Help Test a Potential Cura Feature!

[neotko 1,417]

Not until next-next week. I'm getting ready for vacations. But yeah I can do some test.

[lars86 42]

I'm in the midst of printing a rocket ship (with smoke) that takes all day. Will test you .6mm overlap first thing in (my) morning

 

Good luck with the print!

I just printed the 0.6mm overlap and saw even better results. It is now hard to tell the actual seam from the slight ripple from the print head settling in after the travel move.

I guess I need to make a 0.8mm version now.

[lars86 42]

Okay, I went ahead and cleaned up the original post for clarity. I also took out the code which had no coasting. Since it appears to have no appreciable difference on the seam, it was a redundant print.

[DaHai8 74]

Latest Test Results:

The Coast .6 is showing the least zipper, but I have some concerns on the head jerking.

I think what is happening is that at the end of the layer, after the coast, you're moving the head back .2 to .6mm from whence it came.  There is an audible 'snap' on my printer when this jerk occurs. It's kinda disconcerting. But more than that, I think it's not helping the finish.

What if you instead continued to move the head .2 - .6mm forward (just following the beginning of the path) instead of going backward. I think this might result in even more seam inprovement as there wouldn't be an abrupt stop and retrace.

Of course, this opens the issue of what if then end of the layer is not where it started and there is nowhere to go forward - which it probably why you do a retrace backward...

Hmm.....

[DaHai8 74]

Survey Says:

 

Would you say that your seams generally bulge outwards, pucker inwards, or both?

Both, it looks like a zipper line. It even uglier when they don't line up and are spread all over an otherwise smooth face.

 

What machine are you using?

JG Aurora A3 Prusa I3 Printer. PLA 1.75. Heater Bed, etc. etc. etc. I love this printer

Fixed income. Can't afford a Ultimaker. Wish I could...

 

What slicer and version are you using?

Cura 2.1.2 - Best one on the market and getting better every release (much more so than CrapWare, er Craftware, whatever.). Do I get extra points for saying that? It is true.

Edited July 8, 2016 by Guest

[nallath 1,120]

Cura 2.1.2 - Best one on the market and getting better every release (much more so than CrapWare, er Craftware, whatever.). Do I get extra points for saying that? It is true.

 

Of course:

[lars86 42]

Hahahaha!

I know I'm a picky pain in the butt on the Cura Github, but it is really out of love! I know you guys have cumulatively put in an absurd amount of time and energy on this piece of software. We seriously appreciate it, and plain and simple, it makes great prints.

[lars86 42]

Latest Test Results:

The Coast .6 is showing the least zipper, but I have some concerns on the head jerking.

I think what is happening is that at the end of the layer, after the coast, you're moving the head back .2 to .6mm from whence it came.  There is an audible 'snap' on my printer when this jerk occurs. It's kinda disconcerting. But more than that, I think it's not helping the finish.

What if you instead continued to move the head .2 - .6mm forward (just following the beginning of the path) instead of going backward. I think this might result in even more seam inprovement as there wouldn't be an abrupt stop and retrace.

Of course, this opens the issue of what if then end of the layer is not where it started and there is nowhere to go forward - which it probably why you do a retrace backward...

Hmm.....

 

You are correct. In order to have an aligned seam (especially important in this test), you need to do a quick reversal after the overlap segment. Once we can make an aligned seam almost disappear, I really think that you can switch to a random seam placement, and lose it all together.

I run fairly aggressive travel speeds, and think in these test files I have that set to 190 mm/s. What I would suggest, is for you to turn your XY Jerk, and/or XY Max Accel settings down. I don't know where the defaults are set on your printer, but I feel that the UM's come set too aggressively.

My machine is quite rigid, and low friction, so I can get away with more aggressive settings in general. I currently run at XY Jerk = 14, and XY Accel = 2700. If you want to be a little more conservative, setting them to 12 and 2000 will net nice motion too.

You may also want to put a hard limit on XY travel speed. You could set that around 160 to calm down the fast travel moves.

[DaHai8 74]

Darn! That coupon is irredeemable here....

[lars86 42]

556 views... 1 participant. Killing it!

[DaHai8 74]

lars86, got a Coast + .8mm you'd like me to try?

[lars86 42]

lars86, got a Coast + .8mm you'd like me to try?

Here it is! I finally was able to write a regex search and replace to change the 116 overlap positions. Cura had shifted the start/end points, very slightly throughout the print, which made it hard. You will probably need to add the XY shift code in for your smaller bed, and push through your script. If you want to send back your processed versions of the 0.6 and 0.8 overlap code, I'll add them to the original post.

https://www.dropbox.com/s/76ngzpusmzcwjlt/Overlap_0.8.gcode?dl=0

[Titus 170]

Just read the topic, I can imagine your irritation with the amount of response. Though it is holiday season :)Some other probable causes:

Can you specify for which printer the gcode is? UMO/UMO+/UM2 are quite different from each other in terms of gcode input etc.

Also, I don't really have trouble with seams, so the need is less for me, and possibly others to test

Edited July 13, 2016 by Guest

[lars86 42]

Just read the topic, I can imagine your irritation with the amount of response. Though it is holiday season :)Some other probable causes:

Can you specify for which printer the gcode is? UMO/UMO+/UM2 are quite different from each other in terms of gcode input etc.

Also, I don't really have trouble with seams, so the need is less for me, and possibly others to test

 

Good point Titus, I forgot that the UM2 uses different gcode flavor. In looking at that more, the differences seem to be: no real header code, retracts with G10, and extrusion amounts in mm^3, not mm. I'm not sure there is a clean way to quickly port all my modded code for UM2. I think the best bet would be to regen the code for UM2, then alter to add overlap. I really wish the seam location was exact, but it wobbles a little, and that makes it harder to automate the changes.

I would be far more inclined to jump on this and do it, if the interest level wasn't so dismal. It's really sad, because before the new forum software rolled out (wayyy before it was ready), the community was pretty awesome. Now it seems all the doers were pushed away by a clunky interface, and were replaced with browsers.

When you say that you don't have "trouble" with seams, what do you mean exactly? Seams are never really trouble, in my eyes, but if they are visible, they are ugly. That probably bothers some people more than others. My goal is to be able to challenge someone to find the seam on my prints, which is a fairly high bar. Would you really say that an aligned seam is hard to find on your prints?

[Titus 170]

No more that I either don't have seams, or don't notice them. The last time I remember having a seam was an extrusion test. Not really a print I'd mind to have it on.

I'm not sure what you changed in the Gcode, but perhaps @nallath can help to get you to port it for um2.

[lars86 42]

Mind sharing a macro photo of one of your seamless prints? I'd say that the more likely case, is that you have random seam placement in your slicer settings. This just distributes them all over. They will look like tiny holes in the surface if your seam is dimpled, or pimples on the surface in the other case.

I can see what needs to change to move from UM to UM2 code, but it's a bunch of work. If only one person with a UM2 wants to try, and they already feel like their seams are perfect, I'm not sure I'd invest the time.

[DaHai8 74]

Sorry for the delay. Here are the latest results:

To me, it appears the .8 overlap is not any different than the .6. So I think that's about as good as it gets - at least on my printer.

The 1.75mm filament converted gcode files can be found here: http://bitman.org/CoastTest_1.75mm.zip

Cheers!

[tourdetour 0]

Hi, I've tried your Control gcode on my Ultimaker 2 upgraded to an Ultimaker 2+... it seems to have done nothing but scratching my buildtak... A bit of fillament is coming out but the nozzle is too close from the bed (other models sliced with cura work fine). Does the gcode overwrite the Z level set in the printer before printing ? Is there a way I could generate these gcodes ? That would be easier I think.

[lars86 42]

Hi, I've tried your Control gcode on my Ultimaker 2 upgraded to an Ultimaker 2+... it seems to have done nothing but scratching my buildtak... A bit of fillament is coming out but the nozzle is too close from the bed (other models sliced with cura work fine). Does the gcode overwrite the Z level set in the printer before printing ? Is there a way I could generate these gcodes ?  That would be easier I think.

 

Sorry about that! I just added the note to the original post, that these files are Reprap style gcode, for Ultimaker Original (and + I think), but don't work on the Ultimaker 2.

I can work on another set for the UM2, but you are the first who has expressed interest.

[tourdetour 0]

Sorry about that! I just added the note to the original post, that these files are Reprap style gcode, for Ultimaker Original (and + I think), but don't work on the Ultimaker 2.

I can work on another set for the UM2, but you are the first who has expressed interest.

 

Hummm wait a minute, Reprap gcodes should work fine on an Ultimaker 2, I've already printed some Reprap models using octoprint. Maybe the only problem is that Reprap overrides machine settings so the bed isn't leveled properly.

Your models don't seem too long to print, and as you need help, well, I can help you. You idea seems really interesting. I print a lot of twisty puzzles, and removing those seams would help for the turning quality and for the look.

Finally, I think converting gcodes from one flavor to another should be a cura feature. That would be really handy.

And good luck for your project !

By the way, I used to use cura 15.04.6 until a few days ago because I was very disappointed with the first version of cura 2 I downloaded on Linux (It didn't print well, was slow and the buttons were too big). But now with cura 2.1.2 (on mac) it's absolutely awesome, the UI is really good and understandable and clean, it's much faster and it prints well.

[ultiarjan 1,223]

Using reprap gcode on UM2 will not impact your bed leveling......

[tourdetour 0]

Using reprap gcode on UM2 will not impact your bed leveling......

It seems to override it just for the current print, you can also specify the temperature in the Reprap file and stuff like that.

[lars86 42]

The issue is, which you already found out, the UM2 expects the extrusion values in mm^3, and the older machines use just mm. The relative distances are way off, so you got massive under-extrusion.

[tourdetour 0]

Can't we get the ratio and just adjust the feeding rate on the printer ?

As the filament is a 2.85mm of diameter cylinder,

V = pi*1.425^2*h = 6.3794h

So if we use a feeding rate of 638 %, should it work or am I missing something ?

[lars86 42]

You are correct, it should just be a straight multiplier. I calculated at one point, and it was 6.__x, so your calc is probably right. If you download the files that use the flow multiplier to correct for 1.75mm filament, you could just edit the flow multiplier M code in the header for the 638 value, and let us know if it works.