Thatās 100% the same noise I was getting.
I actually wrote to Jason at Instruo with a recording of my noise and Iāve posted his reply below in case it helps.
Iām still concerned thereās an issue with my module too, but it did seem to improve when i moved it from my main intellijel case into my tiptop happy ending kit.
Let me know if trying the below helps.
Thanks for getting in touch and taking the time to record an example and message.
Gain staging is most likely going to be the solution here.
The Arbhar does have a noise floor which is comparable to many other large digital based modules out there.
The audibility will depend largely on setup and different cases/power supplies seem to behave quite differently.
The general level of signal to noise ratio that the Arbhar and Lubadh achieve in practice is somewhere around -72dB to -76dB. (Measured against a 1KHz sine tone as reference).
These levels are compatible to the audible bleed that can occur between analogue oscillators that share power from the same supply/bus board.
How are you experiencing the noise floor compared to unity gain audio signals?
A good test is to use an oscillator for reference. Tune to something comfortably audible and monitor at a comfortable level.
If you are using a desktop mixer following your modular, I would advice monitoring the tone directly into that with gain adjusted so thereās no distortion at any stage.
Once you have that reference, capture that raw tone into the arbhar and monitor grain playback through the same signal path.
Make sure the output level on the arbhar is up full (itās an analogue attenuator so unity is at full).
Experiment with the Arbharās input level control when finding a sweet stop. The preamp will drive at a certain point so maybe start somewhere around the 2oāclock position.
When playing grain back, try a low intensity, full length and gaussian curve to begin with.
The goal is to approximately match the original raw signals for amplitude and tone.
Let me know how the noise floor compares in this case.
Have you compared functioning signal to noise ratio?
Have you also tried powering just the Arbhar on its own in your system and/or trying a different power supply/case?