I currently have a RAKK DAC mk 1 with USB input from my Mac Mini. This goes directly into the grids of my tube line stage.
I'd like to upgrade the DAC board. What can I use that I can also output directly to a tube stage, bypassing any solid state output device?
I like the ESS range - can this be done with any of them, e.g. ODAC etc?
I'd like to upgrade the DAC board. What can I use that I can also output directly to a tube stage, bypassing any solid state output device?
I like the ESS range - can this be done with any of them, e.g. ODAC etc?
Indeed, but there has to be some point on the board to take the signal out, and that might not be that simple.
You would take the signal directly off the DAC chip. For the bigger DAC chips you can just buy the chip and use point to point wiring. The small SMD ones you may need to get a PCB.
My favorite DAC chip is the AD1865, and it is big, so that is my recommendation. I also like PCM1794, AD1852, and the ESS chips.
This has been my experience for the best DAC sound. But there are two camps:
Depending on the DAC the I/V resistor can be sized differently. But in general the best I/V converter is a passive I/V followed by a phonotype preamp gain + filter + buffer. You also want to parallel the I/V resistor with a small cap to knock out the HF crud as early as possible.
Most of my experience is with the PCM 1704, here I use a 30 ohm I/V+cap followed by a classic Jfet gain (with NFB 36V rails) then filtering then diamond buffer, it measures -and sounds better than any IC opamp I have tried (2H -96 down.)
Using tubes it is "tougher" as the PCM1704 loses performance and sounds harsh with much lower (higher I/V resistor) than 30 ohms. For best high end performance you want to think of the design as a phono type preamp gain/noise structure a gain over 100 is ideal. But as with everything there are a few exceptional DAC chips that allow a large I/V resistor without voltage compliance issues.
The PCM179x series, the AD1865, the PCM56, and the ESS, are four DAC off hand that allow a large enough I/V resistor to use a tube gain without losing 16 bit SNR ratio. Calculate the LSB and you will see what I mean. I don't think true hirez is feasable with tubes, but they do a nice job of analog dithering to 14-16 bit.
Check out audio design guides latest AD1865 and ESS designs for the line-level type passive DAC analog stages.
Depending on the DAC the I/V resistor can be sized differently. But in general the best I/V converter is a passive I/V followed by a phonotype preamp gain + filter + buffer. You also want to parallel the I/V resistor with a small cap to knock out the HF crud as early as possible.
Most of my experience is with the PCM 1704, here I use a 30 ohm I/V+cap followed by a classic Jfet gain (with NFB 36V rails) then filtering then diamond buffer, it measures -and sounds better than any IC opamp I have tried (2H -96 down.)
Using tubes it is "tougher" as the PCM1704 loses performance and sounds harsh with much lower (higher I/V resistor) than 30 ohms. For best high end performance you want to think of the design as a phono type preamp gain/noise structure a gain over 100 is ideal. But as with everything there are a few exceptional DAC chips that allow a large I/V resistor without voltage compliance issues.
The PCM179x series, the AD1865, the PCM56, and the ESS, are four DAC off hand that allow a large enough I/V resistor to use a tube gain without losing 16 bit SNR ratio. Calculate the LSB and you will see what I mean. I don't think true hirez is feasable with tubes, but they do a nice job of analog dithering to 14-16 bit.
Check out audio design guides latest AD1865 and ESS designs for the line-level type passive DAC analog stages.
I've used Voltage out DAC chips from ESS, Cirrus, TI, AKM into a transformer or cap. I like transformer better.
Since most of the voltage output DACs have a low output impedance and don't mind (some prefer) driving a low impedance load, you could use a step up transfo to hit the grids directly. You'd want to know that input impedance of your tube, of course.
Volume control could be done between the transfo secondary and the grid, or use digital volume control if you are comfortable with that.
Since most of the voltage output DACs have a low output impedance and don't mind (some prefer) driving a low impedance load, you could use a step up transfo to hit the grids directly. You'd want to know that input impedance of your tube, of course.
Volume control could be done between the transfo secondary and the grid, or use digital volume control if you are comfortable with that.
Hi Pano,
Out of all the DACs you've used, what would you recommend for a build right now? I'd need a built board - can't do surface mount.
Andy
Out of all the DACs you've used, what would you recommend for a build right now? I'd need a built board - can't do surface mount.
Andy
I have to say that the good old CS4396 in hardware mode plays very well with a transformer output. I know it's not the latest, shiniest darling of the DIY crowd, but it never failed to please at audio shows where I took it or sent it. {EDIT, fixed dac part number}
I find the ESS chips a bit more refined and subtle, but they lack the dynamics and strength that the Cirrus chip seems to bring. A matter of taste. Overwhelmingly listeners choose the transformer coupled CS8416 when I let them listen to it.
If you don't do surface mount (I understand) look for a board with the chip already in place. A good power supply and a transformer output will make just about anyone happy.
I find the ESS chips a bit more refined and subtle, but they lack the dynamics and strength that the Cirrus chip seems to bring. A matter of taste. Overwhelmingly listeners choose the transformer coupled CS8416 when I let them listen to it.
If you don't do surface mount (I understand) look for a board with the chip already in place. A good power supply and a transformer output will make just about anyone happy.
pano experience with cs4398 ? play iron and native dsd
Ps
pcm data need e different filter then DSD in the out ,another point to thinks about
Ps
pcm data need e different filter then DSD in the out ,another point to thinks about
Probably. I'll have to go back and check my notes. I've used so many of these that the part numbers get confusing.
You are correct. Oops! Too many part numbers.
I used the CS4396 chip, CS8416 was the receiver. I don't think I've tried the CS4398.
I used the CS4396 chip, CS8416 was the receiver. I don't think I've tried the CS4398.
Iout into a fixed voltage, wouldn't driving the cathode make more sense?
For example: http://www.tubecad.com/2006/07/blog0072.htm
For example: http://www.tubecad.com/2006/07/blog0072.htm
Last edited:
- Status
- Not open for further replies.