AudioSector-chip amp kits, dacs, chassis

[joshuajoshua]

Originally posted by Peter Daniel
Thanks Bob for clearing that up.

The DAC output section of my DAC is very similar to 47Labs DAC, which was measured by Stereophile: http://stereophile.com/digitalproces...00/index5.html

Thanks for the link!

Joshua

[Bobken]

Hi Joshua,

I am quite sure that you aren't arguing, and I didn't intend to suggest that you were, which is why I said "with respect etc".

I just wished to make it clear to you that in this particular case there is no question of "likely to be", it is the value of these shunt resistors which *determine* the output impedance of this DAC.

Basically, it is the impedance measured between the 'live' output terminal and ground (this is rarely the same as whatever is in series with the output, and it certainly isn't the same in this circuit), so in a straightforward case like this it is easy to determine simply by checking the schematic.
In truth, it will be a little lower than the exact value of the resistors used here, as there will be some much higher impedances in parallel with these resistors, within the TDA chip, itself, also referenced to ground. However for practical purposes, they will have very little effect on this output impedance in my opinion.

Anyway, I don't wish to hi-jack Peter's thread with this correspondence, but as I just saw your query when I was looking at this thread earlier and I had some time on my hands, I thought I might save Peter some trouble.

Regards,

[joshuajoshua]

Originally posted by Bobken
Hi Joshua,

I am quite sure that you aren't arguing, and I didn't intend to suggest that you were, which is why I said "with respect etc".

I just wished to make it clear to you that in this particular case there is no question of "likely to be", it is the value of these shunt resistors which *determine* the output impedance of this DAC.

Basically, it is the impedance measured between the 'live' output terminal and ground (this is rarely the same as whatever is in series with the output, and it certainly isn't the same in this circuit), so in a straightforward case like this it is easy to determine simply by checking the schematic.
In truth, it will be a little lower than the exact value of the resistors used here, as there will be some much higher impedances in parallel with these resistors, within the TDA chip, itself, also referenced to ground. However for practical purposes, they will have very little effect on this output impedance in my opinion.

Anyway, I don't wish to hi-jack Peter's thread with this correspondence, but as I just saw your query when I was looking at this thread earlier and I had some time on my hands, I thought I might save Peter some trouble.

Regards,

Hi Bob,

Thanks for the explanation! Ok, so the output impedance is the impedance seeing from the input of the stage after.
Yes agree, for practical purposes we'll take the lowest impedance which is the external resistor parallel with the DAC output, it should be close....and don't need to plot the output impedance across some range of frequency as well...

No such of hi-jack thing in my opinion, this is the beauty of the forum - learn from different perspective!

Regards,

Joshua

[Bobken]

Hi Peter,

I hadn't seen your response when I started my last post (I got interrupted half way through writing it!) and I should apologise for my 'interference' in your thread.

I now see from your reference to Stereophile that their measurement was 2k3 for the 47 Labs version, and if they used this same 2k7 value as yours for the 2 shunt resistors concerned, this is rather lower than I would have guessed.

However, as I just said, any impedance to ground within the TDA chip's circuitry will have some effect on this issue, and needs to be considered in the equation as this will be effectively in parallel with these (output shunt) resistors in this case.

Next time, instead of saying "determined by" or "based on", I should remember to say "will be mainly determined by, or mainly based on".

Regards, and a happy Christmas to you and the family.

[Peter Daniel]

The 47Labs DAC is pretty close to this design: http://www.fedde.nu/audio/images/nonoz3.gif and if I'm not mistaken they use 3K resistors for I/V as in attached pic.

Happy Christmas

[Bobken]

Hi,

I seem to have dug myself into a bit of a hole here, so it will teach me to mind my own business in future!

The matter is a little more complex (isn't it always?) in that there will be variations at different frequencies (to satisfy Joshua's query), and strictly-speaking the output impedance should be stipulated at a particular frequency. This is why it is called output impedance, and not output resistance.

However for our purposes in a situation like this, it is basically what I have outlined, and few (including Stereophile, it seems) will even indicate at what frequency their measurements have been made, anyway.

This Fedde circuit is rather different as you have several extra caps and resistors in this part of the circuit which are not in Peter's, and these all have paths down to ground and will reduce overall impedances at certain frequencies as a result.

Regards,

[Peter Daniel]

Whenever someone asked me about DAC's output impedance I also always specified it as 2k7

[Bobken]

Hi,

One final comment, if that pic of Peter's is a genuine 47 Labs Gaincard interior, then I for one am appalled at the constructional quality and layout.

Compared with anything you ever see of Peter's own work, that looks more as if it has been put together by a real amateur. Also, using parts 'flying in the air' like that is sometimes beneficial to keep certain signal-paths short etc., but this has hardly achieved anything worthwhile in this regard, if at all.

It just looks like a haphazard 'first-attempt' for an initial circuit-trial, like a prototype breadboard design.

[joshuajoshua]

Originally posted by Bobken
Hi,

I seem to have dug myself into a bit of a hole here, so it will teach me to mind my own business in future!

The matter is a little more complex (isn't it always?) in that there will be variations at different frequencies (to satisfy Joshua's query), and strictly-speaking the output impedance should be stipulated at a particular frequency. This is why it is called output impedance, and not output resistance.

However for our purposes in a situation like this, it is basically what I have outlined, and few (including Stereophile, it seems) will even indicate at what frequency their measurements have been made, anyway.

This Fedde circuit is rather different as you have several extra caps and resistors in this part of the circuit which are not in Peter's, and these all have paths down to ground and will reduce overall impedances at certain frequencies as a result.

Regards,

Ok, 2k7 ohm is a good approximation of the output impedance

Thanks,

Joshua

[Bobken]

Originally posted by joshuajoshua


Ok, 2k7 ohm is a good approximation of the output impedance

Thanks,

Joshua

You are very welcome.

Interestingly, if you add the 50k resistors at the output like Fedde's R9 & R10 (and I think Peter's PCB allows for these), this will lower the nominal output impedance from circa 2k7 to not much above 2k5 Ohms, with no other changes.

Regards,