I read a few places in search for this answer that 1.2 is a desired width as it is not simply placing it straight from the nozzle and then moving along… it gives some ‘smush’ to make sure it adheres to the layer below.
I’m not sure what the 1.2 you’re talking about is… Is it a setting in luban you can control? It seems like you’d be talking about the flow multiplier itself, which if so, you should just set it to 1 and then expect .4mm walls, but that’d only be if luban was configured to do .4mm line widths.
In cura/prusa/super, both line width and flow are able to be calibrated, so you set the width to .4, and then the flow to different values such that the printer is printing .4mm wide lines.
If we don’t even know what width luban is attempting to print, you can’t really calibrate the flow to hit that width. We could just “guess” that they aim for .4 at 100% flow multiplier, but it could easily be configured to be attempting to get to .38 or something, meaning that any calibration you try to do is counterproductive.
Hi,
@brent113 , you mentioned sending Gcode to the machine on the fly, meaníng during print. I failed to do this on 2 levels: First, Luban will hide the console once you are conencted to the machine and will only let you change a pre-defined set of parameters like extruder / bed temp.
However, even when un-hiding the console with the developer tools (luban just sets the corresponding div display value to none, which can be removed), it doesn’t accept the M221 or even the Get Commands. I just get the error message : “Machine is printing now, movement rejected.”
Is there another way of sending the commands?
I’m connected over WiFi , Luban version is 4.4.0
Not sure what you are trying to do @fpsst.
It is possible to print from the touchscreen and editing your settings via console over wifi or serial connection.
The only problem is that your print would maybe interrupted because luban and the machine thinks you are printing from luban.
The machine goes in pause mode after disconnecting, so you need to press resume again (this was NOT tried with the latest firmware or Luban!).
It would be great if you could verify and let us know 
.
Just referring to this one. I’d like to know if there are ways to send commands over the console during a running print.
Luban seems to replace the console with a menu for only a limited set of parameters which are meant to be changed during print.
This question is more out of curiosity than having an actual use case, however the M221 would have come in handy a couple of times in the past
Good day all, first-time 3D printer owner and poster in this community, thus I welcome any feedback.
I found this post after thinking about doing the same, thus sharing my experience. My calibration journey started after I, unknowingly, build my A350T with an enclosure, hooked up the filament supplied with the printer and immediately printed a benchy using Luban’s in-built “Strong and Precise” setting. 4 hours later the filament had snapped, and I had to start the nearly 10-hour print again…
I still remember looking at the benchy after it completed the 9+ hour print and thinking “precise my arse…” and “…this is what I get for 9 hours?”. The print looked terrible and looked like a spider was nesting on Noah’s ark. The image below is after I did some cleaning, rubbing and breaking of the stringy lines.
So, long story short, I considered the money I just spent buying the printer, the warnings I received from SWAMBO (She Who Always Must Be Obeyed) with regards to “…this thing better not become a white elephant” I set off on doing research, reading on this forum, and watching multiple YouTube videos…
Similar to @brent113, I too found Teaching Tech’s calibration video, however, I was extremely hesitant to attempt any of his code, not having any gcode experience (and his warning didn’t help), until I watched his video “Superslicer in built calibration guide for better 3D prints” and I knew that my relationship with Luban may have come to an end.
In Superslicer V2.4.58.5 the in-built calibrations have the following options:
Thus it does not have the below tests that @brent113 mentions:
- PID Autotune, which from this thread does not seem applicable.
- Stepper Motor Drive Current, again not relevant
- Acceleration tuning – I have not done this, and am just using the defaults at this point, I will look into this in future
- Linear Advance – I have not done this, and am just using the defaults at this point, I will look into this in future
The process that I followed is shared below, if you are looking for more details or have comments please feel free to reach out. Considering my and @brent113’s experience, I am considering doing a proper right-up of all calibration items for a first time SM 2.0 owner that are recommended to prevent your first benchy from breaking your heart. This however will not be this year as I have some catching up on prints I need to do.
-
E-steps calibration: Similar to everybody else my printer was set to 212.21. Following the guide on this forum: Summary: Extruder calibration a must I was able to determine my new E value to be 230.
During later stages, and as my confidence in gcode grew, I noticed that the filament diameter was set to 3.00mm, and thinking I was going to one-up everybody, I changed this to 1.75mm, redid the test and found my E value needs to be 679.9. I then started doubting myself backed out and changed it back to 3.00mm. This change seemed to screw everything up as entering the gcode G1 E100 F050 then only delivered 20mm of filament, and regardless of what I did I couldn’t get it “proper” again, so in came the factory reset, set E to 230 again, and another lesson learned in humility. -
Bed levelling: So previous lesson on humility learned out the window, I found this video on YouTube of how to do a 10 x 10 grid auto bed levelling using the gcode G1029 P10, and again thought this must be beter than what everybody else is doing.
Man (or woman) did I struggle… Not so much with the gcode but rather the final step of the calibration card and (to this day) I am still not sure I have been doing it right. I then read on the forum (Somewhere) that the bed had to be heated before doing the mesh levelling process, so using either gcode M190 S65 or the touch screen I pre-heated the bed, unaware that using the G1029 gcode does not retain the bed temp, and I only saw the bed temperature dropping after the 10 x 10 grid was roughly 40 points in…(Note: there is a way to modify the software discussed in the forum to allow for the bed temp to remain on, but I did not go with this. More information can be found here and here.)
So, in the end, I gave up and just did the 5 x 5 grid with heated bed level from the touch screen, as I still do not have the faith in my gcode ability, or could not find a proper guide on how to do a manual mesh levelling with a heated bed. From reading other places in the forum, this may be a potential upgrade to a future firmware.
Outcome of my 10 x 10 unheated bed:
And the outcome of the heated 5 x 5 bed:
Fittingly, as it is so important, this is the area that I still consider to be the weakness in my calibration and something that I would like to continue to improve on and welcome any advice. Some ideas would be:
• Shims to improve the level
• Glass bed – which brings its own set of issues
- Flow calibration: @brent113 has more knowledge on this topic than I was able to learn these past few weeks, and potentially his calibration process is much better. The Superslicer process is visual only, where you print several items generated by Superslicer and then inspect and decided on the value you deem most fitting.
Something to note here, which isn’t necessarily clear in the description or the Teaching Tech video, and I got caught out on this, is that the second iteration of the flow test (2% steps) is downwards only, unlike the 10% steps which are up and down from the current value. Thus, if you consider your “decent” value to be between 0 and -10 (like me) then you should go with the upper value for the 2% run (in my case, the default value had to remain unchanged). It turned out that I did not change this value and kept with the default.
- Temperature tuning: So, a slight difference between the Teaching Tech and Superslicer, the Temperature tuning is suggested before you do retraction calibration, this is because Superslicer’s retraction calibration has the functionality to print at different temperatures, allowing you to investigate what temperatures above and below your calibrated temperature will look like. I believe this is a personal preference, and I just kept with the SS process.
Printing the temperature towers is again a visual inspection with a combination of destructive test/s. For the Snapmaker PLA, the 200deg appeared to have the best results but 210 was very (very) similar, so I decided to go with 205, my thinking that the small temp increase will assist with bed adhesion and compensate for my lack of bed levelling ability.
-
Extruder Retraction Calibration: Following the Superslicer guide I decided to print 3 piles (3 different temperature settings 190; 200 and 210) at 0.1mm per step and 20mm height, thus limiting retraction to 2mm total. From this came a few interesting observations:
a. 210 had overall less stringing than the lower temps, which was the opposite of what I was expecting
b. The seam on 200 was noticeably worse than 210 although the stringing was very similar
c. For some reason the stringing ended fairly early (0.5mm up) but then recommenced at the upper levels, which I cannot quite explain
Noting the above, I then changed the following values:
• 70% Retract amount before wipe
• Length 1.5mm
• Retraction speed 40mm/s
• Deretraction speed 20mm/s
• Minimum travel after retractions 2mm
• Retract on layer change (ON)
• Wipe while retracting (ON)
• Wipe inside at end (ON)
• Wipe inside at start (ON)
Noting the “poor” seam I also did some quick searching and found somewhere (may be from Teaching Tech or Makers Muse’s YouTube channels) that an “Aligned” seam delivers dramatically improved seams for both cylindrical and edged shapes.
I then mistakenly (maybe because it was 02:00am) I reprinted the towers form Superslicer, thinking that my settings will prove if it worked or not, which I realised after I hit print, that it would not, as it constantly changes the retraction value (as part of the test). This, however, is only partially true, as I noticed a significant improvement on the seams which I believe is a result of the wipes and the “Aligned” value.
-
Bridge flow calibration: This, again, is a visual test, and to be honest I was not particularly impressed with the results (in terms of quality, not necessarily capability) but in the end decided on a 90% Bridge flow ration.
-
Ironing Calibration (Over Bridge flow calibration): For this test I was “confused” to be honest. My test results delivered, what I can best describe as “random” results. Example, 100 seemed reasonable, 105 was poor, 110 was roughly between 100 and 105 in quality, 115 was on par with 105, 120 had the best results and 125 was second best. By this time I was feeling like a calibration so decided to just select 120 and move on, but I am still perplexed by the randomness.
-
Ironing Calibration (Top flow calibration): The world started to make sense again as I had a linear improvement in results, and 120 again was the “winner”
So, calibration is done and I’m free to print the world, or at least so I thought. I came to realise:
• I have done all my calibration on the “normal” profile, and I have no idea how some of the results will change for the (in Superslicer) “high detail” or “optimised” profiles, which, from what I can determine is predominantly meant to be the same expect of layer height (resolution) and printing speed. Recommendation: Copy settings into your other profiles to start with.
• This was done for 1 x filament, some of these calibrations will have to be redone for each different filament if results are to be reproduced.
• I printed the last and final calibration cubes (1 x 30mm and 1 x 20mm), the 30mm with my calibrated “normal” profile and the 20mm with the “high detail” profile with updated parameters (thus, just speed and layer thickness are different). I am not sure if it is the reduced speed or the smaller object (= smaller gaps) but at this point, I would not consider the “high detail” profile to have higher print quality. Yes the resolution is clearer, but the edges are not as “sharp” anymore and there are more imperfections
• This showed that my x and y axis is extremely close but my z-axis has some dimensional issues, which I am yet to resolve (figure out); it is however what lead me to this post as I was searching the forum for XY calibrations.
I clearly have a lot left to learn and potentially more calibration to be done, but I never thought I would be so happy to see a cube in my life… In my humble opinion, it is a dramatic improvement. I wanted to share my experience as well, to help people that may be dejected like I was after my first print, and show the true capabilities of this printer, if properly calibrated.
Secondly, although this is my first post, I have read so many comments on this forum, and it has assisted me so much in my learning that I want to use this opportunity to thank everybody, this forum is invaluable!
1 Like
@Waterboy Some things to consider/keep in mind…
The first step in preventing stringing is making sure the filament has been properly dehydrated. Many filaments are hygroscopic, meaning they absorb moisture out of the air, and this moisture in the filament leads to all sorts of problems with prints (stringing, bubbles, adhesion, blobs, etc). Even when they are brand new, fresh right out of the “dry” packaging, they should still be dehydrated/dried before use.
Ideally, e-steps is calibrated with the nozzle/hotend removed. As it is a linear distance that the electrical motor moves, it is not subject to the resistance of the nozzle/hotend or filament viscosities (these are where flow calibration comes in). This is unfortunately a much overlooked and untaught piece of information.
Retraction can be specific to each type of filament (PLA, PETG, ABS, etc.), the brand, and even color (the same is true of best print temperature setting); chosen print temp can also have an affect on best retraction settings. For example, my eSUN PLA+ Grey prints best with 1mm@10mm/s (at 210 temp), while my Overture PETG Black is best with 2mm@50mm/s (at 240 temp). Retraction is also a type of extruder distance move, so is highly dependent on a proper e-steps calibration.
The linear advance calibration is not entirely necessary if you are running the latest firmware, as the default is VERY close to the ideal setting. It is still a fun/good test to run for learning, but you are unlikely to see much difference in your print. Old versions of the firmware were far off and desperately needed it done, but no so anymore. The default k-factor (linear advance) is 0.04, I calibrated mine to be best at just 0.05; change to prints is unnoticeable.
There are some other setup/calibration steps that can be done, like tramming, backlash compensation, and x/y/z-steps, but these will usually require purchasing extra tools to measure and adjust, and can be time consuming.
In your bed levelling, a delta (difference from highest to lowest points) of 0.4-0.6mm is quite common for the stock machine. Here is an abbreviated write-up of the process I went through for my bed levelling; it made a HUGE difference for me, I was able to reduce my delta to as little as 0.14mm (a mere 0.005in). I am tinkering with trying to get it even lower, but we will see.
https://forum.snapmaker.com/t/tramming-shimming-and-3d-print-bed-levelling/28307