[L0rdGwyn]

Finished both channels of my PCM63 DAC and the digital front end power supply board.



The PCM63 is the Holy Grail of vintage Burr-Brown R2R chips. Each board uses two of them in parallel, the two boards will stack on top of each other, similar to my PCM56 DAC. The output of each board are the current output pins of the parallel chips. I will again be using a Sowter step up transformer for the I/V stage.

I am again planning a tube output stage, but a different flavor this time around. I have a healthy stock of NOS E182CC tubes, a very good sounding audio dual triode. Each section has a relatively high gm of 15mA/V. I'd like to use a pair of E182CC on the output with the sections wired in parallel in a transformer coupled cathode follower output buffer with a 1:1 output transformer, something like this.



Being completely transformer coupled and with a triode well suited for buffer service, I am expecting very good results with this setup. However, since a cathode follower has no voltage gain, the Sowter step up transformers will be responsible for converting the current output of the DAC to a line level output voltage.

Without getting too heavy into the technical details, these current output DACs perform best if presented with a low impedance load. Generally, less than 100ohms is best for PCM63, otherwise it will start generating audible harmonic distortion. To reach 2Vrms output with a 1:10 step up and parallel PCM63, roughly an 8.2K resistor will be needed on the secondary of the step up transformer, presenting the DAC chips with an ~82ohm primary impedance.

This should work well in theory, but I plan to measure the distortion on the secondary of the step up transformers with a dummy load before committing to this I/V stage. If the measurements look good, then I'll order parts for the output buffer and design the chassis. This time around, I plan to include some tube aesthetics with the E182CC buffer tubes sticking partway out of the top plate, should look pretty cool.

I have next week off before starting my new job, going to test this and also devote some time to the Aegis project.

 

[heliosphann]

Finished both channels of my PCM63 DAC and the digital front end power supply board.



The PCM63 is the Holy Grail of vintage Philips R2R chips. Each board uses two of them in parallel, the two boards will stack on top of each other, similar to my PCM56 DAC. The output of each board are the current output pins of the parallel chips. I will again be using a Sowter step up transformer for the I/V stage.

I am again planning a tube output stage, but a different flavor this time around. I have a healthy stock of NOS E182CC tubes, a very good sounding audio dual triode. Each section has a relatively high gm of 15mA/V. I'd like to use a pair of E182CC on the output with the sections wired in parallel in a transformer coupled cathode follower output buffer with a 1:1 output transformer, something like this.



Being completely transformer coupled and with a triode well suited for buffer service, I am expecting very good results with this setup. However, since a cathode follower has no voltage gain, the Sowter step up transformers will be responsible for converting the current output of the DAC to a line level output voltage.

Without getting too heavy into the technical details, the current output pins of these DACs perform best if presented with a low impedance. Generally, less than 100ohms is best for PCM63, otherwise it will start generating audible harmonic distortion. To reach 2Vrms output with a 1:10 step up and parallel PCM63, roughly an 8.2K resistor will be needed on the secondary of the step up transformer, presenting the DAC chips with an ~82ohm primary impedance.

This should work well in theory, but I plan to measure the distortion on the secondary of the step up transformers with a dummy load before committing to this I/V stage. If the measurements look good, then I'll order parts for the output buffer and design the chassis. This time around, I plan to include some tube aesthetics with the E182CC buffer tubes sticking partway out of the top plate, should look pretty cool.

I have next week off before starting my new job, going to test this and also devote some time to the Aegis project.

Pictures pretty. Rest makes my head hurt.

 

[L0rdGwyn]

Pictures pretty. Rest makes my head hurt.

Trust me, it makes my head hurt too lol. Designing a DAC, even a relatively simple one, is no cakewalk. But I did most of the brain work on my PCM56 DAC, so I've been able to lean on that work for the second build. Ultimately, I wanted to build something with the PCM63, but given they run around $150 a chip, needed to be confident in my design before committing to it, and I am very happy with my PCM56 build.

 

[jgwtriode]

How will this one compare to the one you brought with you to the last dealer/enthusiast meet and put up against the DCS! That was
a very impressive result btw and perhaps a bit surprising, but not really knowing how incredibly well designed your projects are and
your knack for making things sound amazing!

jgwtriode

 

[L0rdGwyn]

How will this one compare to the one you brought with you to the last dealer/enthusiast meet and put up against the DCS! That was
a very impressive result btw and perhaps a bit surprising, but not really knowing how incredibly well designed your projects are and
your knack for making things sound amazing!

jgwtriode

The digital front end is the same for both, the power supply topology is also the same, however some adjustments have been made to account for the higher current draw of parallel PCM63 vs a single PCM56 per channel. PCM63 is a higher performing chip compared to PCM56, lower distortion and noise, higher theoretical bit depth as well - it is a 20-bit chip vs 16-bit for the PCM56. Both use Sowter step up transformer I/V conversion, however the PCM56 DAC uses a tube voltage gain stage on its output, CCS loaded and capacitor coupled, as such the Sowter I/V transformers are only wired for 1:5 step up ratio. The PCM63 DAC instead will use a 1:10 step up ratio and a transformer coupled tube output buffer as opposed to a voltage gain stage. Given the PCM63 output stage is completely transformer coupled and uses a tube buffer, it should in theory generate less distortion, have better clarity, realism, and excellent low end dynamics.

So basically, the new DAC uses a higher performing DAC chip, two in parallel, and an output stage with a cleaner signal path and lower theoretical distortion. We'll see if the measurements match expectations.

 

[George Hincapie]

The digital front end is the same for both, the power supply topology is also the same, however some adjustments have been made to account for the higher current draw of parallel PCM63 vs a single PCM56 per channel. PCM63 is a higher performing chip compared to PCM56, lower distortion and noise, higher theoretical bit depth as well - it is a 20-bit chip vs 16-bit for the PCM56. Both use Sowter step up transformer I/V conversion, however the PCM56 DAC uses a tube voltage gain stage on its output, CCS loaded and capacitor coupled, as such the Sowter I/V transformers are only wired for 1:5 step up ratio. The PCM63 DAC instead will use a 1:10 step up ratio and a transformer coupled tube output buffer as opposed to a voltage gain stage. Given the PCM63 output stage is completely transformer coupled and uses a tube buffer, it should in theory generate less distortion, have better clarity, realism, and excellent low end dynamics.

So basically, the new DAC uses a higher performing DAC chip, two in parallel, and an output stage with a cleaner signal path and lower theoretical distortion. We'll see if the measurements match expectations.

Forgive my newbie question; is the intention, if all works as you hope, to build some of these DACs to sell, or is this a personal project only?

 

[L0rdGwyn]

Forgive my newbie question; is the intention, if all works as you hope, to build some of these DACs to sell, or is this a personal project only?

Sorry, but this is a personal project only.

 

[George Hincapie]

Sorry, but this is a personal project only.

I wish you the best of luck with it; those boards look like art!

 

[L0rdGwyn]

I scored a pretty sweet deal on some tubes the other day, guy from Italy had a box of 100 NOS D3a, selling them in sets of 10. He accepted my offer for 125 Euro per lot, so I grabbed 30 of them. 12.50 Euro per tube is a crazy good deal for D3a, these are becoming more and more rare and expensive, I was stoked when this listing showed up one of the best audio tubes ever made IMO.

I have an LCR phono stage build on the horizon, these oughta keep me well stocked for a decade or two.

 

[Zachik]

I scored a pretty sweet deal on some tubes the other day, guy from Italy had a box of 100 NOS D3a, selling them in sets of 10. He accepted my offer for 125 Euro per lot, so I grabbed 30 of them. 12.50 Euro per tube is a crazy good deal for D3a

Uh oh... @L0rdGwyn is becoming a @bcowen

 

[L0rdGwyn]

Uh oh... @L0rdGwyn is becoming a @bcowen

Well I only hoard certain tubes, @bcowen hoards all of them

 

[Zachik]

Well I only hoard certain tubes, @bcowen hoards all of them

Actually, @bcowen only hoards GE tubes, telling us they're crap so he can buy them all by the pound!

 

[whirlwind]

I scored a pretty sweet deal on some tubes the other day, guy from Italy had a box of 100 NOS D3a, selling them in sets of 10. He accepted my offer for 125 Euro per lot, so I grabbed 30 of them. 12.50 Euro per tube is a crazy good deal for D3a, these are becoming more and more rare and expensive, I was stoked when this listing showed up one of the best audio tubes ever made IMO.

I have an LCR phono stage build on the horizon, these oughta keep me well stocked for a decade or two.

Great score

 

[pravous]

Well I only hoard certain tubes, @bcowen hoards all of them

You are all too late I have already cornered the @bcowen market for the Aegis market!

 

[bcowen]

Well I only hoard certain tubes, @bcowen hoards all of them

LOL! Except your 'certain' tube hoard collection probably outnumbers my 'all' tube hoard collection by a factor of 10.