Guys... I'm so sorry, I haven't logged into my account for over two years.
Good job on making new drawings of the schematic and the board layout. I remember vaguely that the choices in grounding came from reading Guido Tent's whitepaper on grounding. Therefore I did not use split ground planes. I did a double ground plane in separate layers enclosing the signal layers, connected with multiple vias. So every ground is connected directly to a thick ground plane, and I just made sure that the power, digital and signal grounds were tidy.
When it comes to the I/V. I can agree that it's a simple one. But it definitely does not sound like the NE5534. At all. But it would be interesting to compare it directly to an OP stage.
I still use this dac in my main setup, and I listen to it every day.
Good job on making new drawings of the schematic and the board layout. I remember vaguely that the choices in grounding came from reading Guido Tent's whitepaper on grounding. Therefore I did not use split ground planes. I did a double ground plane in separate layers enclosing the signal layers, connected with multiple vias. So every ground is connected directly to a thick ground plane, and I just made sure that the power, digital and signal grounds were tidy.
When it comes to the I/V. I can agree that it's a simple one. But it definitely does not sound like the NE5534. At all. But it would be interesting to compare it directly to an OP stage.
I still use this dac in my main setup, and I listen to it every day.
Finally I completed my AD1862 DAC with the IV stage as shown in #36:
It's dual mono, one channel is on the lower level of PCBs, the other channel is upstairs. There are many jung style super reulators (ALWSR), it sounds superb. The digital filter is the Kusy Audio 16x interpolation filter. Separate transformers for analog (large toroid) and digital (blue Talema). The third (smallest blue) transformer is for a housekeeping mcu that observes power loss and digital error and controls a mute relay.
I have one question: do I need to worry about the IV resistors value at the end of the diamond buffer? I now use 3K7, but iirc lower values will result in less distortion - but only if the DAC chip itself drives the IV resistor, right...?
It's dual mono, one channel is on the lower level of PCBs, the other channel is upstairs. There are many jung style super reulators (ALWSR), it sounds superb. The digital filter is the Kusy Audio 16x interpolation filter. Separate transformers for analog (large toroid) and digital (blue Talema). The third (smallest blue) transformer is for a housekeeping mcu that observes power loss and digital error and controls a mute relay.
I have one question: do I need to worry about the IV resistors value at the end of the diamond buffer? I now use 3K7, but iirc lower values will result in less distortion - but only if the DAC chip itself drives the IV resistor, right...?
Looks absolutely fantastic 🙏 I can imagine that it sounds lovely. I will try to compare my dac with some more modern AK and ESS dacs soon. I am hoping it holds up, but I really have no idea.
As for the IV resistor I have no idea. The original design from Pedja Rogic uses a 1.5k resistor. But I don't know how much it impacts the sound. The only way to know is to switch and compare. Putting a 1.5k there shouldn't impact the gain too much. I use 300Ohm and it's still fine.
As for the IV resistor I have no idea. The original design from Pedja Rogic uses a 1.5k resistor. But I don't know how much it impacts the sound. The only way to know is to switch and compare. Putting a 1.5k there shouldn't impact the gain too much. I use 300Ohm and it's still fine.
Thanks for your answer! How much output voltage do you get with 300 Ohms? I am getting 5.6Vp-p with the 3K7 resistor if the DAC is connected to my preamp. If my calculations are correct with 1K5 and the TDA1541A that outputs 4mA one would get 6Vp-p.
I finally upgraded the IV stage (inspired by this thread). The sound cleared up considerabely. I consider my DAC project completed 😎
Yes, this is the DAC just before fitting the new IV stage:
I have added separate PSUs for the USB interface (made a huge difference that I never expected), an output mute relay, and a loss of AC detector with a housekeeping mcu.
I have added separate PSUs for the USB interface (made a huge difference that I never expected), an output mute relay, and a loss of AC detector with a housekeeping mcu.
Hello,
So the change is the output filtering of AD1862 with the shunt cap ? Which value did you prefer in your setup, please? Wima KP2 ?
Yeah, USB deserves cares and own input as output supplies if galvanic isolation between them.
RN60 resitor for the I/V I see ? You could swap this polystyren for a tin foil LCR one (Farnel), according the resistor you migth hear an improvment)
thanks
So the change is the output filtering of AD1862 with the shunt cap ? Which value did you prefer in your setup, please? Wima KP2 ?
Yeah, USB deserves cares and own input as output supplies if galvanic isolation between them.
RN60 resitor for the I/V I see ? You could swap this polystyren for a tin foil LCR one (Farnel), according the resistor you migth hear an improvment)
thanks
Thanks for the hints. The change was adding current mirrors to the IV stage. The filter cap was behind the IV stage, now I have it in front of. I am using a 330pF sytrene. I didn't try other caps, I am not a boutique component tester kind of guy ;-) But I know styrenes are fine, as are RN60s. Tin foil with 330pF~470pF is available...?
Last edited:
I don't remember but I assume as I use 1 nF Styren tin foil LCR 2% (a little less zingy). Pedja is using normal one as the one you have. RN60 is fine. I use Ayrton Perry type wirewound one from RHopoint, but a carbon comp is also okay to shunt current fastly (that's what Pedja is using) over metal film. I do prefer carbon type as well to my ears at least. I think it should haveto see with one's own whole setup at the end.
I just noticed that there is some DC offset at the output of the IV stages: one channel has 13mV and the other channel 23mV. I guess the reason is I didn't match the transistors... Is such an amount of DC offset negligible?
I wonder if it is worthing the try to use at input only a matched double Jfet in spite of the BJT for this AD1862 discrete I/V?
2K output impedance and 1 mA maybe makes it okay enough for BJT inputs very low impedance side as you did (emitters) ?
Sommeone told me Pedja used toshiba discrete for the AYA5 with base input, but I don't now about the base impedance of those Toshibas SMD BJT and if there is a need if so of a low Rbb when the load is relativly high impedance (2K for AD1862) ?
If keeping the BC550/560 combo, worthing a try to double diode piggyback to ground to help on the spikes or a RC between te ad18622 and the I/V ?
2K output impedance and 1 mA maybe makes it okay enough for BJT inputs very low impedance side as you did (emitters) ?
Sommeone told me Pedja used toshiba discrete for the AYA5 with base input, but I don't now about the base impedance of those Toshibas SMD BJT and if there is a need if so of a low Rbb when the load is relativly high impedance (2K for AD1862) ?
If keeping the BC550/560 combo, worthing a try to double diode piggyback to ground to help on the spikes or a RC between te ad18622 and the I/V ?
Last edited:
Hi Diyiggy,
This is the IV stage I ended up with.
This is the IV stage I ended up with.