Extruder Calibration a must

The first time I ran the extrude code it only feed about half way through. After I changed it to 419.22 I tried again and it did about the same. I changed it to the value chazr33gtr said to change it to and after testing it feed exactly 100mm.

After setting all that I went to do another print and it wasn’t extruding enough to adhere to the bed and at that point I was done failing for the day and will come back to it later.

I’ll try recalibrating the bed for the millionth time but every time I calibrate to 0.1mm it never seems to make a difference when printing the first layer. I wonder if it was the model I was printing though, but I’ve had the same thing happen with other models and have been been able to get a good first layer.

Are these extruder calibrations only applicable for Luban or do they transfer over if I were to switch to Cura? I’m honestly very annoyed with settings in Luban not being consistent with any other slicer program.

The extruder E-steps are stored in the firmware so the slicer has no effect.
A value of 400 is really not normal. As suggested I would try to swap out the nozzle.
Mine is about 251 for PET-G.

For me the extruder calibration was a real game changer!
The bed adhesion is SO MUCH BETTER now. I had no big problems in the past but on some places the filament just wasn’t sticking well and I had to lower the nozlle so much that I always had an “elephant foot” on my prints. My Snapmaker Originals wasn’t as bad calibrated as the A250 but since I’ve calibrated all my 3 Snapmaker machines all of them are producing even better prints and I don’t have to even look if the first layer sticks to the bed. Only the leftovers from purging sometimes are hanging on the nozzle and I have to pull them away with some tweezers.

I’ve been ignoring this important step of printer preparation for years since my prints did look ok and I just thought that the Snapmaker print sticker simply isn’t as good as the glass bed of my Ultimaker 3 (I’m now selling!). Oh thank you god for letting me read this forum post a few weeks ago!

2 Likes

Thanks for explaining this. I was wondering the same regarding E value as I am still waiting on my A350.

Between 10 and 11 above you’re missing a power cycle. The machine must be turned off and on again for the saved firmware settings to be applied, and to read back correctly. (In my case I tried reading them back before power cycling, they were old, told myself oops I needed to power cycle, did so, and upon re-powering found the settings had not applied, either, perhaps because of the check before cycling[?]).

So to be safe, make the change, power cycle, then get back in with Luban and check. Repeat as necessary until it does echo back the right value.

This is mostly true, but the “flow” value does affect the extruder steps per mm. I I only mention because a lot of people mess with his setting rather then calibrating the machine.

A power cycle should not be necessary, m500 saves the current settings. M503 reports saved settings.

You just love correcting me :wink:
The slicer has no effect on the E-step value stored in the machine except:

  • flow rate
  • custom G-code where new E-step values can be set (maybe usefull for different materials)
1 Like

@rojaljelly its not just you, i love correcting everybody :upside_down_face: but never trust an Atom, we make up everything :joy:

5 Likes

Thank you for the guide!! I was quite close to the example calculation you provided, had to add 13 mm

1 Like

How do you purge to get rid of any clogging filament? I imagine after all the retractions eventually the nozzle will get some residue sticking and blocking a clean flow of material…

Just park somewhere up high, set to extrude temp, and keep tapping the ‘load’ filament button to push clean in. Unless you’re already truly clogged a good unrestricted foot of filament should I’d think carry most debris out with it.

If it’s gotten so bad you need to disassemble and clean that’s another story.

1 Like

Thank you guys for the guide.

I have tried to recalibrate my extruder, and had to do it several times until I decided to stop and come here for help.
I did a test after each calibration and it just wouldn’t stay at 100mm extrusion.

The initial M503 report I received was as follows:

G21 ; (mm)
M200 D3.00
M200 D0
M92 X400.00 Y400.00 Z400.00 E212.21
M203 X150.00 Y150.00 Z50.00 E25.00
M201 X3000.00 Y3000.00 Z100.00 E10000.00
M204 P1000.00 R1000.00 T1000.00
M205 B20000.00 S0.00 T0.00 J0.02
M206 X-7.00 Y-5.00 Z0.00
M420 S1 Z0.00
G29 W I0 J0 Z9.00000
G29 W I1 J0 Z9.00000
G29 W I2 J0 Z9.00000
G29 W I0 J1 Z9.00000
G29 W I1 J1 Z9.00000
G29 W I2 J1 Z9.00000
G29 W I0 J2 Z9.00000
G29 W I1 J2 Z9.00000
G29 W I2 J2 Z9.00000
M301 P13.00 I0.10 D17.00
M851 Z1.00
M900 K0.22
M603 L0.00 U60.00

My first try of calibration: 15mm left out of 100mm
New extruder E value: 249.55

After the calibration, I tried to test the extrusion again with the “G1 E100 F300”

My second try of calibration: 5mm more of 100mm
New extruder E value: 237.66

My third try: 7mm more of 100mm
New extruder E value: 222.11

My fourth try: 11mm left of 100mm
New extruder E value: 252.39

My fifth try: 5mm left of 100mm
New extruder E value: 265

My sixth try: 2.5mm left of 100mm
New extruder E value: 270.3

If base on the latest E value of 270.3, comparing to my initial E value of 212.21, that means there is a 22% of inaccuracy? Comparing to the first calibration which was 15% of inaccuracy?

Am I doing something wrong here?

I don’t know if it was weirdness at my end or not. But I tried to use the M500 (save) command, then looked again (M503), and it hadn’t saved. I did a power cycle and it still hadn’t saved.

So I set it again, power cycled, THEN tried to look again. This time it finally had saved.

You have verified with M503 that it has saved a prior result before trying another time, right?

@rtrski

Yes, I saved every time with M500 and double checked with M503 before trying again.

I decided to reset the E value to the value it came initially, ie 212.21, and restarted the calibration. The result is the same as the first time I did it.

  1. Changed the E value back to 212.21
  2. Recalibrate with result of: 15mm left of 100mm.
  3. The first time when I did the calculation for the new extruder E value, I did a 212.21 x 1.176 (three decimal place). This time I did a 212.21 x 1.17 (two decimal place) and got a result of 248.28.
  4. Changed and saved the new extruder E value.
  5. Decided to do a test with a result of: 5mm left of 100mm. Again, same as the second test previously.
  6. Decided not to change to the new value.

Now am trying to do a print of the benchy and see how it goes.

Yea, you want to be a little careful when measuring/marking 100mm then attempting to extrude that much. When it’s “short”, it’s fine, but if it’s “long”, there’s no way to see (assuming you are measuring from the top of the module like i did).

If you use something larger than 100mm (like 120 or 150) for your markings, you then still attempt to extrude 100mm, you should still be well within ease of measuring.

And just to be clear, each new calculation uses the new E value. it looks like 2Bpencil’s calculations are doing that, though it took a minute to determine that “left of” was under extruding and “more of” was over extruding. Those values are “off” though as you can see in the latest post because of rounding.

speq-extrude-sample

So i use a program call SpeQ Mathematics, which allows you to do all kinds of stuff, like input formulas. You can see here a = E, b = expected extruded value, c = remaining distance to extruded value (so “past it” being negative). So using 2Bpencils’ value of 249.55 is what got closer to 237 for the next E by setting the negative 5 for the “distance”. Again just to help people understand what’s going on.

But yea, you shouldn’t round when you don’t have to, esp not in a formula. So the initial value you should have gotten with 15mm left is 249.66 for your next E @2Bpencil (not 249.55).

I can say after working with mine, then redoing it after swapping hot ends (and finding it over extruding slightly) that if you’ve been printing for a bit, check to see if your nozzle is actually extruding properly. it could have hardened bits in it, the tip could not be proper anymore (esp if you’ve gouged the plate every now and then), bring it up in temp a bit to make sure everything melts properly, make sure the filament isn’t getting “snagged” on anything as it’s being pulled from the spool, etc.

For the stuff I print though, it’s not for any type of business purposes, so as long as it’s “close enough” that’s good enough for me. Hell i have been printing things for a long time without doing this and those models were fine! It can matter which filament you use too. the “better” filaments tend to melt and flow easier, making this less noticeable.

2 Likes

Hi all,

thank you for the detailed guidance,
I’ve tried it myself and am getting some odd results

first attempt was short by 20.5mm
second attempt was short by 10mm
third by 8

after a few attempts Im getting consistent feeding at M92 E309.18, which works out around a 35% error

I aslo checked against the speed at which the snapmaker loads/unloads at F200
and got different extrution distances again.

does the calibration have to be at F300?

it seems that the feed gear can be prone to slipping as well
if I didn’t let out enough filament and it had to pull from the spool I found it would under extrude
and I’d get different results again

Going to try some calibration prints

@StuCol92

I wonder about thew extruder gear… it doesnt look that great, is very small in diammeter and the number of teeth is not high… I think the filament can slip quite easily. I would have thought for the 2.0 version they might have gone for a dual drive setup… but hey ho. I think Im going to customise mine to a BMG / E3D hot end very soon

1 Like

So far for me at least it’s been very effective. Even when I had a roll of filament that kept having tangles/snags/wrapped over itself, it only stopped feeding when I couldn’t even pull it by hand.
There is a problem with the designs feed path that makes flexible filament not work well when loading. A printable adapter someone came up with solves this though.
-S

@sdj544 that was me with the flexible adpater :slight_smile:

Yeah i not had an major problems with it… but i just thought it would be a more ‘high end’ design

You can also calibrate at lower speed.
A extruder step value of 309.18 seems a bit high.
Maybe you should swap out the hot end (or only the nozzle). There might be a partial clog or you have set a too low temperature - can’t say from the distance.

I’ll see if I can home my print settings in with cura, trying print temps between 195 and 205
getting reasonable quality but terrible layer adhesion
Upping the temp helps with adhesion but then overhangs are affected
it probably doesn’t help that I’m using a cheaper brand of filament
Geeetech silk black.

I’m hoping to not have to replace a hit end so soon after getting the printer
replacing the nozzle is a good idea though

I’ve gone back to the default value of M92 E212.21
and my print quality has returned to ‘normal’

this is all good learning and experience for me