What I am talking is oriented more towards what does G0001 ---->G9999 do as a very crude example. I’ve seen some forum posts where people try to modify that start and stop sequence and blow up their printer. Apparently a given g-code may be used to do different things by different programmers
OctoPrint using G91 to do something is no guarantee that Bambu Labs will do the same thing with G91 (I just grabbed a random number off my printer)
Sorry, I misunderstood your question.
Yes, I agree with you, there is no clear list of G-code functions used by Bambulab… that’s why I try not to touch anything in the Start/End Gcode
Studio, on your computer.
Handy is not a slicer, it passes your print request to the Bambu servers which send g-code to your printer. The options for those prints are limited, so it is unclear to me whether that code is generated fresh for each request or if the servers store a pre-sliced file for each possible combination.
I’m no programmer, but this topic and this site have helped me interpret and make trivial changes in some Start and End g-code.
1 Like
Yeah that is fair Bambu doesn’t have a list of the commands anywhere but the Bambu firmware is based on the Marlin firmware so you could take a look at their wiki to see generally what each command does. Also Bambu does a decent job of commenting the start and end Gcodes to say what it is doing.
But also Bambu could reuse or overwrite what a code command does in their firmware so it’s a your mileage may very kind of thing.
Seem to be having the same issue on a brand new P1S. I have an older P1S too, on which I tried the same print, same profile, same settings etc. but no issues on that one. (Things to note are that my old P1S has an older PEI sheet and has an AMS+buffer, new P1S does not).
I’m at a loss so far as well: bed seems absolutely flat, tried the bed from my other P1S (less but still same issue), replaced tubes, checked inside of the toolhead, front/top panels open, aux / chamber fans off, did the tramming procedure and of course ran full calibration many times.
No idea what it is, but my gut leans to it having to do with either cold spots on the bed, or my old P1S not having the same issue because it has an AMS buffer. (Does the buffer even do anything when not using AMS / external filament? If it does, guess it might also make non-AMS printers more reliable?)
If you suspect this, you can try to tease out the issue by setting your heatbed on the printer’s pad, take a walk for 15 mins, then try to send the test. Small irregularities will entropy themselves out.
(I may have forgotten to mention I am experiencing this on the stock, Textured PEI sheet)
Good tip! Tried this, no change… Here’s two pics of tests I printed after 15min heatsoak at 60C:
(Same settings, same print)
This is on the far right corner of the bed, testing square is 10x10cm.
A much higher bed temp also didn’t help, guess it’s not cold spots? I got an industrial ruler to check bed flatness, and at first glance there actually seems to be a dip… but I’d need to measure more accurately to see how deep it actually is. Doubt the bed mesh leveling wouldn’t be able to handle it though…
It’s very weird to me that the adhesion seems fine, even though it’s underextruding pretty heavily. Guess that again would rule out cold spots. Maybe the lack of a buffer, after all?
There are limits to how much compensation the auto mesh leveling can provide. With the fault being on one edge, there is a good chance that manually levelling the bed can get it close enough that auto levelling can fine tune the rest.
Good tests, I think you’re in the realm of bed tramming (manual level) now.
If you’re REALLY avoiding the tram process:
-
Try rotating the plate (like, clockwise 90, see if the bald patch travels with the rotate) and/or using a different plate to see if the problem is a banana plate or the whole tram that needs it
-
Back in the day, we used to put layers of masking tape under the plate to add 0.25mm layers of thickness and resolve minor bed mesh unlevel problems
If you’ve only got one plate that you use… You might be able to get away with some quick and dirty shenanigans by doing the same. See the following horrible hacked together example of lifting +0.50 in a corner…
Good luck!
1 Like
Oh yeah! I remember haha, I’m coming from the good old Creality CR-10 days! I actually did try the masking tape trick already (though with aluminum foil instead), in addition to swapping plates, but something weird happened while I was running tests;
One of my test prints suddenly went from this…
To this!
Not 100% perfect, but this seems very fixable with some flow adjustments.
…So what happened? When removing the foil from under the bed, I noticed the magnetic sheet got a bit smudgy from all the messing around. I cleaned it with some isopropyl Alcohol… and that seems to have fixed it??
I did another production print (the one I got stuck with when encountering this issue) and it came out nearly perfect. To be honest, normally I’d be extremely skeptical that the tiny amount of finger grease could affect the plate that drastically, but it does seem to be the case… I went back and dropped four fingerprints on the magnet just to see how it would affect the print, and this time the print didn’t even stick at all in that area! Cleaned it again, and back to perfect…
So I guess to everyone out there struggling with this issue; at least give cleaning the magnetic sheet UNDER the buildplate a try as well?
The buffer I believe acts like a shock absorber for the speed of the AMS feeder drive motor, the AMS Internal Hub Motor, and the extruder drive motor, to give finer control of filament flow/pressure in the extruder nozzle.
AMS main functions and workflow
1 Like
Yeah, that’s what I thought too! But does it also do anything when there is no AMS? Like, does it physically even do anything then, or does it just get disabled when selecting external filaments?
Because I would say if it does still work with external filaments, it might also make regular filament spool pulling more reliable. I don’t actually know though, just some thoughts.
(And that didn’t seem to be the issue in my case, at least, after all!)