But then you're adding active circuitry, which may produce more distortion, and certainly more noise than would a coupling cap.
Your circuit still doesn't eliminate the D.C. issues as connected to the following preamp or amp box. You can see that the virtual-ground (vGND) is still offset above the signal ground. I should think this would cause problems over an unbalanced interconnection. However, an balanced interconnection would seem to work.
Last edited:
Agreed. I just wanted miro1360 to explain a little more about what he was thinking. The schematic he just posted doesn't look like it will solve the problem, if that's what it is intended for.
I think the same about miro's circuit. However, I believe it could be made to work if the signal-grounds were isolated from the saftety-ground (chassis ground) within all unbalanced interconnected boxes. But, then, you might be better off simply utilizing a coupling cap.
Last edited:
Its true that you can always chose one point in a circuit to call ground, even if it isn't a very wise choice. Thus, one could choose to call ground things that would require to either float the whole power amp, or the dac would have to float. The digital source feeding the dac would also have to float with the dac (if it floats) or be galvanically isolated. And the opamp responsible for holding two different power supplies and chassis apart at two different commons (perhaps one earth ground and one virtual ground) would have to be able to drive any parasitic capacitance those things might have relative to each other and or relative to earth ground (including inside power transformers). I don't see all that shown in sufficient detail in the model to evaluate the practicality of the whole thing, but I suspect one would probably be much wiser to forget the whole thing and move on to another type of dac design.
Last edited:
Dear Toots ...
You never did show us the bottom side of the PCB.
Speaking of which ... is this a 2-layer design?
Probably should consider 4-layer (with dedicated pwr and gnd planes).
Why? See:
EEVblog #1176 - 2 Layer vs 4 Layer PCB EMC TESTED!
Dear Toots:
You never did show us the bottom side of the PCB.
Speaking of which ... is this a 2-layer design?
Probably should consider 4-layer (with dedicated pwr and gnd planes).
Why? See:
EEVblog #1176 - 2 Layer vs 4 Layer PCB EMC TESTED!
it's not that easy either, this point must be capable to hold enough current, , opamps that can drive the headphones directly are a good choice
yes, one can do it, for a higher price ... I think he wanted to do it in dual mono, but he chosed the path of simplicity, the only thing I did not like is the capacitor in the signal path ...
but why to generalize the whole thing like you do? ... I like his simplicity, thats why I pointed out the possibility of a virtual ground, which removes this cap and preserves all the simplicity without changing the whole design of a single power supply to dual one ...
If it was a good, practical solution then I think we would see it used more often. It have doubts about how well it would work in practice. Have you built this and tested it before? If so, what dac, digital source, power amp, etc., did you use? Did you do tests to verify if it actually ended up sounding better and or measured better than just using output capacitor coupling with a decent quality capacitor?
I know similar ideas have have been worked on before. In one case it was to remove DC blocking capacitors and or transformers from condenser mic phantom power circuits. Some guys built a preamp that could do it using methods much like you have proposed, but it was complicated to make work and never went anywhere. I don't think it sounded any better than the conventional circuits with DC blocking capacitors, otherwise it might it might have turned into a product, or a good patent or something. But, no. It turned out not to be worth the bother.
Last edited:
capacitor is much more practical for majority of people and for manufacturers, why not to spend on design? good capacitor gives beautiful bass and that's what many consuments are looking for ...
minority of hi-fi users tend to remove the capacitor from signal path
I will not give you the joy and I will say no
do your own research, or stay with imagination, my opinion is not so relevant I know the guys, who still struggle with the right capacitor
I do not want to underestimate your instant imagination about circuitry sound
your google skill is even better .... patent/US5291121
and the global product knowledge is admirable
Interesting discussion chaps, thanks.
@miro's suggestion takes us part of the way towards a solution to eliminate the caps but its not the full enchilada. That is - it gets rid of the 'gross' offset (half the supply) but then there are inter-channel offsets to deal with, which here could easily be 100mV or so. So in conjunction with opamp servos I think it has a good chance of working.
There's also the issue of common-mode noise to consider - the ground between DAC and pre/amp is going to be potentially dirty with HF common-mode currents and I'd rather not send those through an opamp's output - it becomes inductive at higher frequencies meaning its going to present significant impedance. So I'd want to add an LC to the virtual ground's output, the C can then take the HF CM currents harmlessly away.
All in all a much more complex (therefore expensive) solution than a couple of caps - I shall listen to a few kinds and go for the best. Cap differences are fairly subtle in my experience.
@miro's suggestion takes us part of the way towards a solution to eliminate the caps but its not the full enchilada. That is - it gets rid of the 'gross' offset (half the supply) but then there are inter-channel offsets to deal with, which here could easily be 100mV or so. So in conjunction with opamp servos I think it has a good chance of working.
There's also the issue of common-mode noise to consider - the ground between DAC and pre/amp is going to be potentially dirty with HF common-mode currents and I'd rather not send those through an opamp's output - it becomes inductive at higher frequencies meaning its going to present significant impedance. So I'd want to add an LC to the virtual ground's output, the C can then take the HF CM currents harmlessly away.
All in all a much more complex (therefore expensive) solution than a couple of caps - I shall listen to a few kinds and go for the best. Cap differences are fairly subtle in my experience.
Oh, you're perfectly welcome, sir.
Yarn | Oh, you're perfectly welcome, sir. ~ Star Wars: Episode V - The Empire Strikes Back (1980) | Video clips by quotes, clip | 879491b6-0519-4fc5-bcd1-14746cad5b0a | 紗
"Oh, you're perfectly welcome, sir."
Yarn | Oh, you're perfectly welcome, sir. ~ Star Wars: Episode V - The Empire Strikes Back (1980) | Video clips by quotes, clip | 879491b6-0519-4fc5-bcd1-14746cad5b0a | 紗
I've made one or two minor tweaks to component values for this version, based on measurements of the first three v1 PhiDACs. This is the schematic I'll be putting together kits for (barring any unforeseen howlers of course
BOM for kits to follow shortly. Mouser BOM will be along after that.
Hi Abraxalito, I'm still in at the conditions you proposed beginning January.
Cheers,
JM