Trying to increasing print surface quality (E3D on Flashforge?)

[Dargonfly]

I have printed a lot in layer heights of 0.1[mm] and low speeds (25~50) and noticed that my prints had a 'wavy' pattern in the vertical direction.

This could have several causes:
- 'loose' X & Y belts; (tried tightening, still appeared on straight lines)
- mechanical jerking (tried changing acceleration settings and lowering print speeds overal - resulting in less waves at corners, but still present on straight lines)
- uneven filament (measured my filament at several places with my Mitutoyo, seemed to be good)
- uneven heating of filament (temps on Flashforge appear to be very constant, no real way of checking the output - could conduct a test with extruding at a MUCH higher temp than I usually use)
- damage / junk in nozzle (does not appear to be the case)
- nozzle size too large for amount of extruded filament (this could only be tested by using a smaller nozzle)
- general vibrations
- printing too fast for extruder steppers (not the case)
- printing too slow for extruder steppers (uneven extrusion due to low speed)
- feeding gear is damaged / too little teeth

This last one got me intrigued so I tried printing as usual but moving the bed way downwards (printing in the air) and noticed that the extrusion was a bit 'wavy' by itself: (thick - thin -thick - thin - thick - ...)

This is probably the issue I'm looking at, so that means it has to be one of these causes:
- uneven filament
- uneven heating of filament
- general vibrations
- printing too slow for extruder steppers
- feeding gear is damaged / too little teeth

As mentioned, I've already conducted some tests for the above, but will redo them a bit more serious now that I've narrowed down the cause (which hopefully is really is the cause!!)
Another thing to note here is that the wavy pattern on the final product (a cube for testing purposes) is very consistent with its location. This already rules out the 'uneven filament' because that would mean that it should be more 'random'. But since the extruder starts doing the same thing EVERY layer, it has to be one of the causes listed in bold above.
General vibrations I have already minimized as much as possible.. but this is more of the final cause if everything else fails (which means that that would be the limit of the surface quality.)

Uneven heating I will test by heating the filament MORE, why? Uneven heating could be a problem if different parts of the filament are in different 'phase' stages, if I heat everything way more then it should all be in the molten phase I'm looking for.

The 'printing too slow' is a bit more difficult to check, because yes; this doesn't happen at faster feed rates, but I want to be able to extrude with these layer thicknesses.
So I was thinking to get a bowden setup here, then running the motor 10 times as fast (in software), and using a gear (10:1) to allow the feed rate to be correct. This way any 'steps' of the stepper motor shouldn't be visible anymore.
Regarding that: has anyone thought about replacing the extruders with some E3D's? http://e3d-online.com/E3D-v6/Full-Kit
Does anybody know which voltage is used on the current extruder heating elements? 24V I guess? Would also allow me to print at 0.25[mm], or even better: use 0.25[mm] for shells and 0.8[mm] for infill with dual extrusion.

Regarding the feeding gear: this one works well. it 'might' create some 'step-feeding' of filament due to limited amount of teeth. But I do not want clogging of the extruder either due to a bad feed-gear.

Lots of text I know, but these were just some ideas, not perfectly organized either. I hope that somebody has some thoughts regarding this! If there is interest I will keep you guys updated with pictures of the tests I'm conducting (and the wavy pattern I'm talking about.)


EDIT: there's also general settings that influence this behavior like amount of shells, infill, layer height, etc. but for maximum quality and some durability of the model I often find myself printing at 0.1[mm] layer height, 0.4[mm] layer width with 2 shells and 15% infill - if the height-width rate isn't good, then switching to E3D would allow me to use a 0.25[mm] nozzle. Infill has almost no effect on surface quality with 2 shells. 2 shells looks different then 1 shell.

EDIT2: here is a topic that has pinned down the most (most topic disregard it as z wobble, which is not the case here) http://forums.reprap.org/read.php?1,214990
this picture shown (exaggerated) what I'm talking about:
(NOT MY PICTURE - MY RESULTS ARE A LOT LESS VISIBLE, BUT SIMILAR TOO THIS)

[Mjolinor]

My guess here would be that you are using micro-stepping for the extrusion stepper and the stepper motor is a cheap one.

Micro-stepping is fine in theory but in practise it is never accurate due to build and component quality inside the stepper.

The fix is to use a geared stepper and/or to use a top quality stepper motor though the problem is still there no matter how much you pay for a stepper motor, it is just not as bad.

[Mjolinor]

Bit of arithmetic:

If you have a 13 mm toothed gear then one step is 0.2 mm of filament.
At 1.75 mm diameter that is 0.4912 mm^3 of filament.
With a 0.3 mm nozzle that equates to an extruded length of 1.74 mm per step.

If you are using half micro-stepping then each half step should be 0.869 mm of filament but due to inaccuracies in the half step one will be more than the other.

If your "oscillations" are 1.74 or so period then that is probably the problem.

Attached a picture of something on my printer that seems to show the same thing. The oscillations are just under 2 mm apart so that seems to fit.

[curious aardvark]

I don't get those lines.

But I tend to print as fast as I can. Doesn't seem to effect quality.

That said - I couldn't care less about that sort of thing anyway :-)

[Mjolinor]

I think that faster printing will lower the effect due to elasticity in the molten filament within the melt chamber. Greater layer height should also make it less visible I think.

Someone should check my arithmetic.

[Dargonfly]

I've added a file to show you exactly what I get on my prints (this is 0.15[mm] layer height) though, much worse on 0.1[mm]!)

My guess here would be that you are using micro-stepping for the extrusion stepper and the stepper motor is a cheap one.

Micro-stepping is fine in theory but in practise it is never accurate due to build and component quality inside the stepper.

The fix is to use a geared stepper and/or to use a top quality stepper motor though the problem is still there no matter how much you pay for a stepper motor, it is just not as bad.

Ok! that sounds like a very plausible cause too!
But.. I don't know if I'm using micro-stepping, and I do not know whether my stepper is cheap or not (willing to buy new ones though!)

Do you perhaps know what I need to look for in a stepper motor (for Flashforge Creator)?
And do you think this is for feeding or for X & Y too?

Bit of arithmetic:

If you have a 13 mm toothed gear then one step is 0.2 mm of filament.
At 1.75 mm diameter that is 0.4912 mm^3 of filament.
With a 0.3 mm nozzle that equates to an extruded length of 1.74 mm per step.

If you are using half micro-stepping then each half step should be 0.869 mm of filament but due to inaccuracies in the half step one will be more than the other.

If your "oscillations" are 1.74 or so period then that is probably the problem.

Attached a picture of something on my printer that seems to show the same thing. The oscillations are just under 2 mm apart so that seems to fit.

I got lost in your math... completely.

But I will measure my oscillations (which look quite similar to yours!) as soon as I get home. The only difference being that I use a 0.4[mm] nozzle.

But in conclusion: I need to scratch all my thoughts and focus on the stepper motors that feed the material? Better motors = more consistent feeding, or can I change the firmware somehow regarding this?


Thanks for the help btw, really appreciate it!

And yes, larger layer height means more extrusion so the variations between steps are much less noticeable. (at least, that is my observation from all this)

[Mjolinor]

Of the two possibilities the geared stepper will provide better results though I do not know if you can change the extrusion parameters on the Mightyboard or not. A more expensive stepper will only reduce it, you cannot cure it that way and you will be gobsmacked at how much a UK or German made stepper costs. You also run the risk of fraud however you buy it. English stickers on Chinese goods are too damn common of late.

http://www.aliexpress.com/item/EU-Fr...054390403.html

These are what you want and it is a good price but you are going to have to do some engineering to fit them because they are longer. It may not even be possible to fit them on the head and you will increase the weight and therefore the ability of the head to respond to requests for speed change (more inertia). Another potential problem with a geared extruder is the possibility of backlash in the gears. A planetary set minimises this but you may have to recalculate your extrusion retraction.

I think they have to be using micro-stepping because a minimum extrusion length of 1.74 mm is way too much. They are probably using 1/16 microsteps, you could find that in the specs somewhere I think if you need to know.

Gearing down your stepper means that speeds will be reduced so if you applied the same logic to your X Y steppers it would certainly improve things but I think the maximum speed the stepper drivers can pulse at may mean that you wait for ever to make things. I don't think the improvement in quality would be worth the expense and hastle. You would be better to put your money into getting rid of all the plastic bits inside the printer and using chains instead of elastic bands to drive the axes. It's all down to cost in the end. If you want to spend as much as a Stratasys cost then you would be better buying a Stratasys.

[Dargonfly]

Of the two possibilities the geared stepper will provide better results though I do not know if you can change the extrusion parameters on the Mightyboard or not. A more expensive stepper will only reduce it, you cannot cure it that way and you will be gobsmacked at how much a UK or German made stepper costs. You also run the risk of fraud however you buy it. English stickers on Chinese goods are too damn common of late.

http://www.aliexpress.com/item/EU-Fr...054390403.html

These are what you want and it is a good price but you are going to have to do some engineering to fit them because they are longer. It may not even be possible to fit them on the head and you will increase the weight and therefore the ability of the head to respond to requests for speed change (more inertia). Another potential problem with a geared extruder is the possibility of backlash in the gears. A planetary set minimises this but you may have to recalculate your extrusion retraction.

I think they have to be using micro-stepping because a minimum extrusion length of 1.74 mm is way too much. They are probably using 1/16 microsteps, you could find that in the specs somewhere I think if you need to know.

Gearing down your stepper means that speeds will be reduced so if you applied the same logic to your X Y steppers it would certainly improve things but I think the maximum speed the stepper drivers can pulse at may mean that you wait for ever to make things. I don't think the improvement in quality would be worth the expense and hastle. You would be better to put your money into getting rid of all the plastic bits inside the printer and using chains instead of elastic bands to drive the axes. It's all down to cost in the end. If you want to spend as much as a Stratasys cost then you would be better buying a Stratasys.

Thanks for the help!
I think that geared stepper would allow for more fluid feedrate. I will take pictures of what I mean with the current feedrate. It's possible to increase the feed in software, so that shouldn't be an issue.
Will get back to this with pictures tonight! (and hopefully an idea on how to apply these motors in the current build.)

I don't think mass is a problem, because I already print really slow; and if it becomes a problem I'll just lower the max. acceleration settings.
Also, I don't think I need to change everything to Stratasys quality. It's just that this is currently my largest bottleneck quality wise. (printing at 0.05 looks perfect in X & Y direction, but creates these ugly vertical osculation.)

[Mjolinor]

If you go to a 10:1 planetary geared stepper then the feed rate steps will need to be 10 times as much and the oscillations will be 1/10 of the size, probably not visible at all I would think.

If you wanted to get inventive then you could probably try to even out the steps on your existing steppers by adding small resistors in the wires to finely adjust the current during the microsteps but that would be a lot of trial and error and take lots of time I think.

[Mjolinor]

I have been thinking about this and if the cause is as I suggest then printing a flat panel will show the same pattern on the horizontal surface.

I suspect however that you do not have this pattern on a horizontal surface.

Just more thought to complicate things.