Congrats on your first frames!
Given that the observation is Doppler corrected the first thing is to check if the TLE are updated, for example here is a recent case with such an issue.
In this case the TLE was “16 hours, 11 minutes” old, so for a satellite without any propulsion and on its altitude (very low drag), on 16h we don’t expect any major change, so TLE are good.
Next to check is local station time, as you already checked, the system was synced, so this is not the problem.
Another check is if the transmitted signal is like that, we have seen in the past transmitters/transmissions that are affected from different parameters, like temperature, and result on drifting.
This check can be done through observations from other stations (the real power of the Network), so checking for example a recent one or even better one closer to your observation time, we can see that not any drift happens.
If you notice a little better, the signals actually follow a vertical line and they don’t drift diagonally, however there is a significant drift on the center frequency around 1-3KHz but this applies to all the transmissions during the observations.
For the latter drift, given that it is observed in different stations/observations, we need to update the transmitter drift field in DB with a suggestion and linking to this observations in citation field.
As for the diagonal drift, this is a kind of optical illusion created by not receiving the whole width of the transmitted signal, in other words in each transmission part of the signal hides behind the noise floor creating signals with different widths which results into the impression of diagonal drift.