I have a PCM2902 that I'm using between the computer and my Sansui integrated amp and I love the way it sounds.
I built it around the TI datasheet...it uses an LM317 for the supply, rather than the USB bus.
I'm simply using 10uf caps and some 100k bleeders on the output.
What is the (real--practical) value of having an opamp buffer stage on this chip?
I mean I know the theory of less loading but is it a problem with the PCM2902?
I suppose using an opamp gives you the ability to "tailor" a sound?
I believe in no frills -- but is an output stage a "necessary frill"?
Thanks.
I built it around the TI datasheet...it uses an LM317 for the supply, rather than the USB bus.
I'm simply using 10uf caps and some 100k bleeders on the output.
What is the (real--practical) value of having an opamp buffer stage on this chip?
I mean I know the theory of less loading but is it a problem with the PCM2902?
I suppose using an opamp gives you the ability to "tailor" a sound?
I believe in no frills -- but is an output stage a "necessary frill"?
Thanks.
Also.... (teach me)...
Do I want a higher drive voltage?
And a lower output impedance?
The 2902 shows a .5 v p-p output around a center voltage of .5v....
does this mean that the output swings between 0 and 1 volt?
The output impedance is 10k.
Is it desirable to get a lower figure?........I know most amp inputs are
usually a higher Z.
Are the opamps really used more for getting more robust drive levels?
Thanks
Do I want a higher drive voltage?
And a lower output impedance?
The 2902 shows a .5 v p-p output around a center voltage of .5v....
does this mean that the output swings between 0 and 1 volt?
The output impedance is 10k.
Is it desirable to get a lower figure?........I know most amp inputs are
usually a higher Z.
Are the opamps really used more for getting more robust drive levels?
Thanks
The output swings between 0.75V and 0.25V.. Divide Vpp by 2 and add and subtract that value from the dc offset to get the overall swing.
You need to buffer the output of this chip, a 10K source impedance is quite high and makes it prone to external noise pickup. Low load impedances may reduce the output level and increase thd as well. Generally for good linearity the input impedance of the succeeding stage should be about 10X higher than the source impedance for best linearity.
You can add some gain as well if needed. About 12dB (4X) would give you something approaching the old line level standard of 700mVrms.
Use AC coupling with a very good film cap between the DAC output and the op-amp input.
The AD8610 is an excellent device for operation at low-ish voltages and sounds good. A cheaper alternative might be the AD8510. For good performance you need higher rails - at least 9V if single ended supply is used, perhaps +/-5V or greater with bipolar supply rails.
Note that you could use a simple discrete amplifier circuit in place of the op amp if you preferred or even a tube.
Kevin
You need to buffer the output of this chip, a 10K source impedance is quite high and makes it prone to external noise pickup. Low load impedances may reduce the output level and increase thd as well. Generally for good linearity the input impedance of the succeeding stage should be about 10X higher than the source impedance for best linearity.
You can add some gain as well if needed. About 12dB (4X) would give you something approaching the old line level standard of 700mVrms.
Use AC coupling with a very good film cap between the DAC output and the op-amp input.
The AD8610 is an excellent device for operation at low-ish voltages and sounds good. A cheaper alternative might be the AD8510. For good performance you need higher rails - at least 9V if single ended supply is used, perhaps +/-5V or greater with bipolar supply rails.
Note that you could use a simple discrete amplifier circuit in place of the op amp if you preferred or even a tube.
Kevin
- Status
- Not open for further replies.