Best Way to cut Supply to USB Socket

[jkeny]

Originally Posted by fmak

Good idea. However, in the USD version, bus power is regulated down to 3.3 and 1.2V via 1117 regulators and I think this accounts for the 'good' sound. It is these regulators that need improving and I may get round to it. The onboard closk is also disntinctly 'inferior' when compared to a reclocked output into my UA2192 which also offers a clock cleaning option.

The dac version may be different.

I know that the 1117 regs which feed 3.3V & 1.7V to the Cypress & Xilinx chips & the crystal are isolating these, somewhat, from the raw USB supply but still an improvement in raw supply will improve matters.

In the US model, the PCM1793 DAC chip are fed by this 3.3V & by 5V & OP275 output stage is fed by +/- 5V. This 5V (+/-) is derived from the raw USB supply through an involved circuit of inductors & some chips I don't recognise which cleans up the 5V supply and provides a -5V supply

My next round of mods to this beast will involve a low jitter Crystek clock for the crystal but also a clean supply for the 5V, 3.3V & 1.7V.

I see mine moving away from a neat portable unit to rehousing it in a bigger box with all the ancillaries that are warranted - power supplies, I2S to other DAC, transformer output stage, etc

[fmak]

Originally Posted by jkeny

I know that the 1117 regs which feed 3.3V & 1.7V to the Cypress & Xilinx chips & the crystal are isolating these, somewhat, from the raw USB supply but still an improvement in raw supply will improve matters.

In the US model, the PCM1793 DAC chip are fed by this 3.3V & by 5V & OP275 output stage is fed by +/- 5V. This 5V (+/-) is derived from the raw USB supply through an involved circuit of inductors & some chips I don't recognise which cleans up the 5V supply and provides a -5V supply

I see mine moving away from a neat portable unit to rehousing it in a bigger box with all the ancillaries that are warranted - power supplies, I2S to other DAC, transformer output stage, etc

I actually took the 'messy' route by modifying a double screened usb cable so that the input socket to the converter is fed by a double regulated 5V supply. This alters the 90R impedance of the digital line but improves sound significantly (bass resolution, imaging etc, and removes a degree of colouration in the overall sound when compared to the rme 9362).

Try it.

[jkeny]

Originally Posted by fmak

I actually took the 'messy' route by modifying a double screened usb cable so that the input socket to the converter is fed by a double regulated 5V supply. This alters the 90R impedance of the digital line but improves sound significantly (bass resolution, imaging etc, and removes a degree of colouration in the overall sound when compared to the rme 9362).

Try it.

I thought it might improve matters even in your transport only 01 USD model - I wonder if it's major influence is a cleaner PS to the 24MHz crystal rather than the Xilinx or Cypress chips?

[fmak]

Originally Posted by jkeny

I thought it might improve matters even in your transport only 01 USD model - I wonder if it's major influence is a cleaner PS to the 24MHz crystal rather than the Xilinx or Cypress chips?

There is more to it than that. I tried two PSs, a 7805 and a LT 323 with identical 40 uV rms noise level fed by a 317 supply. They sound quite different!

The LT chip has greater hf clarity whereas the 7805 less resolving.

I am out of Jung regulators at present but will make one up and rig these up for measurement.

By the way, the 01USD is 100 kHz out on fs.

[jkeny]

Originally Posted by fmak

There is more to it than that. I tried two PSs, a 7805 and a LT 323 with identical 40 uV rms noise level fed by a 317 supply. They sound quite different!

The LT chip has greater hf clarity whereas the 7805 less resolving.

I am out of Jung regulators at present but will make one up and rig these up for measurement.

By the way, the 01USD is 100 kHz out on fs.

Well, it's not unusual for different PSs to sound different even in digital only systems. But, as I asked on AA have you tried running the output through your reclocker - does it still improve the sound as much as before?

Also you mentioned on AA that when you bypassed the on-board 1117 regulators it made the sound worse - can you say more about this?

Also, are you telling me the word clock output from the 01USD is 100KHz & this has something to do with the whine I'm hearing when connecting another DAC to I2S? Can you also say more about this?