small perimeters destroyed as print begins 1.1.7 and 1.2..5

My designs are more and more using a hidden nut and a M3 hole. So each print begins with a small perimeter. In both 1.1.7 and 1.2.5 the initial layer destroys the small perimeter as soon as it is created by pathing directly through the print and either distorting the hole or ripping it free of the bed completely. Both ruin the print. I have to check the pathing in pronterface and if it crosses, I just can't make that model. After a dozen failed attempts with different configurations I have given up improving my design for now. My older version paths clean, so I am stuck using it. This is 1.1.7 since it works better.
That stupid test piece prints, but the same piece in completed form refuses! Every time it knocks the second hole off the bed. I have "avoid perimeters" checked.

The code ignores small perimeters. Especially in 1.1.7 where even the speed is ignored for small perimeters - they print at normal perimeter speed (1.1.7)

---

Socrates ~ The Amsterdamman
slic3r-1.2.9

I have the same problem!

---

Socrates ~ The Amsterdamman
slic3r-1.2.9

paths better on a 45degree angle, not my first choice, but better than failure

Edited 1 time(s). Last edit at 01/22/2015 03:06AM by Amsterdamman.

---

Socrates ~ The Amsterdamman
slic3r-1.2.9

The checkbox "Avoid crossing perimeters" ("Print Settings">"Layers and perimeters">"Quality") is designed to stop exactly that from happening. Be prepared however for ridiculously long slice times as Slic3r can hang in the "exporting to file" phase for many hours (it eventually gets there). Due to the long slice time, beware of copying a partial G-code file to the printer before Slic3r has finished slicing (BTDTGTTS). Prints take a bit longer as the nozzle is sent on convoluted paths for non-printing moves (so as to avoid crossing perimeters!).

However it sound like your main issue is that the first layer is not adhering too well. Small perimeters adhere well enough on my printer to withstand the passage of the head on a non-printing move. Maybe try setting your Z zero (home) position slightly lower, or use a coating on the bed (e.g. hairspray or PVC pipe cement). You could also try slowing down the first layer speed ("Print Settings">"Speed">"Modifiers").

One other suggestion is to look into the G-code file with a text editor and try to find the move(s) that knock your perimeter off. You can then manually change each offending move to 3 or 4 separate moves that result in a dogleg around the small perimeter. As it is only an issue on the first layer, it is not that onerous a task.

As a last resort, you could cover the bottom of the holes with a 1 layer thick solid in the design stage, then drill them out after printing. This is what I do when I have a small hole through a bridged section, otherwise the printer attempts to print the hole perimeter in mid-air, and of course the bridging fails miserably.

Dave

Edited 1 time(s). Last edit at 01/22/2015 08:12AM by dmould.

Thanks Dave,
I already do have the "avoid crossing perimeters" checked. It takes longer, but no ridiculously much. I'm on a iMac 10.7.5 for me it makes much better prints.
It is also a stick issue, I was trying to use some of my "bad" filament as well, with my "good" spools it has not been much of a problem.
But for me the "avoid crossing perimeters" doesn't seem to work. The nozzle clearly paths through the small perimeter.
1.2.5 correctly uses the speed settings for small perimeters but paths worse.
So I still use 1.1.7 with a 25% slower 1st layer and a 25% 1st layer acceleration as well.
The small perimeters settings don't seem to work in 1.1.7 so they zoom at perimeter speed for the rest of the print, but seem to be ok.
The hole leads to a buried nut and a pause command, so it can't be drilled out after printing.

But as I say, this is more about the function "avoid crossing perimeters" not seeming to work for me in either 1.1.7 or 1.2.5
I found some rough workarounds, but such a great program should shine bright from all the polishing!!
This flaw is the cause of most of my recent print failures and seems to be a bug in an existing function.

---

Socrates ~ The Amsterdamman
slic3r-1.2.9

Quote
Amsterdamman
The hole leads to a buried nut and a pause command, so it can't be drilled out after printing.

You say the hole is on the first layer (the bottom of the print), so I do not understand why it cannot be drilled out. Even if there is an embedded nut a mm or so behind the hole, you can remove a single covering layer of plastic pretty easily - perhaps not with a normal drill bit but certainly with some other tool. A single layer is weak enough to be able to just screw a bolt through it and into the nut, for example. Or use a much smaller drill or router, cut with the point of a craft knife, use a needle file or melt through the layer with a fine tipped soldering iron. Lots of ways.

Dave

Hey Mr. Amsterdamman,

I think I may have been able to solve your issue.

Open up the Slic3r config. Go to Print Settings--->Layers and perimeters--->Advanced--->Seam position select "random"

I did a quick code gen on my machine, you can see in the pic that the travels don't cross at all. Ive also put my version of the Gcode for your review

On the other hand, the fact that the nozzle passing above the perimeter "destroys" it indicates you have an adhesion problem.
Especially on "bad" filaments, a good bed leveling will make the difference.
Depending on the printer type, you may adjust Z axis or bed while executing the skirt.

Another alternative is to provide the design with a "printing pad", which is a single layer completely filled. It can be exactly as the piece or wider (if piece is larger and you also want to prevent warping). After print, you just use a sharp knife to cut around all contours.
Such feature will be present in a future version (accepted as a feature request a couple of months ago).