Sorry, I have no measurement facility. I tried another two leg crystal it's ok.. but with another four-leg type of oscillator which I sure that is working, faster playback again.. I think there is a reason for using ordinary two-leg types of crystals with these cards.
Perhaps the chip runs at half the frequency of the crystal nominal value? I know crystals can run at multiples (harmonics), but I am not sure about fractions.
Your direction gave me an idea..
I measured playback time vs. actual time, then i realize that there is exact 1/4 delay. Then actual clock must be 4/3 of 24.576, which means 18.432 mhz.. Now I have to find a oscillator to try with this value.
The required oscillator frequency is 24.576MHZ, the same as the Xtal that you removed.
Download the data sheet for the DIR9001, and have a look at fig.9.
It should be fairly similar to what you have. You could also try a series resistor of 47-100 ohms in series with the output of your new oscillator.
SandyK
That is what i have done exactly.. I have used 24.576mhz oscillator. Also tried 100ohm resistor in series but it scratched the playback completely. If you sure that i have to use 24.576mhz, then have you any idea of exact 1/4 delay?
terranigma
Sorry, but I do not know what is causing your problem. Perhaps the waveform isn't correct, or it is of incorrect amplitude? It seems to be a pretty standard type of oscillator input as used with many different I.C.s.
I recently replaced a simple xtal oscillator as shown in the DIR9001 data sheet with a 3.3V 24.576MHZ oscillator module in a DAC without any problems.
SandyK
To me that happens if I select bitmatched playback and I forget to change in Windows the 48kHz samplerate to 44.1kHz. Don't know why there are not in sync when bitmatched is selected (but they are sync'ed when I force the sound card on a certain samplerate). But the speed difference is then 48/44.1...
The value of 24.576MHz seems correct since it is 512 multiple of 48kHz (256 of 96kHz), and that's the native samplerates of X-Fi.
The value of 24.576MHz seems correct since it is 512 multiple of 48kHz (256 of 96kHz), and that's the native samplerates of X-Fi.
Last edited:
Yes..that was it. Since i'm using ASIO as output, samplerate and masterclock varies dynamically. Weird thing that in my setup, if you ever play a 44100 track, there was no fallback for normal operation even you play a 48k track. You have to reboot. I just added SoX resampler to DSP list and everything seems fine..
Thank you all.
And hope to software resampler (Sox what i use) would rival X-fi's hw resampler. Because hw resampler completely out of function after removing crystal. Sox resampler seems one of the best out there.
SRC Comparisons
SRC Comparisons
- Status
- Not open for further replies.