Questions on usrp_tv_rcv.py

GNU Radio
1

Thanks for pointing out all theses raw files
for testing the usrp_tv_rcv.py application.
I tried a couple of them and they work fine.

Now I have the following question:

Why is it that the video shown on sdl moving
slowly to the right?

The fact that there is no vertical or horizontal
synchronization should NOT explain this:

I use a decimation of 8 thus the video rate is 8M.
With 625 lines per frame and 25 frames per second, this results in
512 samples per line (exactly), ie, 8,000,000/25/625=512.

So one would have expected to see the corresponding picture
not aligned in the vertical/horizontal dimension,
BUT AT LEAST the misallignment should have been CONSTANT over time.
This is clear in the vertical dimension (no up/down sliding)
but not in the horizontal…

What is the explanation of this?

Thanks
Achilleas

PS: one explanation I can think of is the
instability of the 64M ADC…

2

I guess I am attempting an answer to myself here:

the slow horizontal sliding reveals a
missalignment on the order of
200 samples/sec.
(I measured that by seeing how much the black strip moves
to the right within the 1 sec of video stream).

For a sampling rate of 8Msps this is can be explained by a
drift/jitter of the ADC on the order of 25ppm

Is this within the drift/jitter specs of the ADC?

Thanks
Achilleas

3

On Tue, Jan 23, 2007 at 01:41:29PM -0500, Achilleas A.
wrote:

drift/jitter of the ADC on the order of 25ppm

Is this within the drift/jitter specs of the ADC?

Yes, this is within the specs on the crystal oscillator.

Eric

4

Anastasopoulos Achilleas wrote:

drift/jitter of the ADC on the order of 25ppm

Is this within the drift/jitter specs of the ADC?

Yes, the master clock is spec’ed at +/- 50ppm over temp

Matt

5

On 1/23/07, Matt E. [email protected] wrote:

Yes, the master clock is spec’ed at +/- 50ppm over temp

Matt

Are there any plans on putting a highly stable reference TCXO on the
board? I know CMAC makes a very stable 0.3ppm part that goes up to
80MHz with a 3.3v Square ACMOS output.

Tracking over a few ppm in software can be pretty difficult I am
aware, so I was just curious if there were any plans on having a
“high-performance” USRP with such a stable reference on it.

Brian

CMAC Information: http://www.cmac.com/fcp/products/TCXOs.php