arhi
Posts: 483
Joined: Thu Oct 06, 2016 5:13 pm

Re: Horizontal lines in test tower

btw the other issue that can make this ribbing you have prominent is backlash, so if you have one XY position when head is going CW and other when it is going CCW then it can be backlash, especially if you have springs on your XY axes (for e.g. those spring tighteners for belts that will cause this issue) .. you can theoretically "solve" the problem by making slicer always print outline in same direction (no clue how to make s3d do that, needed it few times myself) but that's just a hack, not a real solution, drop the springs, make stiff tighteners and drop the acceleration/jerk till you get good results

note also that lowering acceleration and jerk will (in a good way) affect precision of the small holes
gcodestat integrates with Simplify3D and allow you to
Calculate print time accurately (acceleration, max speed, junction deviation all taken into consideration)
Embed M117 codes into G-Code
Upload your G-Code directly to Octoprint
open source and unlicence
InspGadgt
Posts: 61
Joined: Wed Oct 18, 2017 7:59 pm

Re: Horizontal lines in test tower

So today I recompiled and reflashed Marlin on my printer. There was some changes I wanted to make. So while I was in there I lowered the acceleration and jerk settings. I don't know if that actually did anything or if just reflashing the firmware made a difference...but my print quality got better. I still haven't got it as good as my Cura print but it is getting there. See attached pic...Cura on the left and Simplify3D on the right. Both done at .20 layer height and at 60mm/s speed. Comparing the settings I'm not finding anything that stands out different between Cura and Simplify3D but some of the settings I am having to guess what they are called in the other software.
Attachments
20171114_234413.jpg
arhi
Posts: 483
Joined: Thu Oct 06, 2016 5:13 pm

Re: Horizontal lines in test tower

InspGadgt wrote:Comparing the settings I'm not finding anything that stands out different between Cura and Simplify3D but some of the settings I am having to guess what they are called in the other software.
compare gcode not settings
gcodestat integrates with Simplify3D and allow you to
Calculate print time accurately (acceleration, max speed, junction deviation all taken into consideration)
Embed M117 codes into G-Code
Upload your G-Code directly to Octoprint
open source and unlicence
InspGadgt
Posts: 61
Joined: Wed Oct 18, 2017 7:59 pm

Re: Horizontal lines in test tower

I'm starting to think that when you leave Cura at 0 and 0 for acceleration, or undefined, that it actually defaults to a lower setting than what the printer is set as. That could be a big reason why the print quality jumped hugely from 15.X to 2.X.

As for the code I had some questions on that. When I look at it in Arduino 1.8.5 I see 2 acceleration settings:
"#define DEFAULT_MAX_ACCELERATION"
and
"define DEFAULT_ACCELERATION"

When I look at the settings on the printer itself after I flashed to the latest Marlin I see something completely different. There is no Max Acceleration in the menu but there is Vmax and Amax...velocity and acceleration The Vmax is set to 400 which is what I have both DEFAULT_MAX_ACCELERATION and DEFAULT_ACCELERATION set to...but Amax is set to 2000 and I can find nothing in the code I uploaded that has a value of 2000. The default setting for the Anet A8 configuration file was DEFAULT_MAX_ACCELERATION was set to 2000.

So right now I'm looking up to see if there is a way to look into the g-code generated by Cura to see if it changes the acceleration and jerk settings.
InspGadgt
Posts: 61
Joined: Wed Oct 18, 2017 7:59 pm

Re: Horizontal lines in test tower

arhi wrote:btw the other issue that can make this ribbing you have prominent is backlash, so if you have one XY position when head is going CW and other when it is going CCW then it can be backlash, especially if you have springs on your XY axes (for e.g. those spring tighteners for belts that will cause this issue) .. you can theoretically "solve" the problem by making slicer always print outline in same direction (no clue how to make s3d do that, needed it few times myself) but that's just a hack, not a real solution, drop the springs, make stiff tighteners and drop the acceleration/jerk till you get good results

note also that lowering acceleration and jerk will (in a good way) affect precision of the small holes
I'm not using a spring tensioner and the belt is quite tight. Besides if that were the case it would affect all the slicers not just Simplify3D
InspGadgt
Posts: 61
Joined: Wed Oct 18, 2017 7:59 pm

Re: Horizontal lines in test tower

I figured out how to look at the G-Code and verified that on the default acceleration and jerk settings in Cura...this does NOT a M201, M203, or M204 command to use a different value from what is on the printer...the issue must be somewhere else.
InspGadgt
Posts: 61
Joined: Wed Oct 18, 2017 7:59 pm

Re: Horizontal lines in test tower

arhi wrote:no accel/jerk control in s3d

if you got this after firmware upgrade (not changing anything else) then it's the firmware settings... not much in the firmware really
- steps/per/mm for XYZE
- max speed for XYZE
- acceleration for XYZE
- jerk for XYZE

so do some experiments with those, drop them big time (for e.g. I find that for most machines accel of 300 and jerk of 4 are great start positions for calibration, then you start raising accel till you start to get lower quality, dial back to good quality and then start getting jerk up .. ) see where you end up.. you can easily experiment ... manually edit g-code, make something like temp tower .. make a tower and then every 20 layer add manually

Code: Select all

M201 X400 Y400
M205 X4 Y4 
... then after 20 layers
M201 X500 Y500
M205 X4 Y4 
... then after another 20
M201 X600 Y600
M205 X4 Y4 
... another 20
M201 X700 Y700
M205 X4 Y4 
....
so do a manual change of the accel/jerk settings troughout print and look for differences in quality
So I tried acceleration at 400, 200, and 100, jerk at 4...no real change in the print.

Now that I think about it...I only really noticed it after the upgrade. Before the upgrade I was printing my 3dLabPrint Spitfire and was having lines as well but I was thinking that was from something else...perhaps printing at too hot of a temp. At the time I was just so happy at being able to print more than one piece at a time I pretty much ignored the lines thinking I would figure it out later. After the upgrade I had some other issues making my prints worse than usual but it looks now like my prints are at least back to where they were before the upgrade. So now I'd like to figure out how to get rid of these lines, get a good print, and still be able to print 4 pieces at a time...(Could never do that with Cura...kept getting holes in my planes).

So...lets go back to scratch on this and ignore the upgrade. How can I get my Simplify3D prints to look as good as my Cura prints?
arhi
Posts: 483
Joined: Thu Oct 06, 2016 5:13 pm

Re: Horizontal lines in test tower

backlash can happen even without springs .. compare g-codes, look for direction of the moves, does cura generated g-code goes always in one direction and s3d g-code alterates?

or maybe you can upload here a g-code that has visible lines and g-code that does not have visible lines and we can take a look :D without guessing what the differences in g-code are :D
gcodestat integrates with Simplify3D and allow you to
Calculate print time accurately (acceleration, max speed, junction deviation all taken into consideration)
Embed M117 codes into G-Code
Upload your G-Code directly to Octoprint
open source and unlicence
horst.w
Posts: 861
Joined: Fri Oct 17, 2014 5:00 pm

Re: Horizontal lines in test tower

horst.w wrote:
...
Can you please attache your factory-file? and also the STL-file

Regards
horst.w
GER
InspGadgt
Posts: 61
Joined: Wed Oct 18, 2017 7:59 pm

Re: Horizontal lines in test tower

Ok here they are
Attachments
test_rectangle.stl
(684 Bytes) Downloaded 162 times
test_rectangle.factory
(3.24 KiB) Downloaded 172 times

Return to “General Discussion and Tips”