Multiplier extremely low

K8200 3D Printer K8200 - Issues while printing
1

Hallo everybody, I have my K8200 ready for a couple of weeks now. After printing a not so nice case and some ugly little test objects I have read about the need to calibrate the extrusion multiplier. Sounds not so unusual so far. :wink:

For this I use a simple box which I slice with 1 perimeter, no infill and no top layer at a layer height of 0.25 mm.

This led me to lower multipliers and my prints do get clearer now, but now I am down to 0.69 which seems to be extremely low. When I reached the first time a multiplier of 0.75 I changed the stock hubbed bolt against one which is said to have a good grip. I thought there was perhaps an irregularity with the stock bolt, but there was no great difference after changing.

Is it possible to have such an extraordinary low multiplier or is it sure that there is something wrong with my extruder.

I really have to now this, because I don’t want to spend hours with trying to optimize a faulty machine, so any help or hint is very appreciated!

Thanks for you attention.

2

Hi walo,

before the extrusion multiplier calibration, did you

  • calibrate the E steps of the extruder (especially with the new transport screw), and
  • measure the real diameter of the filament?

Cheers,
kuraasu

3

Hi kuraasu, thank you for your replay.

No, i didn’t calibrate the E steps, but I was wondering about having “too much” filament, and this with the stock screw, so I did not think about the new screw, which has a smaller effective diameter.

OK, so today I did the following:
I let 100mm of PLA extrude at a speed of 30 mm/min, 95 mm were extruded.

So I changed the E steps value in configuration.h of the firmware (Marlin V2) from 600 to 630 and uploaded it.
Again I let 100 mm extrude, now the result was 99 mm.

With this setup I printed my calibration cube with an extrusion multiplier of 1 and a layer height of 0.25mm, slic3r calculated an extrusion width of 0.42mm. With the micrometer I measured values between 0.78 and 0.83 mm with an average of 0.80 mm.

I changed the multiplier to 0.53 and printed the cube again. Now the result was a wall thickness betwenn 0.52 and 0.6 mm and didn’t look so good, the walls have many holes. Seems as there is to little matter, even if the calculated extrusion width is not reached.

The filament diameter used to be 2.95 or sometimes 2.90, today I found 2.80 mm and used this value for printing.

Am I doing something wrong or this there a problem with my extruder, perhaps with the nozzle?

Well, for tonight I have to give up. :frowning:

Greetings

walo

4

Last weekend I again wasted time and filament by printing useless one-walled objects. Somehow I don’t get ahead with the extruder calibration. Yesterday I tried to measure the diameter of the nozzle with a jet gauge. 0.45mm went fairly good, with the 0.50mm pin I couldn’t get through, and I didn’t want to use force.

Perhaps I should disassamble the nozzle and clean it?

How do you clean the nozzles of your printer? I have a blowtorch and an ultrasonic cleaner here, could I these for cleaning the nozzle?

Greetings

walo

5

Hi walo,

I still don’t get it your extruder seems to extrude twice as much filament as it should.

The nozzle can only block the filament, not accelerate it. Still, you can of course clean the nozzle. The gauge is already a good idea, otherwise a piece of wire with matching diameter (someone suggested a guitar string for this) can be used. If you print with light colors, you can usually see tiny dark dots in the printed material when you have issues with material burning up inside the hotend. Cleaning PLA can be done with a blowtorch, yes, or some (quite nasty) solvents, so usually people tend to opt for the torch. Acetone does not dissolve PLA, but let it swell, so it’s useless here. For ABS though, it’s a good choice as solvent.

For the original hotend, I have found that cycling a few times between pushing and retracting filament is good for cleaning the hotend volume of residual “old” filament.

As for your initial problem, can you retrace and doublecheck the steps for the calibration print?

  • thin wall stl file with known wall thickness (a), or, better: solid stl file (b)
  • slice with (a) same extrusion width as the wall thickness or (b) fixed extrusion width to, say, nozzle diameter +/- 10%, no infill, no top solid layers, 1 perimeter
  • for PLA: auto-cooling enabled, especially if the object is small
  • filament diameter correct (point instead of comma?)

Is the Z height ok, or did you by any chance replace any of the Pololus with a version with different microstepping, e.g. DRV8825?

Cheers,
kuraasu

6

I had similar results when I calibrated my extruder. My E steps had to be changed from 600 to 705 to get the right amount of filament and to improve my 25x25x25 single walled test cube to something closer to what the STL showed.

But I had to drop my multiplier to about 0.78 which didn’t seem right to me. I also suffered bad prints, the finish and infill would look awful like it was not getting the correct amount of filament.

I gave up and reset it all to the default settings the printer came with and all is fine EXCEPT any holes are slightly under the size I draw up.

I will revisit it again when I don’t have any prints to do.

7

Hi kuraasu,

[quote=“kuraasu”]I still don’t get it your extruder seems to extrude twice as much filament as it should.[/quote]Yes, this is what obviously was happening when I did the first prints after assembling my printer. Many strings, no finer structures. The I have read here in this forum about the need to calibrate the extruder and I thought “No problem, this will be done in one hour”. Well, that was at the last days of January.

[quote]The nozzle can only block the filament, not accelerate it.[/quote]I have bought the gauge, because I thought that the nozzle perhaps is too big. In fact, I was a bit disappointed that it isn’t, because I still have not found the reason for my problem.

[quote]- thin wall stl file with known wall thickness (a), or, better: solid stl file (b)/quote usually a coboid of 30x30x*15 mm.

[quote]- slice with (a) same extrusion width as the wall thickness or (b) fixed extrusion width to, say, nozzle diameter +/- 10%, no infill, no top solid layers, 1 perimeter[/quote]Again (b). Extrusion width is set as in the velleman .ini to 170%. Layer height is 0.25mm, sometimes 0.30mm, so slicer calculates an extrusion width of 0.42mm (with the 0.25mm height) or 0.51mm (with the 0.30mm height). The other parameters are set as you said: no infill, no top solid layers, 1 perimeter. After printing I measure the topmost layers with a micrometer.

[quote]- for PLA: auto-cooling enabled, especially if the object is small[/quote]Yes, I’m using the auto-cooling feature. The possibilty to autmatically reduce speed at layers with short printing times is very usefull with some kind of objects.

[quote]- filament diameter correct (point instead of comma?)[/quote]Yep.

[quote]Is the Z height ok, or did you by any chance replace any of the Pololus with a version with different microstepping, e.g. DRV8825?[/quote]No, the printer is assembled as it came from Velleman, except the glass plate with kapton tape.

So I will clean the nozzle this evening by burning it out and later re-assemble and test-print again. I will report here if something has changed or not (can take 1 -2 days).

Thanks for your hints so far.

Regards

walo

8

[quote=“Moorron”]My E steps had to be changed from 600 to 705 to get the right amount of filament[/quote]Seems quite a lot. Did you use the original transportation screw or an improved one?

[quote]But I had to drop my multiplier to about 0.78 which didn’t seem right to me.[/quote]This was the range where I was with the stock screw and 600 E steps. Setting the E steps higher and lowering the multiplier afterwards, didn’t this lead back to were you came from?

[quote]I also suffered bad prints, the finish and infill would look awful like it was not getting the correct amount of filament. [/quote]Here the same. With a common multiplier of 0.9 I never knew that infill has a kind of structure! There was only smeared stuff inside the perimeters. With 0.7 I recovered unseen details in my prints.

[quote]I gave up and reset it all to the default settings the printer came with and all is fine EXCEPT any holes are slightly under the size I draw up.[/quote]Me too. Deleted all my INIs and started all over with the originals. Only the E steps stayed at 660 as kuraasu told me to do, because I have another screw and this seems logical to me. Of course this makes more filament coming out, not less.

[quote]I will revisit it again when I don’t have any prints to do.[/quote]I’ll be here. Just going to clean the nozzle now.

regards

walo

9

Hi walo,

that (0.3 mm, 170% / 0.51 mm) can be a troublesome setting for the original 0.5 mm nozzle.

The reason is the way how Slic3r actually “designs” the extrusion and calculates the volume that shall be extruded. I have noted on multiple occasions that path generation especially for infill fails when these conditions (extrusion widths just slightly above the nozzle diameter) are present.

In other words: if you see missing infill or layers in G-code from Slic3r, check the extrusion width.

How many layers? Can you rule out Z wobble as a possible cause for the extrusion seeming too wide? A calliper is perhaps not as precise, but with its sharp tips you can measure a single layer. And rather make the part a bit bigger, so the time per layer is increased. You can manually stop the print after a few wall layers, in order to not waste too much filament.

Also check the temperature setting (thermistor placed correctly?), perhaps also try to “calibrate” the extrusion temperature for your orange filament. Some temperature-related printing errors look quite similar to over-extrusion, so it’s possible that this is part of the problem.

Cheers,
kuraasu

PS: which region in Germany are you in?

10

[quote=“kuraasu”]that (0.3 mm, 170% / 0.51 mm) can be a troublesome setting for the original 0.5 mm nozzle.

The reason is the way how Slic3r actually “designs” the extrusion and calculates the volume that shall be extruded. I have noted on multiple occasions that path generation especially for infill fails when these conditions (extrusion widths just slightly above the nozzle diameter) are present[/quote]See what you mean, so I’ll start with 0.4mm layer height now.

[quote]How many layers? Can you rule out Z wobble as a possible cause for the extrusion seeming too wide? [/quote]Well, the micrometer can catch 8mm and I used them all. And yes, I have a Z-wobble. If I drive the z-axis up und down the z motor moves a bit, cause the rod is not totally straight. I saw this, but I put this back under “fine tuning, will care about this later”. So you can see a kind of “banding”. The first layers are very dense until the print reaches a height of perhaps 5 mm. Then follows a band of “lighter” layers, again 5mm. After that again a dense band of 5 mm. Well, my cuboids have only 15 mm, so the last band is what I measured yesterday. When taking up the multiplier, the walls get more massive and the diverences disappear, only the first band is to see a bit then, perhaps the pressure in the extruder still is higher because of the bottom layers.

[quote]And rather make the part a bit bigger, so the time per layer is increased.[/quote]Will try 40x40x20 then.

[quote]You can manually stop the print after a few wall layers, in order to not waste too much filament.[/quote]Usually I let it print to the end because of the first, more massive layers. For cooling I let slow down if the layer printing times is below 30s, minimum speed is set to 10 mm/s, which appears to be really slow.

[quote]Also check the temperature setting (thermistor placed correctly?),[/quote]Because I was not so happy with my capability to place the thermistor in its hole, I wrapped a layer of PTFE-tape around the heating block, holding the thermistor tightly at its place perhaps. Worked very good for me.

[quote] also try to “calibrate” the extrusion temperature for your orange filament.[/quote]OK, so my hot-end is clean now, I switch back to the black Veleman filament, from which I still have have a spool here.

[quote]PS: which region in Germany are you in?[/quote] Frankfurt am Main

I burned every rest of filament in the nozzle and the metal tubing. Trying the ultra sonic to get rid of the rest of the coal did’nt work. I scratched ou the burned residues with a piece of wood and filament string, until everything was clean. The 0.5 mm gauge still didn’t fit through the nozzle, perhaps it was too cheap?

Printer is heating up now, I will run over to my neighbour and see if I can get a digital calliper.

Will report soon

walo

11

Hi walo,

no, stay lower. Unless you want to print only very rough, coarse parts, that is. 0.4 mm layer height is at the upper limit of what’s possible with the 0.5 mm nozzle.

Rather set the extrusion width to a different percentage, or enter a value in mm, e.g. 0.5 instead of the calculated 0.51 mm.

Cheers,
kuraasu

12

Hi kuraasu,

[quote=“kuraasu”]no, stay lower. Unless you want to print only very rough, coarse parts, that is. 0.4 mm layer height is at the upper limit of what’s possible with the 0.5 mm nozzle.

Rather set the extrusion width to a different percentage, or enter a value in mm, e.g. 0.5 instead of the calculated 0.51 mm.
[/quote]See what you mean. Not very pretty and rather useless, because from yesterday’s experiments the nozzle diameter was still set to 0.45 mm.

But I have a digital caliper now (even if it is from Lidl) and I find significantly lower values than with the micrometer.

In the moment I have the next print running, as you suggested I went back to a height of 0.3 mm and set the width to a fixed value of 0.50 mm. Before I forget: I took the multiplier up to 0.8 as a starting value.

Results will follow in 20 mins, and then I will have to shut the printer down. :frowning:

13

That’s a real problem in this calibration. Measuring only one layer should give more accurate results, but keep in mind that there may still be wrong values. The wobbling in the print is caused by uneven movement of the Z axis (mostly due to the motor mount bending up and down), so the real layer height varies. Hence, the actual extrusion changes, too, because it is determined by the amount of filament squished out of the nozzle.

It depends of course on the size of the object, but an interval of 5 mm (or 10, for the “restart” of the dense layers) could point to an issue with the transport screw. E.g. if it’s not fully round, or chips of plastic blocking some of the teeth.

I was thinking more of 40x40x5, or 60x30x5, or even lower. But of course the extrusion needs to be stable through those <= 5 mm to get a good measurement, so it depends on your printer.

[quote=“walo”]I find significantly lower values than with the micrometer.
[…]
I took the multiplier up to 0.8 as a starting value.[/quote]
Ah, that sounds already much more likely than 0.5 (if it holds). For later fine-tuning, two digits of precision are the goal (i.e. having the multiplier calibrated down to 1 %, not 10%).

Cheers,
kuraasu

14

Hi kuraasu,

perhaps switching back to the black filament was not such a good idea. The cube looks awfull, many holes in the walls. Looks like the filament is not coming constantly out, perhaps the temperature (185) is to low?
Difficult to measure so, but at most places of the wall I get 0.62 - 0.67 mm width.

And sorry, of course I set the multiplier in 1/100 so ist must be 0.80, not 0.8. Getting tired now switching everything off here, including myself.

Will be back tomorrow, perhaps I can take a picture of the cube, or I try printing again with higher temperatur.

Thanks for your help and regards

walo

15

Yes, that’s probably the case. Try something in the range 190 to 195 °C. The usual “calibration” for the temperature is to go down until you see intermittent extrusion, then go up 5 to 10 degrees again.

If you can get your hands on a flex coupler for the Z axis, it’s one of the best additions to the printer. In combination with a stable motor mount (can be printed), the Z wobble can be reduced a lot.

Not quite what I ment; you can enter or leave as many trailing zeros as you like. But if a later test results in say 0.862, enter 0.86 and not 0.9 …

16

Hi kuraasu,

[quote=“kuraasu”]Try something in the range 190 to 195 °C. The usual “calibration” for the temperature is to go down until you see intermittent extrusion, then go up 5 to 10 degrees again…[/quote]Ok, so will give the black one a second chance this evening and try to find a more suitable temperature setting.

[quote]If you can get your hands on a flex coupler for the Z axis, it’s one of the best additions to the printer. In combination with a stable motor mount (can be printed), the Z wobble can be reduced a lot.[/quote]The mount is laying around here for weeks. I didn’t want to fix the motor without a flex coupler, because I was not sure what would happen when all the tension is induced the rod. Well, a coupler arrived this week and next week I don’t have to work, so I will applicate these parts (and others) then. Just wanted to carry out this little thing with the extruder calibration first, ha ha. :wink:

[quote]But if a later test results in say 0.862, enter 0.86 and not 0.9 …[/quote]Now I understand, of course not. I chose a “round” starting value, but after measuring and calculating the new multiplier I always use 2 digits after, technically rounded.

regards

walo

17

This evening I first was trying to find a good temperature for the black PLA if loaded now. The first single-walled cube printed with irregular extrusion and many nasty holes at 195 °C. Some cubes later with 200 - 205 °C the flow was better, higher temperatures did’nt improve anything.
The holes in the wall which is first printed after lay change have benn bigger than in the other walls. Has been better after reducing retraction to 1 mm.
Made a picture of the first side printed after layer change : dropbox.com/s/yci8bqu96s6z1 … l_back.JPG
and the last side: dropbox.com/s/ojz6xw4rfk5fj … _right.JPG
The width of the last layers is round about 0.60 to 0.65 mm. Layer height in slicer still is set to 0.30 mm, extrusion width is set to 0.50 mm and the actual multiplier is 0.80.

I’m really thinking of pausing the extruder calibration for a while and first implement improvements like z-motor holder and flex coupler. Eventually I should switch back to the original transportation screw.

regards

walo

18

Hi,

yesterday I mounted the holder for the z-motor and a flex coupler. Then I printed again the calibration cube of 40x40x20mm with an multiplier of 0.74. This time the wall-width was much more constant, I measured between 0.54 and 0.64mm with less extremes than before. Extrusion width was manually set to 0.50mm

I lowered the multiplier to 0.70 and printed again. I found width between 0.42 and 0.52mm width.

Today, with the multiplier set to 0.72 I was brave enough to print the cube with a layer height of 0.25 and the 170% setting for the perimeter width which resulted in a target width of 0.42 mm.
The cube printed with a wall width of around 0.52 - 0.54 mm, all four walls and all measuring points giving only little difference.

With 0.25 mm height you can see again the banding of the layers: https://www.dropbox.com/s/d5u00qwt1oggagy/20140413_excal_25.JPG
And I still have a lot of holes, you can see.

I switched back to the original transportation screw and will repeat the printings to see if the banding has gone and what is the influence on the width.

The E steps had to be left at 630. I set this back to 600 when remounting the original screw, but then there were only 95mm instead of 100mm filament carried through.

I will provide the results later then.

regards

walo

19

Hi,

with the original transportation screw I came out with a multiplier of 0.9 - 0.92. I still have problems with printing thin walls, I don’t get them smaller then 0.47mm and wall width differs from point to point. It is not easy to measure because of the holes. For testing I set the multiplier to 1.10 and printed with height 0.25 and width 0.42 mm. The walls have minimum 0.60 mm and still have holes:
dropbox.com/s/4ez3g34s362gz … l_em11.JPG
If you watch the printer work, you see the molten filament pouring out of the nozzle constantly for a while. Then suddenly the flow stops, only some little portions coming out while the object moves ahead, which causes gaps in the line. The gaps is covered by the next layer, a hole remains.
So I think the inconstant flow of filament actually is the reason for the problems.

Anybody has some ideas or speculations what causes this behaviour? Is there something I can check? Possibly something wrong with the hot-end? Any help is very appreciated.

regards

walo

20

Well, this problem is solved, if you want to see it like this. When cleaning the extruder again, the copper barrel broke.

Game over.

regards

walo