gwhite
Posts: 214
Joined: Wed Jul 19, 2017 9:37 am

Re: Gaps in Perimeters (Again)

Sun Jul 25, 2021 6:59 pm

gwhite wrote:
Sun Jul 25, 2021 3:01 pm
Just a quick note: Apparently Version 4.0 has no trouble loading a file saved in 4.1. I closed 4.1, and opened 4.0, and it immediately loaded the file I had just been working with. I've got the G-code on an SD card, and will run a print as soon as the machine is available.
Just got the print finished. It has the same gaps, so the issue is not a bug that was introduced in Version 4.1.

gwhite
Posts: 214
Joined: Wed Jul 19, 2017 9:37 am

Re: Gaps in Perimeters (Again)

Sun Jul 25, 2021 7:00 pm

maker2 wrote:
Sun Jul 25, 2021 4:08 pm
Very interesting. Were there other places on the 1st layer where retractions / immediate restarts occurred to compare with?

And you have 0.2mm of "retraction vertical lift" enabled. Can imagine if the printer had a sticky Z axis /backlash it might not come down again as promptly as the Z axis movement speed of 20mm/s would suggest.
I haven't checked the bottom for evidence of similar effects. I can do that while I'm running Airscape's new file.

gwhite
Posts: 214
Joined: Wed Jul 19, 2017 9:37 am

Re: Gaps in Perimeters (Again)

Sun Jul 25, 2021 7:02 pm

airscapes wrote:
Sun Jul 25, 2021 4:55 pm
The attached factory file was used to print the part in the attached image. Please load and print without changing anything and see if you gaps are gone. I would consider this part perfectly acceptable, this is FFF and it is far from perfect. The more we try and make it so the more issues we create. The makergear printers are very accurate they don't need a lot of software tweaks to do good work..

No gaps .jpg

Left Rear Block Airscapes NO Major Gaps.factory
Thanks! I'll run it as is and report back this evening.

gwhite
Posts: 214
Joined: Wed Jul 19, 2017 9:37 am

Re: Gaps in Perimeters (Again)

Sun Jul 25, 2021 7:57 pm

maker2 wrote:
Sun Jul 25, 2021 4:08 pm
Very interesting. Were there other places on the 1st layer where retractions / immediate restarts occurred to compare with?

And you have 0.2mm of "retraction vertical lift" enabled. Can imagine if the printer had a sticky Z axis /backlash it might not come down again as promptly as the Z axis movement speed of 20mm/s would suggest.
OK, while I'm waiting for Airscape's print to finish, I took a look at other regions on the bottom. There are two internal holes in the model, and the photos I've been taking have been of the perimeter around one hole. I get the same issue of the 2nd perimeter underextruding & leaving gaps around the second hole:
2nd Under Extrusion.jpg

gwhite
Posts: 214
Joined: Wed Jul 19, 2017 9:37 am

Re: Gaps in Perimeters (Again)

Sun Jul 25, 2021 9:17 pm

airscapes wrote:
Sun Jul 25, 2021 4:55 pm
The attached factory file was used to print the part in the attached image. Please load and print without changing anything and see if you gaps are gone. I would consider this part perfectly acceptable, this is FFF and it is far from perfect. The more we try and make it so the more issues we create. The makergear printers are very accurate they don't need a lot of software tweaks to do good work..
We have a winner! Thank you Airscapes! I really appreciate the time you took to help out with this.

I just fished the print off the printer, and the gaps are gone! Here's the top side:
No Gaps (Top).jpg
And here's the bottom, showing a clean start to the 2nd & third perimeters.
No Gaps (Bottom).jpg
Despite all sorts of changes, the only minor "quality" issue I noticed is some "ghosting" of the internal voids that wasn't present with my settings:
No Gaps but Ghosting.jpg
This has Airscapes' version on top of one of mine. The depressed region is pretty small (small enough it's hard to photograph), and for the project at hand, it won't be an issue.

I want to go through the gapless process and see if I can identify the culprit(s) in mine. I've used pretty much these settings on a bunch of projects with good results, and at first glance, the process I used shouldn't break anything like this.

The good news is that I won't have to wait for S3D to fix anything, and I can avoid PrusaSlicer a bit longer...

I'll report back if I identify exactly what parameter(s) made the difference. I suspect it will be only one or two, and with luck, I can have no gaps and no ghosting.

airscapes
Posts: 374
Joined: Tue Mar 06, 2018 10:35 am
Location: Philadelphia PA Area

Re: Gaps in Perimeters (Again)

Sun Jul 25, 2021 11:03 pm

gwhite wrote:
Sun Jul 25, 2021 9:17 pm

Despite all sorts of changes, the only minor "quality" issue I noticed is some "ghosting" of the internal voids that wasn't present with my settings:

I'll report back if I identify exactly what parameter(s) made the difference. I suspect it will be only one or two, and with luck, I can have no gaps and no ghosting.
If you don't see a change, those settings did not need to be tweaked.. just because they are there does not mean they need to be changed. This software is made to work with all kinds of printers and what is needed for one (Bowden remote drive) is not needed for direct drive.. But you are going to procced to fiddle no matter what.. Have fun, I know I didn't..

gwhite
Posts: 214
Joined: Wed Jul 19, 2017 9:37 am

Re: Gaps in Perimeters (Again)

Mon Jul 26, 2021 9:25 am

Your process is definitely simplified considerably from mine. I've spent a lot of time over the years making adjustments to deal with various print quality issues. The changes have accumulated over the years, and I could easily have slowly waltzed off into the weeds, or to the point where I stumbled across a bug in S3D.

For example, the reason I have the vertical lift is left over from a project with a tall skinny part, with several "spires". Without the vertical lift, it would occasionally clip one of the spires en route to another one, and actually knocked the print loose. I didn't see that it did any harm, so that got left turned on. In retrospect, it probably adds a bit more wear on the Z drive, and I will leave that off until I need it again.

The values I have for retraction and thin wall settings are not just an accident. I do a lot of printing with embossed text on the bottom of things, and getting the text to come out reasonably clean requires tuning those just right. I've spent a lot of time printing sets of test parts, sweeping one parameter at a time to get an acceptable result. Of course, it my be that I've been fighting whatever caused the gaps in this job without knowing it.

In any event, I'm going to export all the settings into XML, and then into Excel from my profile, your gapless profile, and the default PLA profile that came with the printer. I can then do a detailed comparison of all the different settings to see what the differences are. I can then take the gapless process, and slowly add in some of my settings to see if they help anything. If not, out they go.

I'll post a summary of the process differences once I have that tabulated cleanly.

airscapes
Posts: 374
Joined: Tue Mar 06, 2018 10:35 am
Location: Philadelphia PA Area

Re: Gaps in Perimeters (Again)

Mon Jul 26, 2021 11:43 am

Yes, managing one off changes can be difficult and they can bite you when you don't expect it.
I tend to save factory files, I have no default process or profile.
My factory files are named for the part and include the nozzle size, layer height and type of filament.
I will just load a factory file for a similar part, remove the old model, import the new and save it with a new name so as not to accidently overwrite the original. Then reopen the new name and make any adjustments that I think are needed for this particular model. Lets face it, the only reason to change anything if you already have the filament, nozzle and layer stuffy set, would be model specific.

Even this method is not perfect and you can miss a boxes checked you don't need checked this time when looking over each tabs..

I was trying to pull the things off or your process one at a time but was burning my entire day so I just ripped out all the stuff I know is not really "needed" to get a good solid part..
Again, I prefer imperfection in finish over imperfections in structure and accept that this is a very imperfect fabrication method.

airscapes
Posts: 374
Joined: Tue Mar 06, 2018 10:35 am
Location: Philadelphia PA Area

Re: Gaps in Perimeters (Again)

Mon Jul 26, 2021 11:51 am

BTW, I would think the main issue was the negative restart distance .. the nozzle and drive on the m2 works very well when retracting and restarting. Anything but zero in this field will cause under or over extrusion after the retraction.. same goes for coasting.. not needed with this hot end/drive setup that I have ever found in any part with any nozzle size. Different printers have different mechanical issues that need software to over come.

gwhite
Posts: 214
Joined: Wed Jul 19, 2017 9:37 am

Re: Gaps in Perimeters (Again)

Mon Jul 26, 2021 12:17 pm

I'll find out.

Extra restart distance is VERY sensitive, but a value of -0.05 shouldn't break anything. It does improve the quality of seams noticeably if you have something fussy, but if you use it, you should do a careful test to find the right value. I have a 9 process file that runs a 3 x 3 grid with varying retraction & ERD to home in on the best settings when I need something super clean. Sometimes it's for esthetics, but it can also improve the fit of mechanical things depending on where the seams end up. I use dovetails a lot to build larger pieces, and having the ERD right can sometimes minimize the hand fitting required to get a tight joint.

I now have your gapless profile, my gappy profile and the default PLA profile exported and converted into tables for easy comparison. This afternoon I will get them all into Excel, clean up the extraneous stuff and do a detailed analysis of the differences I need to explore.

Return to “Troubleshooting and Bug Reports”