Why a samplerate of 2.4 MSPS may not be the best samplerate for cheaper RTL dongles

[thewraith2008]

I have observed that while using dongles with standard crystal which have higher PPM values while using the samplerate 2.4 MSPS, a signal become visible in the center when the PPM value is moved away from 0. The more the PPM moves from away from 0, the frequency increases. see GIF.

Testing with cheaper RTL dongles with higher PPM.
This is more visible when you have the FFT resolution set above 32768.
I have also seen the signal described using the RTL-SDR.COM V3 dongle with TCXO but to lesser degree where PPM value is not 0.
Can be seen when using with SDR#, SDR Console and HDSDR.
SDR# 'Correct IQ' has no affect.

Same can be observed at samplerates:
1.8 MSPS
0.900001 MSPS except signal is not nulled when PPM is at 0.


I have created a modified RTL frontend for SDR# that I can more selectively set a more favorable samplerate where this does not occur and spikes/spurs are reduced.

I'm finding that using a non standard samplerate is also stopping other problems I've seen with the standard SDR# RTL frontend.

 

[Ubbe]

I see the same at 2.4 using a v1 RTL-SDR. But I use 1.024 and in that setting there's no problem.

/Ubbe

 

[jonwienke]

This will fix the problem at any sample rate or PPM setting:

 

[slicerwizard]

"SDR# 'Correct IQ' has no affect (sic)."

 

[Ubbe]

I tried "Correct IQ" at the 2.4 sample rate at it didn't change the anamoly one bit.

/Ubbe

 

[jonwienke]

Weird. It worked for me, but I'm using an older version. Bug?

 

[thewraith2008]

Using the "Correct IQ" option should not be required when using the SDR# RTL frontend as it already applies IQ balancing to the IQ data before it make it available to SDR#.

I'd still like to know out of curiosity why this signal appears and where it comes from when changing PPM while on 2.4 MSPS.