Tír Na HiFi

Sorry, I don't understand your question?
Edit: Re-read your post & understand your question now.
Powering the 48MHz clock @1.8V & then dividing it via a flip-flop to 24MHz for sending onto USB hub chip may not result in as good a delivered clock signal as running the 48MHz clock @ 3.3V, dividing it via flip flop to 24MHz & then using a capacitive divider to reduce the signal to 1.8V or it may be the other way around - only testing will tell :)

hit a snag with 1nf its drawing too much current to gnd meaning the clks not operating at least not properly

so either i use the other half of the 74 ff to generate a higher current output signal to use the 1nf cap divider on
or i buy even more parts to accompany some 100/120pf caps

data sheet of xmos says it could be programed for 12/24/48/96mhz
its a pity we couldn't change that

i don't like having to put clk signals though caps it changes the sound

nige2000 wrote:hit a snag with 1nf its drawing too much current to gnd meaning the clks not operating at least not properly

so either i use the other half of the 74 ff to generate a higher current output signal to use the 1nf cap divider on
or i buy even more parts to accompany some 100/120pf caps

Ah, that's a shame!

data sheet of xmos says it could be programed for 12/24/48/96mhz
its a pity we couldn't change that

Yep, I'm working on that :)

i don't like having to put clk signals though caps it changes the sound

Sure, that's what I was thinking.

if you get the xmos working at 24mhz are you going to use 1.8v for clk power rather than a signal voltage divide

nige2000 wrote:if you get the xmos working at 24mhz are you going to use 1.8v for clk power rather than a signal voltage divide

Yea, both USB2412 USB hub chip & XMOS ucontroller can both use 1.8V 24MHz clock, I believe?

jkeny wrote:

nige2000 wrote:if you get the xmos working at 24mhz are you going to use 1.8v for clk power rather than a signal voltage divide

Yea, both USB2412 USB hub chip & XMOS ucontroller can both use 1.8V 24MHz clock, I believe?

probably the best outcome if it can be done

does xmos have to be reprogrammed or just adjusted

nige2000 wrote:

jkeny wrote:

nige2000 wrote:if you get the xmos working at 24mhz are you going to use 1.8v for clk power rather than a signal voltage divide

Yea, both USB2412 USB hub chip & XMOS ucontroller can both use 1.8V 24MHz clock, I believe?

probably the best outcome if it can be done

does xmos have to be reprogrammed or just adjusted

Reprogrammed - so the whole firmware has to be done, not just tweaked like hardware - it's not like the firmware can be downlaoded & reprogrammed & DiYinHK or any XMOS DAC manufacturer will give away the source.
Still, it's not impossible, just takes a bit of time

Another out of the box solution, use a transformer to lower the signal amplitude.

A 1:2, or 1:1 with center tapped secondary should work.

BTW, I thought of this because I noticed that the really expensive neutron star clocks use a transformer, looks like it is right before the clock output

http://www.newclassd.com/index.php?page=36

They are using a 1:1 ethernet transformer for isolation, with a cap and diode on the output side, another thing to try.

Randy

randytsuch wrote:Another out of the box solution, use a transformer to lower the signal amplitude.

A 1:2, or 1:1 with center tapped secondary should work.

BTW, I thought of this because I noticed that the really expensive neutron star clocks use a transformer, looks like it is right before the clock output

http://www.newclassd.com/index.php?page=36

They are using a 1:1 Ethernet transformer for isolation, with a cap and diode on the output side, another thing to try.

Randy

transformers were used in series with a dc blocking cap in paul pangs clock modules
i disliked its effect as it lacked detail compared to been direct but it sounded smoother which is probably likable to some
on the scope it changed a cleanish square wave into a sine wave
actually so far ive preffered a square wave as long as there isnt significant preringing

nige2000 wrote:
transformers were used in series with a dc blocking cap in paul pangs clock modules
i disliked its effect as it lacked detail compared to been direct but it sounded smoother which is probably likable to some
on the scope it changed a cleanish square wave into a sine wave
actually so far ive preffered a square wave as long as there isnt significant preringing

How does the cap divider look with your scope?