ah, thaanks, I didn't got he tested the AD8597 in lieu of the D77. Going to re read...
As usual I had great pleasure to re read Fran's testimonies about his DACs stages.
I hadn't noticed it was indeed a passive I/V with gain after in lieu of the conventional inverted input with virtual said ground.
Btw, is the virtual ground typology not better about having a bigger and flatter bandwidth ? At least the (only, with good noise figure as well) good side of the opa's evil for this transimpedance task ?
Had someone the chance to experiment both with a same opa like the AD777 for illustration ? Anyway not hard to try. I can understand you use passive resistor to ground with tubes for instance...
@EUVL , curious to have your subjective opinion (sounding) about Sergio's discrete if you may.
I hadn't noticed it was indeed a passive I/V with gain after in lieu of the conventional inverted input with virtual said ground.
Btw, is the virtual ground typology not better about having a bigger and flatter bandwidth ? At least the (only, with good noise figure as well) good side of the opa's evil for this transimpedance task ?
Had someone the chance to experiment both with a same opa like the AD777 for illustration ? Anyway not hard to try. I can understand you use passive resistor to ground with tubes for instance...
@EUVL , curious to have your subjective opinion (sounding) about Sergio's discrete if you may.
Fran told me that we can still support PCBs for the FC CEN, if there is still interest.
You can PM Fran, but he is busy so be patient.
It is unlikely that there is sufficient interest to justify a GB for matched devices.
So you will have to match yourselves.
The AD8429 is indeed less costly than the AD797, and there's no real difference in performance.
See Fran's audition report linked earlier.
So maybe you should try them first.
Advantage of the AD797 is that the PCB is done for you, and you can plug in direct to a DIP8 single opamp footprint.
www.diyaudio.com/community/threads/dac-ad1862-almost-tht-i2s-input-nos-r-2r.354078/post-7319860
Remember you MUST remove Riv and Civ in the normal opamp IV circuit.
Patrick
You can PM Fran, but he is busy so be patient.
It is unlikely that there is sufficient interest to justify a GB for matched devices.
So you will have to match yourselves.
The AD8429 is indeed less costly than the AD797, and there's no real difference in performance.
See Fran's audition report linked earlier.
So maybe you should try them first.
Advantage of the AD797 is that the PCB is done for you, and you can plug in direct to a DIP8 single opamp footprint.
www.diyaudio.com/community/threads/dac-ad1862-almost-tht-i2s-input-nos-r-2r.354078/post-7319860
Remember you MUST remove Riv and Civ in the normal opamp IV circuit.
Patrick
Last edited:
Hi,
Not a big fan of direct passive I/V with Gain after with semi-conductors.
Hummm, I found some genuine k170/J74 pairs, but how much of the two Mosfet one should buy to have a match according your experience, please ?
What finally was the subjective listening feeling of the Sergio's discrete VS the FC CEN if you had time to build it ?
Thanks
Not a big fan of direct passive I/V with Gain after with semi-conductors.
Hummm, I found some genuine k170/J74 pairs, but how much of the two Mosfet one should buy to have a match according your experience, please ?
What finally was the subjective listening feeling of the Sergio's discrete VS the FC CEN if you had time to build it ?
Thanks
Hi @diyiggy
K170/j74 - not being glib, but the bigger the number the more matches you will get. You might be better off just buying some matched sets if you don't have enough to match.
Sergio's discrete is still on the workbench - the delay here is just getting time to match devices more than the build.
K170/j74 - not being glib, but the bigger the number the more matches you will get. You might be better off just buying some matched sets if you don't have enough to match.
Sergio's discrete is still on the workbench - the delay here is just getting time to match devices more than the build.
Hello Fran, thank s for your feed-back.
Do you have some to sell left, please ? I have first to check if my K170/J74 are paired enough.
Sergio's sim had crazy good THD number without matching in his sim. You had bad result in real world since populated ?
I tougth of try it with non matched KSA with a little higher Hfe and even quieter than the BC559 & its sibbling...
cheers
Do you have some to sell left, please ? I have first to check if my K170/J74 are paired enough.
Sergio's sim had crazy good THD number without matching in his sim. You had bad result in real world since populated ?
I tougth of try it with non matched KSA with a little higher Hfe and even quieter than the BC559 & its sibbling...
cheers
Don't forget that in the sim, devices are alwyas perfectly matched - they use the same model so are 100% identical.
You can un-match sim devices by hand by duplicating models and slightly changing some values but that's a sciese in itself.
Jan
Thanks
I didn't know that. I was said in some shematic the unmatching was mostly increase H2 and thermal issues.
The only member who construct it, I am aware about is @bohrok2610
I didn't know that. I was said in some shematic the unmatching was mostly increase H2 and thermal issues.
The only member who construct it, I am aware about is @bohrok2610
It depends on the particular schematic. If you have a balanced or push-pull stage, perfect matching minimizes 2nd harmonic distortion.
So in a sim, where matching is perfect because devices are identical, you get unrealistic low (2nd harmonic) distortion.
Jan
So in a sim, where matching is perfect because devices are identical, you get unrealistic low (2nd harmonic) distortion.
Jan
It has been measured :
www.diyaudio.com/community/threads/distortion-measurements-of-a-collection-of-iv-converters.415094/
But measurements don't always say anything about subjective impressions.
Else we only have composite opamps or Class D, and nothing else.
Fortunately, DIY is more fun than that. 😊
Patrick
www.diyaudio.com/community/threads/distortion-measurements-of-a-collection-of-iv-converters.415094/
But measurements don't always say anything about subjective impressions.
Else we only have composite opamps or Class D, and nothing else.
Fortunately, DIY is more fun than that. 😊
Patrick
Is it that one ? Was made for aan ESS9018, Maybe it is less good with a 1 to 2 mA output dac chip only ?
Hummm, should be this one : I wonder how much is the input impedance, must re read the thread : https://www.diyaudio.com/community/...rtion-with-very-low-distortion.217459/page-47
Do you think according to your knowledge, those BC337-40 / 327-40 could be swapped by transistor with same noise figurebut twice gain for the weaker dac chip that output 1 or 2 mA, please ?
I have in mind the KSC1845/KSA970, I assume the resistors had to be changed and a sim remade?
I have in mind the KSC1845/KSA970, I assume the resistors had to be changed and a sim remade?
Questions for Sergio's circuit should be addressed to him, at his thread :
www.diyaudio.com/community/threads/dac-i-v-convertion-with-very-low-distortion.217459/
Patrick
www.diyaudio.com/community/threads/dac-i-v-convertion-with-very-low-distortion.217459/
Patrick
Having had these sitting in a box for too long, I finally got around to stuffing the boards.
Was able to trim the output offset on the first board to a few mV and the input offset is less than 1mV. On the second board, the output offset was also trimmed to a few mV but there is -10mV offset on the input. I guess it may down to a poorly matched pair of jfet's. Component values are all standard/according to the instructions. In any event, could anyone suggest if/how the input offset may be trimmed (R58?). Of course, this may be a non-issue for the DAC chip and only an OCD issue😊
Cheers, Soren
Was able to trim the output offset on the first board to a few mV and the input offset is less than 1mV. On the second board, the output offset was also trimmed to a few mV but there is -10mV offset on the input. I guess it may down to a poorly matched pair of jfet's. Component values are all standard/according to the instructions. In any event, could anyone suggest if/how the input offset may be trimmed (R58?). Of course, this may be a non-issue for the DAC chip and only an OCD issue😊
Cheers, Soren
Last edited:
@EUVL, @woodturner-fran,
please, do you think the Bjt pairs can be swapped by the smd : HA101F-GR & HN1C01F-GR ? or the heat is too much for smd here ?
Thanks
please, do you think the Bjt pairs can be swapped by the smd : HA101F-GR & HN1C01F-GR ? or the heat is too much for smd here ?
Thanks
Got this sorted out and have been listening to it this afternoon. I am sure I am going to like this one 😊
Thanks for sharing this, EUVL.
New pair of FC Cen: I have input offset of 8mV on one channel and started trouble shooting. These are built following the BOM/boards from @woodturner-fran and using a P+N quad with idss=8.5 and an N-quad with idss=9 (as reported by punkydawg). The output offset was easily set to a few mV for each channel but measuring the voltages on the boards show some curious differences. Maybe it is insufficient current through the mirror on one board, as @EUVL outlined in post #80? Any suggestions?
Both boards produce music.
Both boards produce music.
