This DAC have these modules:
Total: 11 power lines.
I can imagine three different modes of powering the DAC:
I can see some obvious differences in these approaches:
With 1) only two diode bridges is good for lowering high freq. switching noise, but all GNDs are tied together.
With 2) there hare 10 diode bridges and GNDs can be independent at PSU level.
Has been stated that PCM1704 is a great performer but very sensible with PSU quality and implementation. On the other hand, BB datasheet says:
“The PCM1704 only requires a ±5V supply for operation. Both positive supplies (+VDD and +VCC) should be tied together at a single point and connected to a single +5V analog power supply. Similarly, both negative supplies (–VDD and –VCC) should be tied at a single point and connected to a single –5V analog power supply. No advantage is gained by using separate analog and digital power supplies. It is more important that the analog supplies used to drive these pins are as noise and ripple free as possible to reduce coupling of supply noise to the output.”
Obviously, all GNDs will be tied somewhere, in the PSU or elsewhere, but this will make large differences in final results.
I have attached two extreme versions of feeding.
I appreciate any suggestion.
Regards,
Luis.
- USB interface (no PSU)
- Shift Register (+VD0)
- DAC = 2 x pcm1704 (+VD1, -VD1, +VD2, -VD2, +VA1, -VA1, +VA2, -VA2)
- I/V converter: Battery charger, channels L and R: +VB1, +VB2
Total: 11 power lines.
I can imagine three different modes of powering the DAC:
- One transformer with two secondaries, that feeds 11 voltage regulators (positive and negative.
- Six transformers: 1 for Shift Register. Four for DAC, 1 for charge batteries.
- An intermediate solution between 1) and 2)
I can see some obvious differences in these approaches:
With 1) only two diode bridges is good for lowering high freq. switching noise, but all GNDs are tied together.
With 2) there hare 10 diode bridges and GNDs can be independent at PSU level.
Has been stated that PCM1704 is a great performer but very sensible with PSU quality and implementation. On the other hand, BB datasheet says:
“The PCM1704 only requires a ±5V supply for operation. Both positive supplies (+VDD and +VCC) should be tied together at a single point and connected to a single +5V analog power supply. Similarly, both negative supplies (–VDD and –VCC) should be tied at a single point and connected to a single –5V analog power supply. No advantage is gained by using separate analog and digital power supplies. It is more important that the analog supplies used to drive these pins are as noise and ripple free as possible to reduce coupling of supply noise to the output.”
Obviously, all GNDs will be tied somewhere, in the PSU or elsewhere, but this will make large differences in final results.
I have attached two extreme versions of feeding.
I appreciate any suggestion.
Regards,
Luis.
One very stable and quiet +/- power supply for both the analogue and digital sides of the PCM1704s is all you need. Having more supplies does not help and can make things worse if your supplies are not quiet enough. In fact, all devices needing similar +/- voltages should be powered from the one supply. Observe proper decoupling practice.
This is how it is implemented now. A single dual PSU of +/- 5V for feeding PCM1704 and Shift Register.
Another PSU of 20V + 20V for battery charging.
I have noticed what huge difference in powering channels L and R with two independent pair of batteries of Output Stage (I to V converter Sen I/V). When Output Stage was powered with a sole battery for both channels, sound was very inferior.
This encourage me for triying to find proper powering solution.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.