• Welcome to the new Distech Automation Forum!
  • Feel free to post questions, comments, and feedback.
  • Ask us anything!
Hello There, Guest! Login Register


Thread Rating:
  • 1 Vote(s) - 5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Post processor script development (S3D/Prusa Slic3r support)
#61
Trying out new version now, I see the initial layer print of the ooze tower has been slowed since the version I had, good idea!
I do note however the actual tower is really really far from the object now.. Why? Personally I think a 5layer skirt touching its base would be a good idea to keep it stable.

thanks again !

PS: What is the actual difference between the Peek and PTFE profiles btw? Im sure this is written somewhere!
 
Reply
#62
The tower should be 5mm away from the object's borders. And border means the furthest coordinates from object center so if the object is wider on top, the base and tower will seem to be far away from each other. Also skirt counts as object. Not going to change that anytime soon, would have to implement logic that knows different parts of the object. Maybe if there's existing library for that, haven't checked...

Skirt could be wider, maybe even follow the slicer skirt settings.

PTFE and PEEK differences in filaswitch aren't documented Smile. Basically the g-code that retracts the filament in before tool change is bit different, PEEK has some extra cooldown if I remember correctly. I've used both with PTFE liner and they work. PEEK might even be more consistent? Just a gut feeling, I could be wrong...
 
Reply
#63
(10-03-2017, 04:13 PM)spegelius Wrote:
(10-03-2017, 03:52 PM)EskimoRuler Wrote:
(10-03-2017, 07:16 AM)spegelius Wrote: My guess is that Smoothie handles G90 differently, it sets absolute positioning for extruder too and doesn't respect previous M83. Should be easy to fix, I'll just add M83 at the end of tower gcode block as you did manually.

I tried finding some information on if smoothies' G90 sets absolute for the extruders too, but I couldn't turn up any direct info on it. So given whats happening I'd say you are correct.

I've only been using Smoothie for about 2 years, so I don't have experience with other firmwares. And I've never played around with absolute/relative positioning. Is using M83 at the beginning of the gcode supposed to hold true through the reset of the job, even after switching between the too?

It works like that on Marlin and apparently Repetier, although I've only used Marlin myself. M83 is set only once, at the beginning of g-code. Adding it after every G90 shouldn't have any adverse effects in this case.

Hey spegelius,

Just wanted to let you know that I posted this question on the Smoothieware Google Plus page and wolfmanjm confirmed that G90/G91 will set absolute and relative for everything on smoothie. And that I'd have to use the M83 after each G90.

But the script is working great! I feel my benchy came out really well and the maker coin perfectly! Thanks again for this.

https://www.dropbox.com/s/c13bw19p5eydfy...1.jpg?dl=0
https://www.dropbox.com/s/d7sezek70y751g...6.jpg?dl=0
https://www.dropbox.com/s/0yzsm02sjs1vgm...5.jpg?dl=0
https://www.dropbox.com/s/u72bilarq4swwu...3.jpg?dl=0
 
Reply
#64
(10-04-2017, 06:41 AM)spegelius Wrote: The tower should be 5mm away from the object's borders. And border means the furthest coordinates from object center so if the object is wider on top, the base and tower will seem to be far away from each other. Also skirt counts as object. Not going to change that anytime soon, would have to implement logic that knows different parts of the object. Maybe if there's existing library for that, haven't checked...

I think this is over 5mm distance!?
https://imgur.com/a/9Ixlh
 
Reply
#65
[quote pid='336' dateline='1507133140']
Hey spegelius,

Just wanted to let you know that I posted this question on the Smoothieware Google Plus page and wolfmanjm confirmed that G90/G91 will set absolute and relative for everything on smoothie. And that I'd have to use the M83 after each G90.

But the script is working great! I feel my benchy came out really well and the maker coin perfectly! Thanks again for this.

https://www.dropbox.com/s/c13bw19p5eydfy...1.jpg?dl=0
https://www.dropbox.com/s/d7sezek70y751g...6.jpg?dl=0
https://www.dropbox.com/s/0yzsm02sjs1vgm...5.jpg?dl=0
https://www.dropbox.com/s/u72bilarq4swwu...3.jpg?dl=0
[/quote]

Okey great that clears it up. I'll add M83 lines after each G90 so that's taken care of then. And great to see real life results of the script Smile. Looking good

(10-04-2017, 06:43 PM)kodachrome Wrote:
(10-04-2017, 06:41 AM)spegelius Wrote: The tower should be 5mm away from the object's borders. And border means the furthest coordinates from object center so if the object is wider on top, the base and tower will seem to be far away from each other. Also skirt counts as object. Not going to change that anytime soon, would have to implement logic that knows different parts of the object. Maybe if there's existing library for that, haven't checked...

I think this is over 5mm distance!?
https://imgur.com/a/9Ixlh

Hmm, a delta then. I don't have one so understandably my testing with one has been quite limited. Send me the g-code file and I'll see what triggers such large distance. Currently the code simply goes through all the print moves and calculates the min and max of x and y and uses those values. Might be that delta side of the code has broken due to some other change... If only I'd have enough motivation to write some unit tests Tongue
 
Reply
#66
Unfortunately dont have that gcode, was battling other issues so have been changing settings and trying.. and of course the tower is right on 5mm from the skirt now! No idea what was different about that one, but if I find it again Ill send you the gcode!
 
Reply
#67
Pushed latest code to master branch and I'll make 0.12 release out of later today. Got stuck with one PLA brand that doesn't feed properly during material change and been debuggin it, so I don't have full print done to test the changes. It's blue PLA from marwiol.pl, cheap and seems to be somewhat softer? or something, because the feed gear grinds through it quite easily and it seems to leave bit too long string when pulled out from hotend.
Still, the g-code itself works fine and other PLA's work fine, so there shouldn't be any problems.
Major changes to be released:
- Prusa Slic3r support
- fix Smoothieware E relative mode
 
Reply
#68
(10-06-2017, 10:14 AM)spegelius Wrote: Pushed latest code to master branch and I'll make 0.12 release out of later today. Got stuck with one PLA brand that doesn't feed properly during material change and been debuggin it, so I don't have full print done to test the changes. It's blue PLA from marwiol.pl, cheap and seems to be somewhat softer? or something, because the feed gear grinds through it quite easily and it seems to leave bit too long string when pulled out from hotend.
Still, the g-code itself works fine and other PLA's work fine, so there shouldn't be any problems.
Major changes to be released:
- Prusa Slic3r support
- fix Smoothieware E relative mode

Hey spegelius,

Thanks for the update. I just tested your latest push and the M83 after each tower is there and perfect.

I did notice though that there needs to be a M83 after the Raft is created at the beginning of the print. But other than that, it looks good.
Code:
G90
M83
M106 S0
M140 S100
M104 S235 T0
M109 S235 T0
; START SCRIPT START
G31 ; bed leveling
G28 ; home all axes
G30 F2000 Z1.21
T0
; START SCRIPT END
; TOWER RAFT START
G1 Z0.400 F1200; z-hop
G1 X-27.260 Y63.470 F18000; move to raft zone
G1 Z0.2 F1200; move z close
G91; relative positioning
G1 X54.800 E2.1043 F2000; raft wall
G1 Y17.700 E0.6797 F2000; raft wall
G1 X-54.800 E2.1043 F2000; raft wall
G1 Y-17.300 E0.6643 F2000; raft wall
G1 X54.400 E2.0890 F2000; raft wall
G1 Y16.900 E0.6490 F2000; raft wall
G1 X-54.000 E2.0736 F2000; raft wall
G1 Y-16.500 E0.6336 F2000; raft wall
G1 X0.424 Y-0.424 F18000
G1 Y16.900 E0.8099 F1000; raft1
G1 X1.000 F1000; raft2
G1 Y-16.900 E0.8099 F1000; raft3
G1 X1.000 F1000; raft4
G1 Y16.900 E0.8099 F1000; raft1
{...}
G1 X1.000 F1000; raft4
G1 Y16.900 E0.8099 F1000; raft1
G1 X1.000 F1000; raft2
G1 Y-16.900 E0.8099 F1000; raft3
G1 X1.000 F1000; raft4
G90; absolute positioning  ------------------------------"Right after this G90"
; TOWER RAFT END
G1 E-2.0000 F3600
G1 Z0.400 F1200
; process Color1
; layer 1, Z = 0.200
; TOOL CHANGE
; tool H0.200 W0.400
; skirt
G1 X3.538 Y-5.118 F18000
G1 Z0.200 F1200
G1 E2.0000 F1080
G1 X3.604 Y-5.108 E0.0021 F3000
G1 X3.737 Y-5.082 E0.0043
G1 X4.413 Y-4.919 E0.0222
 
Reply
#69
Ah good catch, I'll add it also.

Ok 0.12 is out. Prusa Slic3r support and one bugfix. See post #2 for details:

Got the Mavin finally printed (Slic3r sliced), although it's clear that either the PLA is crap or it has absorbed some moisture. Hopefully it's the moisture, would probably fix the problems with the filament changing. Need to finally build them airtight filament containers...

[Image: 2017-10-06%2020.46.50.jpg?dl=1]
[Image: 2017-10-06%2020.47.21.jpg?dl=0][Image: 2017-10-06%2020.47.21.jpg?dl=1]
 
Reply
#70
(10-06-2017, 05:14 PM)spegelius Wrote: Ah good catch, I'll add it also.

Ok 0.12 is out. Prusa Slic3r support and one bugfix. See post #2 for details:

Got the Mavin finally printed (Slic3r sliced), although it's clear that either the PLA is crap or it has absorbed some moisture. Hopefully it's the moisture, would probably fix the problems with the filament changing. Need to finally build them airtight filament containers...

Awesome! Thanks again!

Just tested it out and the gcode output looks good. I'll do a test print tonight.
 
Reply
  


Forum Jump:


Browsing: 1 Guest(s)