Linien signal on Output 1 of Red Pitaya is noisy

[daqt-dit1]

Issue

The signal generated by Linien on the output port 1 of the Red Pitaya is significantly noisier than the one generated by the built-in Red Pitaya oscilloscope.

Description

I am currently using Linien to lock a laser to a certain frequency from an atomic spectrum. Unfortunately, the signal generated by Linien on output port 1 is noisy and this affects the lock. This is especially true for low modulation amplitudes, which are the ones I currently need to modulate the laser. I have tested this for different versions of Linien and different Red Pitaya boards and the
results always showed that Linien generated signal on output port 1 is noisier.

To Reproduce

1. Install Linien (I have used versions 1.0.1 and 2.0.4).
2. Connect to the Red Pitaya board using the Linien GUI.
3. Set the modulation frequency to output port 1 (if not set already) and record the modulation amplitude and frequency used.
4. Connect output port 1 to an oscilloscope (in my case I used Keysight DSOX2012A) and measure the signal generated.
5. Connect to the Red Pitaya using the Red Pitaya GUI (type rp-xxxxxx.local in the web browser).
6. Use the recorded modulation frequency and amplitude to generate the same signal with the built-in Oscilloscope app
   from the Red Pitaya GUI.
7. Repeat steps 3-6 but for output port 2.

Expected behavior

The signal generated by Linien on output port 1 is significantly noisier than the one generated by the built-in Red
Pitaya Oscilloscope App whilst the signal generated on output port 2 seems to be reasonably clean
(more or less similar to the signal generated by the built-in Red Pitaya Oscilloscope).

[git-father]

I have similar issue with the ouput port 1. For example, when I set the modulation frequency as 5MHz, the output port 1 produces ~4.88MHz. The difference is proportional to the modulation frequency. When I set 25MHz, then the difference becomes ~590kHz, which is 5 times to the 5MHz case. This difference disappear when I use the built-in Red Pitaya Oscilloscope.

I'm using linien for digital servo of PDH locking, and I could find there are also modulation depth-proportional frequency in the error signal while locking(but different frequency with above, for example, for 5MHz case, there are ~610kHz noise which is 5 times larger than the above difference). I'm wondering that does it could be demodulation error, which is due to the difference between the modulation frequency and demodulation frequency.

I found the error frequencies on oscilloscope, FFT of the exported error signal during locking, and linien's PSD measurement.

[Andrew-wi]

+1, have also observed a frequency-proportional frequency offset between the linien GUI setting and the measured frequency output.