Thinking of building my own DAC box. What DAC chips do you think would be good choices? I would like to connect the DAC chips' outputs directly to a tube stage (say an I to V converter resistor feeding a triode SRPP maybe) to drive the analog output jacks. Want to avoid op-amps in the chain. Another issue I would see is the clock reconstruction PLL circuit for SPDIF. Does TOSLINK also require this? It would be nice if the clock was provided directly so I wouldn't have to deal with PLL jitter issues.
You have a fairly wide choice of chipsets. Latest generation of 24-bit DAC chips are all Delta-Sigma designs, with PCM1792, AD1955, and SM5865C being top-spec models from Burr-Brown, Analog Devices and NPC correspondingly. All of these are single-supply and have differential current output with respect to mid-rail virtual ground. Of the traditional R2R ladder audio DACs, the only one still around is PCM1704. A lot of people on this board like old 16-bit chips like TDA1541 connected without digital interpolation filters (non-oversampled).
Digital interface receiver chips (like DIR1703) typically have clock-recovery PLLs built-in, but offer essentially no jitter attenuation in audio frequency band. Additional VCXO-based low-frequency PLL can be used to improve jitter rejection in audio band, but this is more complicated and is rarely seen in amateur designs. Possibly interesting one-chip solution for clock clean-up is CY23FS04 from Cypress Semiconductors, which is a digitally-controlled XO and clock buffer.
As for coupling DAC current output to tube output stages, I think the best way to go is a low-impedance step-up transformer. This gives you low-impedance load for DAC output, high signal input for tube stage, and passive third order RC-L-RC filter all in one 🙂.
Personally, I have designed and build DACs around PCM1730 and NPC5865. (I also used a bunch of commercial DACs, including tubed one from Audio Note...)
Digital interface receiver chips (like DIR1703) typically have clock-recovery PLLs built-in, but offer essentially no jitter attenuation in audio frequency band. Additional VCXO-based low-frequency PLL can be used to improve jitter rejection in audio band, but this is more complicated and is rarely seen in amateur designs. Possibly interesting one-chip solution for clock clean-up is CY23FS04 from Cypress Semiconductors, which is a digitally-controlled XO and clock buffer.
As for coupling DAC current output to tube output stages, I think the best way to go is a low-impedance step-up transformer. This gives you low-impedance load for DAC output, high signal input for tube stage, and passive third order RC-L-RC filter all in one 🙂.
Personally, I have designed and build DACs around PCM1730 and NPC5865. (I also used a bunch of commercial DACs, including tubed one from Audio Note...)
Well, I think there are a lot of different opinions about the best choice for a DAC. I'd prefer chips in production, and it is sometimes stated the delta sigma are not the best. The TDA1541 is a NOS type, as I know the only R-2R DACs available are from TI. The best choise to me seems the PCM1704 but they are expensive when using the highest grade types.
The same with the digital receiver. AKM and Crystal are actually in production. A CS8416 will do the job but care must taken when SPDIF signal is fed in.
I don't like the passive I/V stage concepts regardless if it is a transformer or a simple resistor. A well designed discrete I/V sounds better to me. But this can be stange when using a tube output stage. I think you dont want to have some sand into your DAC box...
PLL is not easy to implement. I agree that the only possible way is using a VCXO, designing a loop filter, buffer the signals with some registers and so on.
The same with the digital receiver. AKM and Crystal are actually in production. A CS8416 will do the job but care must taken when SPDIF signal is fed in.
I don't like the passive I/V stage concepts regardless if it is a transformer or a simple resistor. A well designed discrete I/V sounds better to me. But this can be stange when using a tube output stage. I think you dont want to have some sand into your DAC box...
PLL is not easy to implement. I agree that the only possible way is using a VCXO, designing a loop filter, buffer the signals with some registers and so on.
Most important is to keep the whole thing as simple as possible🙂 .A CS8414+CS4334 combination wont be a bad idea(sound.westhost.com)
Check out this
I built this dac a few years ago, Its impressive. Ive got the tube output boards but still have not messed with them
Tom
Jukka Tolonen's http://koti.mbnet.fi/siliconf/JukkaTolonen/projects/projects.html
http://www.audiodesignguide.com/cdplayer/
http://www.audiocraftersguild.com/AandE/articles.htm
I built this dac a few years ago, Its impressive. Ive got the tube output boards but still have not messed with them
Tom
Jukka Tolonen's http://koti.mbnet.fi/siliconf/JukkaTolonen/projects/projects.html
http://www.audiodesignguide.com/cdplayer/
http://www.audiocraftersguild.com/AandE/articles.htm
- Status
- Not open for further replies.