Have you checked for oscillations. In series resistor at input is usually needed for that opamp, also hows the heat?
When you wire the power supply to the dac make sure the correct voltage goes to the right location. The picture makes it look like the wires just go straight but in fact they cross underneath
Not yet. The heat is ok since supply is +/-5. Total DAC power consumption is 4W (including USB->I2S and simple headphone amp).
But only today I realized that there is no reconstruction filter on this board...
Yeah, this is a NOS filterless dac. You can check @abraxalito dark I/V stage. Nice engineering magic there.
Anyone else that's powering the opamp section of the DAC separately?
I figured it's just to leave the jumpers around the opamps un-populated and wire +/-12Vdc(in my case) to the jumper pads around the opamps. Please let me know if I'm wrong. Currently getting ready to start power supplies for the PCM1702 board. After that I'll complete one of the PCM63 DAC's (either the one with the I2SoverUSB footprint or the standard one).
I figured it's just to leave the jumpers around the opamps un-populated and wire +/-12Vdc(in my case) to the jumper pads around the opamps. Please let me know if I'm wrong. Currently getting ready to start power supplies for the PCM1702 board. After that I'll complete one of the PCM63 DAC's (either the one with the I2SoverUSB footprint or the standard one).
Brijac is right, there should be 1K on the inverting input of the AD811. Otherwise you risk problems with the AD811. Try increasing the supply voltage(then don't forget the heatsink), +/-5V is the lower limit for the AD811 and increasing the voltage improves performance of op amp. Power the AD811 with separate regulators that have a separate secondary or their own transformer for the IV stage.
I hope I helped a little.
Thanks, actually I'm considering development of my own DAC board with separate higher voltage supply for out stage.
Regarding the oscillations - that's interesting, I'll look into it and do couple of experiments. I already know DAC designs w/o such thing. There are small ceramic cap at the DAC output parallel to larger film cap, that's it:
Is this the new Double Crown version of AD1862?? 🤣
https://www.ebay.com/itm/2051498611...sGheTUqSyi&var=&widget_ver=artemis&media=COPY
https://www.ebay.com/itm/2051498611...sGheTUqSyi&var=&widget_ver=artemis&media=COPY
Those forges are reder than the comunism. LOL!
AD811 at 10/12V seems more quiet temperature wise, at least the pdip version,no one had the balls to try it soic, lol ! Notice the pdip socket must not help for osci as it is a fast op amp. I don't know if one checked that with an osci !
The tantalum resistor are well know to smooth the sound, so it is perhaps welcome at the input of the AD811 for PCM1702, PCM56/58/63.... maybe less for an AD1865. AD1862 certainlly stands certainly in between.
So no one has tried the LT1028, yet ? It as pdip package as well (though maybe NLA inthat size).
Strange noone really work on the op828 as a competitor of a standalone I/V buffer chip too VS the op1655 !
op1633, anyone ?
AD811 at 10/12V seems more quiet temperature wise, at least the pdip version,no one had the balls to try it soic, lol ! Notice the pdip socket must not help for osci as it is a fast op amp. I don't know if one checked that with an osci !
The tantalum resistor are well know to smooth the sound, so it is perhaps welcome at the input of the AD811 for PCM1702, PCM56/58/63.... maybe less for an AD1865. AD1862 certainlly stands certainly in between.
So no one has tried the LT1028, yet ? It as pdip package as well (though maybe NLA inthat size).
Strange noone really work on the op828 as a competitor of a standalone I/V buffer chip too VS the op1655 !
op1633, anyone ?
I've got LT1028 in dip and soic package.
I like them as they are very low noise.
I've also got LME49713HA, LME49710HA, OPA827(IIRC), AD797, OPA627 and a bunch of others I don't recall of the top of my head.
I like them as they are very low noise.
I've also got LME49713HA, LME49710HA, OPA827(IIRC), AD797, OPA627 and a bunch of others I don't recall of the top of my head.
I'm late to the party here and only have a very basic PCM56 stage connected so far. More to come on that. But what I can contribute with my 20 year experience of building valve amps and stages, is a couple of alternative valve output stages to consider. These are attached. The choice of Kemet DC Link cathode bypasses and FT-3 teflon coupling caps is the result of a lot of experiments and these are important for the sound quality. Of course caps are a matter of taste and others may have their favourites. There is no step-down on the output because this feeds a valve stage with 100K grid resistor. The valves chosen have a Rp of 10K or 11K so that needs to be factored in. I chose ECC40 and E80CC for this because I had already used them in other contexts and they sounded better than the alternatives I tried. Both have a gain of around x30. I'll give 6N1P a whirl in this at some point - just have to move a couple of wires on the heater connections. Maybe ECC88 if I can find one.
A word on the topology. I have tried the usual alternatives like active loads and SIC diodes in the cathode and rejected them because they added an edge to the sound, however slight. Resistors I always find to be the most natural sound. Like with caps you can experiment with good sounding ones. The anode resistor needs to be close to 5x the Rp of the valve for good sound, so 47K was chosen.
The E80CC and ECC40 were auditioned on a range of acoustic music - opera, orchestral, jazz piano trio (Bill Evans "I Will Say Goodbye") and jazz vocals (Dee Dee Bridgewater "Keeping Tradition"). I don't listen to rock music, sorry! Both were very good. I think ECC40 just edged it on piano tone, which is very natural. It was widely used in European valve equipment in its day. This is a rimlock base valve but bases are available and the cutout is 22mm like for 9 pin valves. These valves are a bit rarer and more expensive than ECC88s but in my view worth it. For E80CC I prefer the Tungsram with the silver pins. Others like the gold pin versions. Andy
A word on the topology. I have tried the usual alternatives like active loads and SIC diodes in the cathode and rejected them because they added an edge to the sound, however slight. Resistors I always find to be the most natural sound. Like with caps you can experiment with good sounding ones. The anode resistor needs to be close to 5x the Rp of the valve for good sound, so 47K was chosen.
The E80CC and ECC40 were auditioned on a range of acoustic music - opera, orchestral, jazz piano trio (Bill Evans "I Will Say Goodbye") and jazz vocals (Dee Dee Bridgewater "Keeping Tradition"). I don't listen to rock music, sorry! Both were very good. I think ECC40 just edged it on piano tone, which is very natural. It was widely used in European valve equipment in its day. This is a rimlock base valve but bases are available and the cutout is 22mm like for 9 pin valves. These valves are a bit rarer and more expensive than ECC88s but in my view worth it. For E80CC I prefer the Tungsram with the silver pins. Others like the gold pin versions. Andy
Attachments
Last edited:
Are you suggesting these as I/V stages? I think the PCM56 (and the other DACs discussed here) need a very low input impedance of the I/V stage (like 30-50 Ohms or so).
Yes - I/V stages. The 30-50 ohms, 100ohms or whatever would be in parallel with the 100K input resistor.
Last edited:
Yes it's CFA, I don't know how it compares to the AD811 as I have no experience with that opamp
Thank you for sharing your experience. I am intrigued by ECC40. Is the output impedance going to be too high for a typical preamp? Also, I am curious if you experimented with a transformer output instead of a capacitor.
The output impedance of 11K would need to see an input impedance of around 100K. The 1 to 10 rule for connecting stages. I haven't used an output transformer because I don't need to in my setup since it feeds a valve stage. The ECC40 would work into any valve preamp with an input resistor of 100K. But not much less. It wouldn't be suitable for a solid state preamp without a cathode follower tube. If you used a step-down transformer to lower the output impedance you would lose the gain. For a 100 ohm I/V resistor you need a tube with a gain of around 28 and if you stepped that down you would need a tube with a higher gain.