Dim3nsioneer 558
You also might want to tag @swordriff directly for getting advise.
Anyway, printing with a 0.25mm nozzle on an UM3 has to be considered as experimental, i.e. one has to find good settings by try-and-error.
Edited by GuestYou also might want to tag @swordriff directly for getting advise.
Anyway, printing with a 0.25mm nozzle on an UM3 has to be considered as experimental, i.e. one has to find good settings by try-and-error.
Edited by GuestHi ! Thanks for posting the question!
I recommend you first try printing it at normal size ( not scaled) and with a layer height of 0.1mm.
When you scale it down, and also have a minimal layer height, (in your photo you state 0.05, which is maybe below what is technically possible= 1 "tick" in Z screw yields more than 0.05mm height?), then you are pushiing it really far!
In your case you have a lot of retrations, and almost no printing inbetween them.
ANY slack in bowden connection; bowden tube is allowed to move as little as 0.1mm, will mess up the continuity of the flowm especially when you are printing such small amounts of palstic.
So when you print and have these retractions, the "same" filament with "a little added" is trying to make its way out of the nozzle.
Bigger nozzle, less of this effect.
My suggestions to make the 025 nozzle work are:
Make sure the bowden tubes are ALL THE WAY down into theur endpoints,
by taking out the horseshoe, push, wiggle, fiddle the bowden tube all the way down,
then while holding it, lift the collett with a plunt knife and insert the horseshoe.
Render the model at 100% size, not 50%.
Layer 0.15 mm
Turn retraction off.
Your speed and temp settings look good!
Now it should print perfectly.
Then, if necessary, turn retraction off.
Print again.
Then, if necessary render the model 0.1- 0.06 LH.
Mind you; the difference between 0.1mm and 0.06 is 40 microns!, and the difference in the print is that
usually 0.1 looks better! this is due to the effect of slack in filament guiding "rail", and the approach of the physical limitations of the machine.
You are now printing a "safe" configuration.
If you add lower LH, smaller model and retractions, the "same filament" is trying to get out of the nozzle, almost nothing comes out, and the filament in the nozzle is staying a long time getting cooked and difficult to print.
Lets hear it!
It looks like 2 different issues:
1) underextrusion on first two photos and also on the final photo at the very top of the print (it stopped extruding completely)
2) The roof supports (room walls) of benchy are - well really bad.
I think problem #2 is because you are printing too hot. 190C is too hot I think. I would try 180C. make sure fans are at full power also. You need more cooling with smaller parts like this. And you might want to print a tower next to benchy for when you get to the top of benchy to let it have time to cool. Also problem #2 might be better if you print slower and with thicker layers. The walls are almost non-existant.
Which slicer are you using? I'm guessing cura 2.* (not cura 15.* or s3d?)???? There are issues with cura 15.* that could cause these if you don't have the settings quite right.
Underextrusion issue: I'm not sure but my leading theory is too many retractions. With a .05mm layer height and with quite a few retractions per layer you are getting the same piece of filament going back and forth through the feeder many times. Potentially 50 times on the same piece of filament. And because you are printing so slow volume-wise you might want to try lowering the feeder spring pressure to the minimum (top value - insert hex driver into the top of the feeder and rotate and you can see the indicating moving up or down - all the way up is the least pressure).
Or just go with thicker layer height. Try 0.1mm. Your resolution in XY is very roughly half the nozzle diameter (0.125) so why not match that in Z and do 0.1mm layer height?
I would start by doing the .1mm layer height and 180C temp. Then if that works, try decreasing layer height and decreasing spring tension. Don't forget to put the spring back to half way when doing .4mm prints though!
Underextrusion could also be clogging I suppose so definitely do a cold pull to remove any particles such as a little piece of teflon or something in the nozzle causing partial clogging. If you have a needle that fits inside the nozzle that would help push any clogs into the filament before you do the cold pull.
Edited by GuestThis issue is a top priority support ticket!
Here is more advice (thanks @gr5 for tripping me on this):
There are practically no retractions in the beginning of the print.
Then it works!
When retractions begin, the same filament is gnawed back and forth in the feeder.
The filament is then thicker, and will not move through the UM3 Bowden, because the
UM3 bowden is not slippery, it is made of PFA, and not PTFE.
Please try this:
Remove filament and bowden tube.
Push "old filament" which has been through the feeder several times manually through the
bowden tube.
Tell us if that was easy or hard.
If it is hard, get a slippery bowden tube withthe right inner diameter.
Waiting...
I finally had some good pregress.
@gr5 and @swordriff , at last i think one of the reasons might be indeed the flow rate being too low + feeding gear too tight. At the beginig i thought the feeding gear might need to be more tighter to give it better grip in extruding filament to solve the under extrusion issue, but no, to the opposite it had to be loosen up to avoid grinding filament which in turn will cause under extrusion. There's no problem with the bowden tube though (i guess).
I can make successful prints without clogs right now, with some bottom-line i found might be helpful to point out:
(i) at 0.06 layer height, print speed 30m/s, flow must be at least 110% to avoid clogging
(ii) at 0.06 layer height, print speed 20m/s, flow must be at least 120% to avoid clogging
(iii) 0.05 layer height is very unstable, 0.06 seemed to be the lowest layer height that gives stable prints.
(of cuz i might be wrong, that's just what I think I get from all the trials and errors on my own machine )
(I miss-typed "line width" with "layer width")
As you can see, on the left is the best print result I can get yet at 0.06 layer height.
However I hv some quesitons:
-----------------------------------------------------------------
I am also testing with 0.1mm layer height prints now at the time of this writing.
It is much more stable and seemed I hv also find the right config to avoid clogged nozzles.
I can now see "ringing" is the most noticeable issue so far on my 0.1mm prints, i hvnt got the chance to took photo yet. I will share the result here soon to ask for advises to improve print quality at 0.1mm layer height.
I am hoping to find 2 sets of print settings at the end of this post in future, one that worked stably for 0.1mm layer height and one for 0.06mm layer height on UM3, so people can have a bottom line to begin with. I think this might help a lot of new comers who uses 3DSolex nozzles, like me
Edited by GuestThe reason slower is better is mostly I believe because the extruder is not so good at stopping and starting. If you have a speed at the "jerk" speed (which is 20mm/sec) or not too much faster (up to around 30mm/sec) then the extruder gets to mostly go at a constant speed (no acceleration and deceleration). Basically the "jerk" speed is roughly the allowed instantaneous velocity vector magnitude change at a corner. So if you go into a corner at 14mm/sec and leave it at 14mm/sec the overall velocity change turns out to be 20mm/sec. I believe it's partly this slowing down and speeding up that causes some of the quality issues. At some point going even slower doesn't seem to improve quality.
I have seen better benchys I think with .4mm nozzle so I'm hoping your .1mm layer height experiment will be even better.
Also in cura 2.* when you set the speed to "30mm/sec" there are probably 4 or 5 other printing speeds that can be things as high as 60mm/sec so I would display all speeds and keep them all at or below 30mm/sec. Also keep the travel speed fast. At least 150mm/sec.
I still think 195C is MUCH TOO HOT. Please try 180C. Please.
Set flow to 100% you should never fix problems by increasing flow rate as it causes all sorts of other problems.
Set temp to 180deg like Gr5 says it's too hot.
Set all speeds to 30mm/s again as Gr5 says there will be hidden ones that will be faster and slower then the one that's shown.
Benchy is u useless model for small nozzle at low layer heights. The strands are small and break easy on overhangs so best to go even slower like 10 - 20mm/s at 175deg.
Listen to Labern! He is the small nozzle printing expert here! I have only used small nozzles for about 5 unrelated prints (that added up to about 20 actual prints).
Edited by GuestAt times when under extrusion happens I thought higher temperature and flow will fix that. I will try now with lower temperature at 100% flow, but most of the print fails when infill happens in my previous attempts, but i will try again.
I noticed in CURA 2.x, infill speed is equal to print speed, while most other speed are half the print speed, so i think you guys are right, 30mm/s is still too fast on a 0.1layer.
I made some successful print at 120%Flow, 30mm/s Speed, 210C Temp, at 0.1mm layer height. However that is not stable, and would only work if Retraction is OFF. I will try what you guys commanded.
Edited by GuestIf it doesn't work then I think you have a problem with the print core. Maybe the PTFE got dammaged or something doesn't line up. It could be some other issue in the drive train. Can You take a photo of your feeder and it's adjustment position.
I will do a print with mine with my hardcore and 0.25 nozzle and post the photo and gcode so you can print it and compare. This way the settings will be correct and will test your hardware.
You will just have to wait for tomorrow as I just started a 12hr shift.
Thanks @Labern , @Gr5 , @swordriff
I just checked my log book (Yes I kept one), my last attempts on the 0.25 nozzle at 0.1 Layer height, 100% Flow, 30mm/s Speed was on the perhaps defective print core.
3DSolex sent me a new one and I am doing a print now at:
0.1mm Layer Height
0.2mm Line Width
100% Flow
30mm/s Speed
180C Temp
It looks good so far and will do one more print to see if it is stable.
Felt like an idiot ha, when I changed so much of the setting before and it turns out how little changes there should be. I continued on testing my new core from the wrong data I gathered on a defective core and perhaps that bought me here (feeling so stupid now). I think things are getting on the right track now. Will post update when I get the print done:)
Here's the fail prints haha:
0.1mm Layer Height
0.2mm Line Width
100% Flow
30mm/s Speed
180C Temp
I am happy with this, or can it be improved?
Just curious.
With PLA, my setting is about 20C lower than the recommended printing temp, so it's 200C - ~20C = ~180C.
If I now start working on ABS, would it be roughly the same? like, 20C lower than the recommended temp and keep every other settings the same?
The print can be greatly improved but this is something you will learn over time.
The temp is lower then the recommended because of the small nozzle, print speed and layer height.
Recommended speeds are normally for 0.4mm nozzle printing from 0.1 to 0.25mm layers.
So if you used the 0.25mm nozzle then yes you would want to drop the temp. But ABS has bad layer adheasion and makes weak prints so I would advise against using it at low temps.
ABS is quite difficult to print. For me it took about 50 prints before I was getting consistent satisfactory results. But you won't get the same quality as PLA I believe. For one thing it helps to enclose the printer to bring air temp up to 35C. Another trick is to use very little fan (about 1% on the UM3 which is about the same as 30% on UM2).
Your benchy is MUCH better but you can improve still more. I see horizontal banding that I associate with Z screw issues. How clean is your Z screw? I would try to remove all the grease and dirt from the top half of the Z screw. Realize that there are 3 helixes and you have to clean all 3. Find some kind of gentle tool (plastic or wood maybe) that sticks into the screw grooves. Try to remove all grease and dust and dirt. Then you can re-apply some grease if you want but probably not necessary (the bottom of the z screw will still have some). One pea-sized drop of grease is enough for the entire z screw.
WD-40 is good at removing grease but can also be messy. If you want to get really serious, then remove the entire screw from the printer and clean it over a newspaper with WD-40 and wipe completely dry and clean. While the Z screw is out you can slide the bed up and down to check for friction in the vertical bearings.
The print can be greatly improved but this is something you will learn over time.
The temp is lower then the recommended because of the small nozzle, print speed and layer height.
Recommended speeds are normally for 0.4mm nozzle printing from 0.1 to 0.25mm layers.
So if you used the 0.25mm nozzle then yes you would want to drop the temp. But ABS has bad layer adheasion and makes weak prints so I would advise against using it at low temps.
So if i am going to tune for a thinner layer, say down to 0.06mm, should I adjust the temperature to even lower, and stay away from adjusting the flow?
I used a repRap Delta printer before and it wont print ABS, so I maybe getting a little ambitious and a bit over excited here about trying to do it on a 0.25mm nozzle.
Would you use ABS on a 0.25mm nozzle or should i be better off forget about it, and be happy with the 0.4 nozzle for stable outputs?
ABS is quite difficult to print. For me it took about 50 prints before I was getting consistent satisfactory results. But you won't get the same quality as PLA I believe. For one thing it helps to enclose the printer to bring air temp up to 35C. Another trick is to use very little fan (about 1% on the UM3 which is about the same as 30% on UM2).
Your benchy is MUCH better but you can improve still more. I see horizontal banding that I associate with Z screw issues. How clean is your Z screw? I would try to remove all the grease and dirt from the top half of the Z screw. Realize that there are 3 helixes and you have to clean all 3. Find some kind of gentle tool (plastic or wood maybe) that sticks into the screw grooves. Try to remove all grease and dust and dirt. Then you can re-apply some grease if you want but probably not necessary (the bottom of the z screw will still have some). One pea-sized drop of grease is enough for the entire z screw.
WD-40 is good at removing grease but can also be messy. If you want to get really serious, then remove the entire screw from the printer and clean it over a newspaper with WD-40 and wipe completely dry and clean. While the Z screw is out you can slide the bed up and down to check for friction in the vertical bearings.
I see, my printer was bought just a month ago and... hopefully it's just becuz me putting it into too much printing. So you think horizontal bending there's the main reason that it didnt went as smooth? could there possibly be other reasons (except Z screw) is causing that?
Never adjust the flow
The only reason your not get so called stable results from the 0.25 nozzle is because you need more time to learn and tweak the settings. This is the reason why UM are not selling it yet as they haven't created the profiles to get good results. Printing with smaller nozzles is not easy and takes a bit of testing.
I just think ABS is a bad material for 3D printing. It's weaker then PLA and if you need temp resistance then high temp PLA like greentec or a co polymer is far better.
You could print ABS with the smaller nozzle but again it's another learning curve and one I wouldn't get excited about.
If you have a strong need for ABS let us know. Otherwise stay away. For one thing it smells quite bad and is probably bad for you to breath the vapors.
So to clarify. Your benchy had a few horizontal lines. It looked like it overextruded there and the filament was spilling out the sides as it printed. I could be wrong there may be other causes. If the horizontal lines stick inward (have holes) then it's underextrusion for a layer or two or whatever. These horizontal lines that stick *out*, if they happen several times, could be a sign that you have bad temperature control: when the filament is hotter it flows better and sticks out more - when temp is cooler it goes inward. The effect is tiny but visible when you reflect light off the side.
Much more likely is that the Z axis is not always moving the same distance. If for example you ask the Z axis to move .06 every time and sometimes it moves .03 then the next layer is effectively overextruding and it sticks out like a larger pancake on a stack of pancakes. The Z axis id quite accurate so it may move less than it is supposed to for one or even 3 layers but eventually things catch up and it will move *more* than it's supposed to and you get some underextrusion on that layer. It makes for some horizontal bandings.
This kind of thing shows up more the thinner the layer height. That is one of a few reasons I recommended switching back to .1mm layers as that is easier to print (for other reasons as well).
So anyway it might not be dirt. It usually is dirt. It could be a bad Z bearing. It could be bad Z nut. One test is to see if it always happens on the exact same layers every time. Or if the pattern repeats every 3mm vertically then the Z screw may be bent in a "C" shape.
Try putting something heavy - about 5kg - like a brick - on the back of your print bed. That should help remove Z errors and make each layer more consistent.
Also watch the nozzle temperature carefully while printing to see if it has a slow oscillation of at least 5C in range from hottest to coolest and if the hot layers are sticking out more.
@gr5 If you look closely then you will notice that not all the over extruded sections are throughout the whole layer. just parts of the layer. This leads me to believe it's not the Z axis. It could be Zhop though as it can cause this. There are also lots of little blobs which could also be from Z hop or retractions or something.
Sticking to 0.1 layers till you get it right is a good idea but you can definitely go down to 0.06 layer height. Just don't go below 0.04.
Never print ABS through a previous "PLA-Nozzle", as the smallest residue of PLA which mixes poorly with ABS, will keep staying inside the nozzle and slowly turn to carbon or жевательная резинка.
Two ways to avoid problems when changing from ABS / PLA and back:
1) Very clean cold pulls
2) change nozzle.
@gr5 If you look closely then you will notice that not all the over extruded sections are throughout the whole layer. just parts of the layer. This leads me to believe it's not the Z axis. It could be Zhop though as it can cause this. There are also lots of little blobs which could also be from Z hop or retractions or something.
Sticking to 0.1 layers till you get it right is a good idea but you can definitely go down to 0.06 layer height. Just don't go below 0.04.
This is on 0.1 layer, so you think the Z hop and retractions is the issue there?
Assume this is not the Z axis problem, or over extrusion, I hv read a few on internet but i couldnt quite understand what the implication of it on a small nozzle yet. What I comprehend is that these are to help with the quality of print by avoiding overlaps and dripping of filament that also causes stringing/ringing, which I dont exactly know how this play to my print. I hv set retraction to 4.5mm, and Z hop 2mm, and I am not sure, to improve this print, should these value go lower or higher?
If this really is over-extrusion issue like @gr5 mentioned, i guess lower temperature might help?
Not so sure which would be the more likely problem, it seems on 0.4mm nozzle things looks smoother at 0.1mm perhaps it's more stable and forgiving. Will try more prints.
Turn z hop off. It makes your print worse.
Also what is your minimum layer time?
Z axis issues give overextrusion and underextrusion and adjusting temp won't help. Only fixing your z axis would help. Looking at that picture it could be the case so you should clean It.
Recommended Posts
fbrc8-erin 302
Link to post
Share on other sites