Great resource, thanks David
On our tests, many users were reporting 100ppm error and BER ~ 5%. Even
with athat sample rate error , this is well within the capture of the
OFDM modem (nice one David !) , even with that steady state error, users
were reporting that non zero BER on otherwise high SNR channels. I have
some captures.
Using IC7300 via USB audio for TX :
I checked the 12 MHz crystal in the 7300 used for USBus was 15ppm low.
that's OK.
I hooked up my R&S modulation analyser , and brought the 'receiver'
audio into the computer via a USB analog audio channel.
I found that using the 'default' offered 44100, but this resulted in a
sample rate error of ~ 100ppm
Choosing 48000 for all sample rates resulted in 7 ppm error on my own
kit. This is something that used to be a common problem many years ago,
this is not a new problem. I suspect 7ppm is the difference between the
USBus 12 MHz crystal clock in the IC7300 (output clock) and the USB
audio card (input side) sample clock
My RTL SDR piped to FreeDV - the clock offset wanders and the
performance is poor. I think the short term stability of the RTL SDR SR
is not as good as required AND also the RTL SDR at 2.4MHz sample rate
was WORSE than at 2.048MHz sample rate (a nice 2^n number) , so there is
obviously some sample rate conversion issues internally in the SDR
software.
-glen
On 7/12/2020 8:00 AM, David Rowe wrote:
Hello Glen,
What is the sample clock offset corrector's maximum offset for pull-in,
and hold in ?
Take a look at the "Specifications" section of README_ofdm
- David
_______________________________________________
Freetel-codec2 mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
