001_03.pdf
It is just a suggestion for something to try, as I think your parallel dac idea is just incredible.
Another option (especially for tube guys that already have a lot of mercury and deadly voltages in the system) is to maybe forget the tube stage altogether. Take a look at the Pass D-1 and use mosfet (or power jfet) to do the iv. You will get very clean current gain, and low ouput impedance.
This is a great thread and very D-I-DAC-tic 😀 
Keep the good work, ecdesigns! Forza!
I hope someday there will be a kit
or a groupbuy 😉 with your aprouval...
Good luck!
M
Keep the good work, ecdesigns! Forza!
I hope someday there will be a kit
or a groupbuy 😉 with your aprouval...Good luck!
M
Hi ecdesigns
Somewhere in this thread you envited Tubee for a listening session! Does this also apply for other interested Dutch guys? I would like it very much to listen to the DAC - actually the whole chain. Also, I will already start looking for places that still sells the TDA1541A.
Erik
Somewhere in this thread you envited Tubee for a listening session! Does this also apply for other interested Dutch guys? I would like it very much to listen to the DAC - actually the whole chain. Also, I will already start looking for places that still sells the TDA1541A.
Erik
Pass D1. I found this on another thread
http://www.diyaudio.com/forums/showthread.php?postid=927855#post927855
http://www.diyaudio.com/forums/showthread.php?postid=927855#post927855