Hello. I have been using the twcs for 3 years in bms and concurr that although it is great by many aspects, this device has weaknesses, the main one being the ministick.
Mine quickly had a bit of plastic friction (with weird noise) and uncontrollable sensitivity. The worst was centering problem : it kept drifting unless I applied subtle pressure, which is horrible for both the fcr and the tgp.
First I tried applying a large deadzone in bms (via the alternative launcher actually). It was better but then it was kind of on/off due to the reduced ‘sampling’ range.
So I started using joystick gremlin to precisely apply curves. The idea was to limit the overall sensitivity (max range), and define precise/progressive deadzones. I got good results, but as it was to slow (but accurate and controllable ) I had to slightly increase the cursor speed in bms config tool in order to keep high efficiency in fcr use especially.
I also discovered or understood then, that I should area track on the AG tgp first, before attempting any fine adjustment. In area track, the tgp is much more controllable.
I would advise to try and tune a bit with joystick gremlin. I tried Target but found it not as versatile or efficient (ergonomics).
As a side note : I tried disassembling the ministick as I thought there was plastic friction external to the ministick. In fact that plastic noise comes from the internals of the mini stick itself (mecanic cage, kind of). Bad quality I guess. I applied a few drops of nyogel 767a on the sides of the cage in it. Wouldn’t say this really changed anything to be honest. Not in the magnitude of the improvement that lubricant gives to the rails of the throttle itself (highly recommended for this part !). And probably risky for the pots/electronic in the ministick. I would not do it again I think.
I hope this helps a bit. I tested many many approaches.