gap between perimeter and infill

1,117 views
Skip to first unread message

rjschave

unread,
Dec 27, 2012, 2:37:23 AM12/27/12
to make...@googlegroups.com

I've been working on tweaking my Replicator 1 with 20mm calibration cubes with ABS.

It took me a long time to realize the filament packing density was causing the bulge on the top of my cubes.  Now that I've got that resolved, I'm working on the gaps that appears between the perimeter and the infill (the gap moves based on the number of shells I have).  The gaps only appear on two sides and those sides change based on the direction the extruder is moving.

With 100% infill, the gap is present on every layer.  I was adjusting the K2 value but the only change I noticed as I moved from a value of 0 to .0050 is that the bottom got smoother and the gaps disappeared on the bottom layer.  The gaps are still present on the top layer so I don't believe the K2 value is the correct parameter to be adjusting.

The cube on the left has a K2 value of 0.  The cube on the right has a K2 value of 0.0050.

Here are some of my other skeinforge settings:

Edge Width over Height - 2.0
Infill Width over Thickness - 2.0
Layer Height - 0.2
Filament Diameter - 1.83 (measured with caliper)
Feed Rate - 30.0
Flow Rate - 30.0
Filament Packing Density - 1.83

Does anyone have any suggestions for removing the gaps?

Thanks.
20mm_cali_cube_k2_bottom.JPG
20mm_cali_cube_k2_top.JPG

Cymon

unread,
Dec 27, 2012, 10:35:01 AM12/27/12
to make...@googlegroups.com
I want to talk about your packing density for a sec. If I remember the default for that is < 1. I set mine at 1 but your way above that. I thought that value was supposed to correct for filament lost in the extrusion process. Am I fundamentally misunderstanding what this value is for?

rjschave

unread,
Dec 27, 2012, 10:49:00 AM12/27/12
to make...@googlegroups.com
Sorry... it was late.  My Filament Packing Density is 0.95.  Anything less and I get bulges on the top of the cube with 100% infill.

Dan Newman

unread,
Dec 27, 2012, 11:36:31 AM12/27/12
to make...@googlegroups.com

On 27 Dec 2012 , at 7:35 AM, Cymon wrote:

> I want to talk about your packing density for a sec. If I remember the default for that is < 1. I set mine at 1 but your way above that. I thought that value was supposed to correct for filament lost in the extrusion process. Am I fundamentally misunderstanding what this value is for?

At the end of the day, it's a fudge factor. You can argue that it's for this effect or
that phenomenom, but actually it's a number of physical effects *and* empirical calibrations
all rolled into one. Yes, ABS has some volatiles which disappear when the ABS is molten
and thus there's some volume loss. And yes ABS has much more die swell than PLA. And
yes, ABS actually shrinks some as it cools. And yes, you may not really get the expected
steps per mm from you extruder for ABS vs PLA owing to how much the pinch gear teeth
sink into the plastic and thus reduce the circumference of the circle associated with a
complete 3200 microsteps of your stepper motor. Etc.

Dan

Fastrack

unread,
Dec 27, 2012, 4:02:46 PM12/27/12
to make...@googlegroups.com
What do you have your Infill Perimeter overlap (ratio) set to (Skeinforge Fill tab)?  I believe the Makerbot default was .3,  I use .25, but from your picture it appears your infill isn't even protruding into the shells.

I personally have not touched the settings for Edge Width or WoT, as I let Print-o-matic handle that.  My cubes are perfectly flat on the top.  I use about .93 or .95 with ABS, depending on the plastic for my packing density.

I would bet Skeinforge is saying with an edge width of 2.0 it doesn't have "room" for another line of infill - yet the plastic can't really fill an edge width of 2.0.  I would try an edge with of 1.48 (or enable print-o-matic) keeping .95 packing density and try a test cube.

Ben

rjschave

unread,
Dec 27, 2012, 4:27:59 PM12/27/12
to make...@googlegroups.com

Ben,

My Infill Perimeter Overlap is .3 as well.  I will try adjusting that value along with reducing the values for Edge Width over Height (EWoH) and Infill Width over Thickness (IWOT) as suggested.

If you specify a nozzle diameter of .4 mm and a layer height of .20 mm then Print-O-Matic will use 2.0 for EWoH and IWoT  ( 0.4 / 0.2 = 2.0 ).

Thanks.

Fastrack

unread,
Dec 27, 2012, 7:53:22 PM12/27/12
to make...@googlegroups.com
Good point... I've never tried .20 layer height.. But you correct on the formula.

I'd try a cube with .27 height and the packing density you have. If that works maybe it's related to the ewoh and wot. Skeinforge calculation error...?

Is the cube 20x20?

Ben

rjschave

unread,
Dec 28, 2012, 1:47:11 AM12/28/12
to make...@googlegroups.com
Ben,

I've been testing with a 20x20x10 cube.

I tried a .27 Layer Height but got the same result.
I tried a .25 Infill Perimeter Overlap.  It didn't remove the gaps, but I liked the surface better, so I may stick with this change.

Based on your comment that Skeinforge may not think it has enough room for another line of infill, I went back to my settings posted above and printed a 21x21x10 cube.  I have no gaps on the top and only 1 gap on the bottom.  The gouge in the corner on the bottom is from the spatchula I used to remove the cube from the platform.  It seems this confirms your idea.  With that in mind, does anyone have any suggestions for getting around this?  Or do I just design around it / live with it?

The one corner edge is not as smooth as the other 3.  This is where the extruder is when the z axis changes.  I believe you guys are calling this a zipper.  Any suggestions for eliminating this or at least reducing it?

Thanks.
21mm cube bottom.jpeg
21mm cube top.jpeg

Dan Newman

unread,
Dec 28, 2012, 10:42:09 AM12/28/12
to make...@googlegroups.com

> The one corner edge is not as smooth as the other 3. This is where the
>
> extruder is when the z axis changes. I believe you guys are calling this a
> zipper. Any suggestions for eliminating this or at least reducing it?

In the Fill plugin use "loops > perimeter > infill" so that the zipper
is inside the perimeter not on the perimeter. Also, you can adjust either
the deprime setting used by the firmware (RepG's Machine > Onboard Parameters),
or if you have deprime=0 then adjust the restraction distance in SF's
Dimension plugin. NOTE that you should not use both deprime and retraction.

Dan

Fastrack

unread,
Dec 28, 2012, 11:00:46 AM12/28/12
to make...@googlegroups.com
So... I just looked a few of my test cubes (I printed a lot of them.. LOL)

I have a slight gap just like yours on each end!   So I think it's the slicer, it can not "fit" another line of infill.  However I have no gaps on the bottom.  I believe your bottom issue might be related to your nozzle height of the HBP.  I used to use a feeler gauge of .005" (.127mm) .. I now use a .008" (.203mm) gauge (back left or right screw on the HBP).  I then throw on my Dial indicator and level the platform... The first layer works out to exactly .27mm.  Of course depending on your layer thickness you'd have to change the nozzle height (different feeler gauge).  I have an no issues with lifting corners, some need a squished first layer to prevent this.

If you truly want to get rid of that.. Infill at 45 degrees will very likely get rid of the gaps (but I personally don't like the look of it.. it's not as smooth).
Fill -> Infill Begin Rotation ->  45 instead of 90

For your fill type... If your doing 100% your best to use RECT instead of HEX.  There's a bug with HEX that uses too much plastic when doing 100%.

I agree with Dan.. 
Loops > perimeter > infill is what I use.. I believe that's Makerbot's default.

As for the Retraction/Restart... That is one thing I spent quite a bit of time tweaking...
Dimension Tab
- Retraction Distance (mm) : 0.5
- Restart Extra Distance (mm): 0.0

I found with retraction distance set to 0 I had a ton of "strings" so I settled on .5

I'm not sure what Deprime is... However I've seen this mentioned a lot in the Sailfish discussions.. I have yet to update to sailfish... I will one of these days :)

Ben

rjschave

unread,
Jan 4, 2013, 2:26:35 AM1/4/13
to make...@googlegroups.com
Thanks for the info Ben.  I'll incorporate this into my next round of fine tuning.

David Wilson

unread,
Jan 4, 2013, 4:19:58 PM1/4/13
to make...@googlegroups.com
Hi Dan...

Sorry I've been away, distracted by Christmas/New year.

I'm continuing to tune my TOM, MK6 - Sailfish - printing with PLA 3mm. The picture below shows an initial start of print.


I'm seeking advice on what configuration to "enhance".
The issues I've numbered in the picture.

1. The initial start of print, after everything comes to temperature - and the extrusion across the print

2. The beginning of the skirt - and the extrusion of PLA - it doesn't matter really for the final print - though is symptomatic of the issue perhaps?

3. The beginning of each layer sees a burst of extrusion... as pointed to by the arrow...


Also, regarding the onboard preference (deprime) - On the TOM - is that the left or right extruder?


- David

p.s. ignore the splodge in the middle of the print - it's where I stopped the printing.

--



Dan Newman

unread,
Jan 4, 2013, 4:37:44 PM1/4/13
to make...@googlegroups.com

On 4 Jan 2013 , at 1:19 PM, David Wilson wrote:

> Hi Dan...
>
> Sorry I've been away, distracted by Christmas/New year.
>
> I'm continuing to tune my TOM, MK6 - Sailfish - printing with PLA 3mm. The picture below shows an initial start of print.
>
> <IMG_2054.png>
>
> I'm seeking advice on what configuration to "enhance".
> The issues I've numbered in the picture.
>
> 1. The initial start of print, after everything comes to temperature - and the extrusion across the print

That's the start gcode AND how SF-50 places the first segment to print. SF-50 as distributed
with RepG seems to favor the Replicator and it's inverted X & Y axes. Net result is that you
get a starting segment which is to the back left of the ToM / Cupcake's build platform. You
can change your start gcode to wait to the front left of the build platform instead of front
right so as to alleviate some of that.

> 2. The beginning of the skirt - and the extrusion of PLA - it doesn't matter really for the final print - though is symptomatic of the issue perhaps?

That may be the extruder ooze which built up? Does your start gcode print an anchor?
(The anchor typically gets that blob.)

> 3. The beginning of each layer sees a burst of extrusion... as pointed to by the arrow…

3 and 2 may be the same effect. Make sure that you aren't using both deprime and SF's retraction.
Also, do you have dual extruders? Is this tool 0 or tool 1? If it is tool 1, then you cannot
set the deprime = 0 for it unless (I think) if you have a LCD display.

> Also, regarding the onboard preference (deprime) - On the TOM - is that the left or right extruder?

It's Extruder A = Tool 0. Supporting dualstrusion on the ToM is still
somewhat experimental. However, since it seems to be working for folks
with Sailfish, I'll put on the to-do list to add the settings to the
onboard preferences for both extruders.

Dan

Joseph Chiu

unread,
Jan 4, 2013, 5:07:09 PM1/4/13
to make...@googlegroups.com
That looks like splodge plugin is turned on, or extra-restart distance under the dimension plugin.



Dan

--



Fastrack

unread,
Apr 19, 2013, 11:36:13 AM4/19/13
to make...@googlegroups.com
I know this is an old topic.. However did you ever fix it?  I installed Sailfish 7.3 and one of my goals was to fix this... However unless I use "too much plastic", ie not a smooth top ... there's always missing filament on each side... on every layer.. including the bottom/top.

Ben
Reply all
Reply to author
Forward
0 new messages