Anyone Using the New Adaptive Pressure Advance in Orca Slicer

[Penatr8tor]

This looks like something worth trying.

Orca Adaptive PA

"This feature aims to dynamically adjust the printer’s pressure advance to better match the conditions the toolhead is facing during a print. Specifically, to more closely align to the ideal values as flow rate, acceleration, and bridges are encountered."

Thoughts?

[mvdveer]

Been there for some time now, but have not used it as yet. Not much out there about it either. So, for myself - I'm waiting  to see the outcome from others.

[BlackNet]

Not many printers outside the Bambu A1 use this feature. Right now, it is exclusive to Bambu, and they may have a patent on it as well, but I am not 100% sure about that.  

[GarthSnyder]

I've been using it for a few of my main filaments. It does improve the first layer. I often had pinholes between walls and infill on the first layer when using a constant PA value. Beyond that, I don't really notice much difference.

Calibration is time-consuming. It's not too different from a regular PA test, but you have to do it at least 12 times under different conditions, then use a spreadsheet to prep the parameter block. There's a batch mode that sets up multiple calibrations, but each test area is just a standard PA calibration print.

Overall I think there's something useful here, but I suspect that the model doesn't really need to be so complex. There are clear patterns in the test results that I don't think the current model takes any account of. (It seems to just be doing separate interpolations along two different axes.) I would bet that if we could collect enough calibration data, the model could be simplified to the point where it would only require three or four test prints. It wouldn't be as accurate as the current model, but I bet it would reach the 80% mark.

[claudermilk]

I have not messed with it. Right now the printers are doing well with the current calibrations so I'm reluctant to fiddle.

[PFarm]

2 hours ago, claudermilk said:

I have not messed with it. Right now the printers are doing well with the current calibrations so I'm reluctant to fiddle.

I feel the same way, got it dialed in and prints looking pretty good, so I'll let someone else do the fiddling.

[Penatr8tor]

3 hours ago, claudermilk said:

I have not messed with it. Right now the printers are doing well with the current calibrations so I'm reluctant to fiddle.

Yup... I'm thinking the same thing but...

I figured I'd throw it against the proverbial wall to see if it would stick.

[Lind301]

Test on my voron. its a bit of process to setup and it needs to be done for every filament/hotend your using now (think also the average room temp will do a difference).
but the end result looks like a bit more precise details but i need to have a blind test to say for sure (the work you put in creates a will that it will be a good thing).

And it might set you up for a nice experience with the coster setting and acceleration to deceleration settings.
Not 100% sure and it can be the slicer having a whoppsey but some moves was way to fast for my build so it made fun layer shift.

But if you do the work and run the test you might hit that "Wait a minute my hotend/extruder/nozzle/bowden is having a issue"   

For it to be "The EPIC thing to use" we need to have lidar/nozzle camera/load cell  to be a every day driver to speed up the process.

The "others" that use the buzzword AI the have a "Few" units out that the get data from and can train there LLM.
Might be for Obico to look into to add to there Spageti hunter LLM.
Empirical solution whack in a nozzle cam feed to obico and start to whack it on the fingers when its bad and give it a kiss when it ok.