So we have a certain degree of dimensional accuracy that’s had by the lidar…
I just realized, after having a picky petg and having to actually run the calibration vs the print-start calibration wasn’t working with it, that gives me an idea.
I see my retraction settings pop out as a secondary issue.
Why not run calibration of say 5 lines, 1 line is 10mm with a 2nd layer that’s 5mm, repeat for the other 4 lines.
This would alllow for adjusting retraction settings 5 ways, then seeing Z where the retractions happen, and picking the best retraction setting depending upon whether retraction or detraction is poking above or below the “point of flatness” where the top of the layer should be. If that makes sense.
This could also be used to analyze Z hop in a similar fashion.
Unfortunately, Bambu does not allow us to access the Lidar data. So usage is pretty limited to what Bambu deems a priority.
As for retraction settings, most available tests need a fairly high number of retracts as stringing does not need to happen on every retract.
So I expect that a test based on a very low number of retracts to give a wide range of settings with a low probability to give the desired result on a larger print. If we could use the Lidar data, we could probably come up with something on a single layer based on post-retract flow analysis, but chances are that we’ll either not see anything or have a wide parameter window.
From
https://wiki.bambulab.com/en/software/bambu-studio/parameter/retraction
Ya I’m not talking stringing, which is still identifiable with 2 posts.
I’m talking about calibrating retraction to do what it’s supposed to do. Leaving the correct amount of filament at last point of application. Which is surprisingly simple. It would look like the following.
Conversely when starting a line, you can identify too much/too little detraction without benchmark prints by the physics “footprint” during the point of application.
Too much retraction can lead to inadequate volume of filament when starting a line, thus it upturns as the nozzle continues it’s application. Only sticking to the plate or previous line when there is sufficient filament to “complete the math” behind the print parameters. Vice versa in the other direction, not enough retraction with too much ooze or flow, results in a tiny bump at point of application.
Retraction isn’t unique to stringing, though it’s a basic identifier most people can see easily. What I’m proposing is using the microscopic camera to verify uniformness and proper geometry, which is what it excels at.
Luckily Bambu listens to us unlike other companies. I’ve seen quite a few things they’ve implemented after the general public give feedback. It’s quite refreshing. 
Yes, I usually use wipe with retraction for which this could be helpful.
A bunch of 1st layer retracts with wipe should enable a pretty good calibration using the Lidar.
Yes, I’m sure we will see this very feature soon, right after we get truly secure cloud access, multiple accounts on a printer, folder support on the SD card, pressure advance settings in the filament profile, LAN-only Handy video monitoring…
I wanna say it only took them a few months to implement real-time spaghetti detection during a print (vs 1st layer) from when I first heard the idea floated on the forum.
I never thought I would see it 