New to the Process: Calibrating the Machine

Enable retraction

You can use the default settings in Snapmaker Luban. And the work speed should be low so that the dwelling time is long enough to retract the filament.

Edwin

Good to know. Looks like I have a little reading to do on GCode.

Edwin, that was one of the first things I did after my first print.

Clewis, I used untergeek’s tower, and along with a 30 minute job cleaning the bed after the print because the initial layer was so thin it would not come up, this is what came out:

I’m not sure I’m reading it correctly though. Following the large loops up and around, and looking for the best print area, I seem to need a full 6mm retraction, and (counting the loops up, starting with the default 10mm, and going up by 10 per increment), it needs to retract at 200mm/s. That can’t be right, can it?

I think I might need to try another test (but with something else, because I don’t want to have to scrape my bed for another half hour).

I wouldn’t give that test any credence, as it is not a proper test. Retraction tests are rather tricky, and the majority of the ones that I’ve seen are fairly useless. For the direct drive extruder on the Snapmaker 2.0, something in the range of 0.4-1.0mm is what you should use. Different filaments may need different retraction, like PETG. It will work better with a 0.8-1.0mm retraction, while a good PLA (not Snapmaker) will work great with 0.4mm.

The better tests only test stringing, and don’t use paper thin walls or needle point towers. As a teacher, this might even be a good exercise for your students.

BEST OF LUCK!

1 Like

Could you please send me the STL file and G-code file so that we can try to reproduce this issue at our end.

As @WilliamBosacker wrote, the retraction test is rather tricky and we need to analyze this problem with certain models.

Cheers
Edwin

Since I started to use a dehumidifier, I no longer get ANY strings at all.
Could be that is contributing to the issue…

I’ve been sharing the test results with the students, and we’ve been talking about them in relationship to what we known about the process. We’re also printing out a small 15mm miniature, for comparison purposes, with each tower we’ve printed (see below). What test would you recommend for just retraction issues?

Here are the GCode Files for the various tests I’ve been running. Test_01_01 and Test 03_01 are retraction test towers, and the 01_02 and 03_02 are small game figures. I’ve run them both at 190, 195, 200, and 205 degree, using:

  • 60mm/s travel speed
  • 1mm Retraction distance
  • 60mm/s Retraction speed.

The best result I’ve gotten at this point, using the Snapmaker PLA included with the machine is below. As you can see, I’ve moved from solid threads to a few smaller threads and lots of fine webbing, so I guess that’s a bit of improvement:

Test_01_01.gcode (407.6 KB) Test_01_02.gcode (1.5 MB) Test_03_01.gcode (380.2 KB) Test_03_02.gcode (1.5 MB)

It’s pretty dry in here, but I can give that a shot.

I use this one: https://teachingtechyt.github.io/calibration.html#retraction

My settings are:

  • Retraction: 1mm @ 60mm/s
  • Z hop of 0.5mm

I believe @WilliamBosacker uses 0.4mm retraction

I thought it was dry here in Denver, but I bought a hygrometer and found out the RH was up in the 80s and 90s in the last couple weeks! I’m running a Sunlu dryer box at 55C (for ABS) on top of the enclosure, feeding through PTFE tube to a Thordsen inline dryer at 90C, and from there through a hole in the top of the enclosure, right over the middle of the X rail, to the printhead in PTFE tube. The roll now turns with no jerking on the X rail, and I can definitely see the difference between dried filament and not dried!

3 Likes

@NTorson while none of the current tests are perfect, the one that @brent113 as provided a link for, covers half of the test. A full and complete test should use retraction, detraction and Z-hop. With PLA the Z-hop can be fairly minimal, if not removed entirely. However, with PETG you definitely should use a Z-hop. Also, different brands of filament act differently, so you need to take that into account as well.

I haven’t had the time to do any Z-hop tests, but I have dialed in my retraction and detraction settings pretty well. Here is what I am currently using:

 BRAND   | Type |   Retraction   | Detraction
OVERTURE | PLA  | 0.4mm @ 65mm/s | 40mm/s
OVERTURE | PETG | 0.8mm @ 65mm/s | 40mm/s
OVERTURE | TPU  | 1.2mm @ 40mm/s | 25mm/s

These are good starting points, and once I have time to do some Z-hop tests, I’m sure that things will even get better. I have pretty much been using PETG exclusively, as I like it’s characteristics better than those of PLA. I’ve started using ABS-like resins and my Elegoo Mars Pro, for printing functional parts that are almost exactly like extruded plastic parts. They also have extremely high temperature resistances, which I need for the headed enclosure of my Ender 5 Plus.

I hope this helps you and your students. Who knows, but your students might even be able to write a program that builds the GCode to test this. :wink:

P.S. I looks like Teaching Tech added “Prime (unretract) speed (mm/sec)” (detraction) and Z-hop to the test, so you have everything together now.

1 Like

I’m did some tests of this one this morning. I ran them at 190, 195, 200, and 205, with the Snapmaker PLA (unfortunately, that’s all I have, at the moment, having shot my budget on the machine and enclosure). I’ve set the retraction parameters for:

1mm Retraction
60mm/s Retraction Speed
0 Extra Restart
40mm Prime Speed
0.5 Z-Hop

All the resulting temps had ‘hairy legs,’ to some degree, but really, there was little real difference between the various temperature ranges. Pretty much all of them came out looking like this:

So it’s clear from this particular test (thanks, Brent) that retraction isn’t really my issue. So I’m kind of stuck at why I can’t get a relatively clean print on my other models (like the ones below) using the same retraction settings. Is there something else in this GCode that I’m missing?

I will say that I don’t believe the 4 spiral test is a particularly valid test, and it definitely should not be used on a printer that hasn’t been dialed in. Think of it this way, we know that FDM printing is not particularly good at printing that type of pointed peak, and it is extremely unlikely that anyone is going to use it for that purpose. So, is it a valid test?

The other model looks melted, like the nozzle temp is 215°C or higher. It also looks like the layer height is very large. Was this printed using the Fast profile in Luban? Tests should always be performed at Normal, or slower settings. When I zoom in, the layer height of the print on the right is more than twice that of the print on the left.

For the 2 prints with stringing, what did you use to slice them? Is it Luban, as you describe in your first post? If so, that may be the issue. Try using either Ultimaker Cura, or PrusaSlicer. Ultimaker Cura version 4.9.0 and later, now have a Snapmaker profile that is built-in, which was created by a community member. PrusaSlicer seems to work slightly better for me, but you will need to create a profile yourself from scratch. PrusaSlicer only accepts profiles from the manufacturer, and Snapmaker hasn’t provided one as of yet.

While Luban does use a very old version of Ultimaker Cura, approximately 80%+ of the settings are hidden and can’t be changed. Also, my experiences with the Snapmaker filament were that prints would look great when printed at ~205°C, but the layer adhesion was terrible. At 215°C the layer adhesion would be close to what it should be, but the appearance would be bad, and stringing would begin at that temperature as well. However, the filament that you have may be rebranded from a different vendor.

Since the latest retraction test came out very well, I have a feeling that changing the slicer may address your issues. If it doesn’t, then we need to eliminate the filament as being a possible cause of the issues, by using a filament that is known to be good. A 1kg spool of a good filament will run you around $23 from Amazon. For these tests, you should only need a maximum of ~250g. There may be some 1kg spools available in the $16 range, but they could also be pretty bad, so it’s hit and miss with them. Also, I would recommend using a neutral color like gray for testing, as it is much easier to see the problems, and dramatically easier to photograph.

Because of the retraction results, I don’t believe that the filament is our current issue.

Thanks for the feedback. I ran the test, and the results where counter intuitive to me. I wasn’t sure how useful it was, but it seemed to be useful. I’ll stop linking back to it.

Point taken (no pun intended).

I’ve kept it between 190 and 205. Nothing higher. Typically I do 4 print runs of the same gcode at 190, 195, 200, and 205 to get a good idea of which temperature is working best. And I am using Luban, but with a custom copy of the Normal Print setting, modified for 1mm/60mms retraction and 0.5 Z-hop. The layer depth is set to .15mm after an initial layer height of .3mm.

Again, I’m using Luban, but that is mainly because it came with the machine and I haven’t the budget for a new program, and the bureaucratic hassle of getting a new piece of software added to the computers in a public school is a nightmare (it took me almost two years to get them to add Adobe Creative Cloud rather than continue using Creative Suite 6). So let’s assume I’m stuck with that for the moment.

I agree. I started to print out a set of miniatures (I plan to use these for my Game Design class), widely spaced apart at 205 degrees, but immediately got some serious webbing issues that were not present in the previous retraction test prints:

I tried to adjust the nozzle temperature down to 195 during the process, to try and correct for that, but the nozzle never cooled down to more than 204, so on the fly adjustments don’t really work.

The sad thing is, before I updated the firmware to make up for the extraction error issue, I got one good print out of the machine (this was a few weeks ago):

As you can see, no spikes, everything printed with a clean separation, and minimal webbing between the antlers (which I would expect). Once I updated, however, everything started getting spikey, and the best I’ve been able to do since then is to dial it back it to the point where it looks like I have a machine full of spiders.

I’m looking at other colors of filament as we speak.

As an aside, I am totally new to this process, as I have said, and the folks on this forum have been fantastic. Much obliged to all of you for helping me through this…

1 Like

It’s prudent to be mindful of assumptions being made.

Unless you’ve verified the thermistor is reporting the temperature accurately, that’s an assumption. The exact number is not important, and a temp tower (like the one from Teaching Tech) is necessary to find out optimal temps for your machine.

On my machine 205 is the minimum acceptable temperature for the layers to fuse while printing. On yours it appears the model is sagging from the heat. Indeed, this has wider considerations where some models may need different cooling or different nozzle temperatures to have better layer adhesion and strength or less sagging or better bridging, etc.

From the perspective of ‘continuous improvement’ your examples of ‘good prints’ seems to have several aspects that can be improved. I’d be interested in which model the bottom middle print is - some of those surface pocks seem plausibly in the model, but if not would be indicative of serious slicer or firmware issues.

Regarding the stringing - have you tried replacing the hotend? Rarely, but sometimes, a burr or machining defect in the nozzle can result in unusually bad stringing.

1 Like

@brent113 has made some good points, and there have been some new updates, which I discuss below.

Ok, from what I see here, I think I have some good news. First off, a firmware update was just released, and it corrects many of the issues that the previous update created. You should upgrade the firmware ASAP, before proceeding. After reviewing the latest pictures and information that you provided, along with something that I just learned about the new 3D Printer Module, we first need to determine which 3D Printer Module it is that you have. Snapmaker apparently started shipping the new 3D Printer Modules recently, and there is a design flaw in the part cooling of the new 3D Printer Modules.

They apparently changed the part cooling air flow, which results in the right side not being cooled. On the bottom of the 3D Printer Module there is the nozzle in the middle, a long cooling vent on the left, and 2 blocks on the right. The larger front block on the right is either a solid block, or a vent. The original parts have a slit for cooling, while the new parts are just a solid block. Which do you have? This will determine which way we need to proceed.

You might want to try one more print with Z-hop configured in Luban. This might make a big difference, and otherwise allow you to move forward. We’re here to help, and it looks like we’re very close to a conclusion. I fully understand the limitations that you need to work with in a school environment, and hopefully both you are your students will fully operational very soon.

All this being said, Luban will probably be fine for your environment. However, some more good news is that Ultimaker Cura and PrusaSlicer are free for everyone to use. Read the licenses and see if they work for your environment. Neither require licensing fees, and there may be special offerings for educational environments. I would recommend that you evaluate both of them on the computer that you are using, and to not install them on the computers that the students are using at this time.

1 Like

I’ll put that on my list of test prints.

I’m trying to knock out issues one at a time, while trying to understand the principles behind the various settings. Once I get the stringing reduced or eliminated, I’ll definitely look at those.

That would be a last resort (out of budget for the year), but if nothing else works, I can try that.

I updated the firmware when 12.0 came out. That’s when my prints started getting hairy. As for the module, both sides have a slit/vent. It’s also very dirty under there, with some sort of splashback/splatter on the area around the nozzle. How do you clean that?

Z-Hop has been configured to 0.5mm on the previous prints.

That is helpful! Right now, I’m running the printer from my machine only, so one is all I need, and one free bit of software is much easier for me to swing.

I guess I had a different thought: I usually troubleshoot by starting with the settings that are tied to physical things, the artifacts I was seeing are indicative of issues with that.

Once the machine is physically doing what it’s commanded to, then one can move up to the next layer of abstraction and begin turning other knobs in the software.

The fundamentals are extruder steps/mm and flow multiplier.

If there are any issues with those, sometimes even endless tweaking will not eliminate issues.

It should be one of the first things you try. There’s one in the toolbox that came with the machine. It’s free and takes about 2 minutes to do.

If having replacement Snapmaker hotends on hand is cost prohibitive, having just spare nozzles on hand is fine too, and they cost in the ballpark of 25 cents each. Changing one takes just a few minutes.
image

1 Like

I thought that was what I was doing. I updated the firmware to fix the extraction error, modified the steps/mm, and tried various temperature settings, as many folks have suggested. I’m pretty sure that I’ve now eliminated Retraction as an issue, as I no longer get thick spikes and strands, only the webbing.

So the question is, where to go from now. You mentioned flow multiplier, and I can see a Flow Percentage in the Material settings (which is currently set at 100%). Tell me more about that.

I totally forgot about the spare, until the moment you mentioned it here. So that I can do. And it looks like I can get the hot end replacement kit from Amazon, so that’s doable (we are restricted to specific vendors when we purchase, which is why I had to wait to for the SM2.0 to arrive on Amazon before I could purchase it, costing me several hundred extra dollars).

Are those MK.8 nozzles compatible with SM2.0? And how do you replace just the end of the piece?