NOAA/METEOR HRPT not syncing correctly

[Crosswalkersam]

Hello,

I tried receiving NOAA / METEOR HRPT with an Airspy Mini @ 3MSPS and Satdump 0.0.39. Its Win 10 with an i5 processor and 16GB of RAM.
The satellite is there and clearly visible in the spectrum when I record a baseband.
When I try to demodulate it however, I get something like this:

The BPSK is clearly splitting into two groups, but connected with two "bridges", and the Deframer does not sync.
However: After 2/3 Minutes, it suddenly starts working:

You can see that the SNR jumps from 6dB to 9/10dB almost instantly, and the deframer syncs fine.
I tried enabling DC-Block, but that did not fix the issue. This happens with NOAA and METEOR.
Decoding with a RTL-SDR@2.56MSPS, I never had that happen. If you need, I can supply the .i16 file recorded with satDump aswell.
Thanks!

[Aang23]

That's weird. Did DC-Block help? Otherwise, yes it will be easier to just look at the .i16 on my end.

[Crosswalkersam]

DC-Block does not help. Here is a folder that contains a recording of NOAA18 at 3MSPS.

[Crosswalkersam]

I just checked, the issue goes away when sampling at 6MSPS. I guess the Airspy behaves a bit weird at 3MSPS?

[zeke800]

Apparently, yes.

[mapaknot]

Hey
Synchronization is interrupted. I changed one setting and everything was fine for me.
#####"rrc_alpha": 0.5, #####

"frequencies": [
           1707
           1698
           1702.5
       ],
       "sampling rate": 3e6,
       "baseband_type": "i16",
       Work: {
           "main band": {},
           soft: {
               "pm_demod": {
                   "symbolrate": 665400,
                   "agc_rate": 0.002e-3,
                   pll_bw: 0.01
                   pll_max_offset: 0.5
                   "rrc_alpha": 0.5, 
                   "rrc_taps": 31,
                   "costas_bw": 0.005,
                   "clock_omega_relative_limit": 0.005,
                   "buffer_size": 8192

Hello,

I tried receiving NOAA / METEOR HRPT with an Airspy Mini @ 3MSPS and Satdump 0.0.39. Its Win 10 with an i5 processor and 16GB of RAM. The satellite is there and clearly visible in the spectrum when I record a baseband. When I try to demodulate it however, I get something like this: The BPSK is clearly splitting into two groups, but connected with two "bridges", and the Deframer does not sync. However: After 2/3 Minutes, it suddenly starts working: You can see that the SNR jumps from 6dB to 9/10dB almost instantly, and the deframer syncs fine. I tried enabling DC-Block, but that did not fix the issue. This happens with NOAA and METEOR. Decoding with a RTL-SDR@2.56MSPS, I never had that happen. If you need, I can supply the .i16 file recorded with satDump aswell. Thanks!