Printing a scale model fence

[VulcanRaven1989]

Hi,

I designed a simple fence in Tinkercad to print as a scale model. I cannot get it to print perfectly. I tried different options in Bambustudio and played around with the patterns which makes it al lot better but not perfect. I always get a mix of straight and blobby lines. I think this is mainly because the pattern the printer uses, it makes small U shapes and connects these together which causes some blobs. I think if the printer would just draw straight lines from top to bottom and side to side it would look much better.

Is there a way to make this work? Or maybe any other suggestions. I am using a A1 mini with the default 0.4 nozzle, i am planning on buying a 0.2 nozzle for even finer lines. I have put in a new roll of PLA to make shure it’s not moist.

This is the design:

This is the best i can get it (yes its bend because i took it out immediately of the bed).

Thanks for the help!

[Olias]

Welcome to the community.

Unfortunately, nothing in your screenshots shows scale or dimensions. Can you post an image of the screengrab of the sliced model as it sits on your plate?

Or better yet, uploading the 3MF file will take the guesswork out of your question and allow for thorough feedback and suggestions.

[Lenyo]

I’m guessing the problem is due to the model being symmetric and you have an overhang between the fence and the fence posts no matter the orientation.

You need to make it so the part that is on the bed is flat all the way across.

So, I would either remove the overhang on the bed side of the fence or print it upright, maybe with a support on the bottom.

[MalcTheOracle]

I’m with @Olias and second everything he said.

Pure guesswork though, these could help.

1. Enable Wall type Arachne in the properties list first tab.

Or

2. Slow down the print.

Or

3. Use a 0.2mm nozzle

I’m guessing that you are hoping for fine details in what might be (by implication rather than photographic evidence) a very small model.

You must ensure all of the bottom surface touches the build plate or have supports. I would recommend flat against the built plate though as supports wouldn’t work well at the size you appear to be going for.

[VulcanRaven1989]

Thanks for the quick replies! I have created a 3MF file of the print when sliced.

I have already enabled Arachne & detect thin lines (when i don’t enable them both some lines do not get sliced).

I use the 50% option when printing, the normal 100% messes up the lines after a while.

The model is indeed very small, maybe i am pushing the limits of my printer but maybe there is a way to get it done.

Fence 0.25mm.3mf (75.2 KB)

[Olias]

Thanks for posting the 3MF file, this really helps a lot.

@VulcanRaven1989

maybe i am pushing the limits of my printer

You are absolutely correct. If you look at the image, your fence rails are 1 filament layer wide. There isn’t a whole lot more you can do, you’re pushing the limit of the technology, you’re asking the filament to move in precision that is as wide as the filament itself. That’s why you are seeing the results you’re seeing.

However, there are two things you can try but they likely will not produce appreciably better results.

1. Slow down the filament speed. In this case, it is the outer wall that you are looking to slow down.
   
   
2. The second is to change the outer wall thickness.
   
   

Although these will improve the print, they will not give you that perfect result you see in your CAD.

[VulcanRaven1989]

Thanks for the help! I have tried the settings but as you already guessed it does not get much better unfortunately.

Would buying the 0.2 nozzle help with this issue?

[MalcTheOracle]

@VulcanRaven1989

Would buying the 0.2 nozzle help with this issue?

As I already said…

@MalcTheOracle

Use a 0.2mm nozzle

Yes.

[hooonk]

Just out of curiosity, i think i would try a 6 or 8mm nozzle.
My line of tought is, if its one line of filament anyways, make it a bit thicker so it probably is more rigid. And play with the temperature/speed.

Also what plate are you using?

[lkraus]

The current nozzle path draws a bunch of rectangles and then connects them, then does the same on the second layer. At each corner the filament is forming a bit of an arc as it changes direction.

It may help to change your model to put the horizontal lines on one layer and the vertical layers on the second, so the nozzle makes longer straight lines.

Whatever you do, with just single lines, the “fence” will be very fragile.

[user_3026326371]

I had similar issues a long while ago, trying to print sort of an air filter grille.
Same problem though - the printed features were barely as wide as the nozzle diameter.
Did not get a perfect result, just to be clear here but maybe the same trick can help you getting a better result >

I recreated the model as a 4 part model.
Basically the first model was two layers thick and only had the ‘fence lines’ of the x-direction.
Second model had only the lines for the y-direction.
The next two just alternated.
Printed in 0.12mm layers.

The idea behind is quite simple - you force the printer to do just straight lines instead of the whole pattern.
It would also help if you could increase the extrusion width to let’s say 0.5 or 0.6mm.

[VulcanRaven1989]

Thank you for the responses. I wil try the 0.2 nozzle just to be shure, i print very small items so it will always be usefull for me.

About the changing of the model to use fixed lines per layer. I understand the idea but not entirely how to create this. Is this something that i have to do in the drawing of the model itself? And if so, won’t the top layers be floating over the layers below creating the need for supports?

The sturdyness of the fence is not so much a issue, it is a static scale model. It is going to be placed and then never touched againt or have any sort of pressure on it.

[user_3026326371]

Ok, let me try to explain this layer thing a bit more…
If you create a model with just the vertical features and another one with the horizontal features of the fence you can mix them.
Basic shape, place, extrude up by 0.1mm or just 0.05mm and repeat in the other direction.
Yes, the slicer WILL complain about floating areas…
But if you design in alternating thin layers the filament will just sage a bit and with a slow speed the result is still quite acceptable.
You only need a few layers anyway.
By doing one direction at a time with just uninterrupted lines the print won’t struggle with the usual extrusion issues for those too fine details.
Might still be more like a spider web but should be printable this way if the 0.2mm nozzle won’t do it the normal way.

[Olias]

I just found a bunch of photos that may be germane to this topic.

A few months ago, I posted something here in response to a question about speed optimization. I can’t find that post, but I just found the pictures.

Let me explain what we are looking at. This is not a CAD-generated file. Instead, it is a cube primitive that is 1mm tall. I set the top and bottom levels to zero and used the grid infill.

So, what exactly does this show? An imported file, whether STL, OBJ, STEP, etc., is always converted to a mesh. That’s how slicers work. However, if the slicer calculates geometries, it will be as precise as the X-Y nozzle movement for that printer allows.

In this example, you can see the limitations of a grid laid down as infill. This demonstrates the limits of mechanical movements. Even with the absolute “best” nozzle movements and calibration, the results, though uniform, also show uniform displacement. I found no way to remedy this.

Here is the sample grid at scale. If memory serves, this was a 75x75x1mm cube primitive.

Here are some zoomed in versions

This side is actually an example printed on a specular(mirror like) build plate that I use when printing optically clear models. This was printed at very slow speeds but you can see how the filament achieved a glass-like smoothness where it was in contact with the plate. I believe this further illustrates what the limits of a 0.04 nozzle can get to because in this case, the layers were laid down at 0.28 layer height which meant now “squishing”. Again, you can see that these columns on the left are very uniform but the intermediate lines, although uniform, still show start and stop extrusion that can’t be avoided.

Here it is under extreme magnification backlit. Note that the grid was uniform in that one can see the alternating shapes but far from the square grid pattern I wanted.

And as you can see here, there were random locations that had some blockage in the grid. This was a precisely calibrated filament too, or at least as far as my skills could get it.

If you click on this zoomed in photo, you can see that the grid pattern is at the 1mm resolution which it appears you were trying to achieve.

[VulcanRaven1989]

I received the 0.2 nozzle yesterday and tried it out. The lines are a little better but not much change unfortunately. So i tried the layer option and created a fence with the vertical lines 0.5 mm lower as the horizontal lines and tried to print this. Below is the result:

The design was quick and dirty and i haven’t played around with the settings + the picture is realy zoomed in but this is a huge difference. With the naked eye you can’t see the sloppy lines and it is almost exactly what is wat looking for! I am going to tinker a little bit with it but this is already a huge step in the right direction!

Thanks for the help and advice again!

[MalcTheOracle]

I am glad you took my advice.

[Olias]

That is indeed much finer resolution. It makes me want to get a 0.20 nozzle just to try out even though I haven’t personally had a use case for it.

However, even though it’s an improvement, if your objective was to create a scale-model of a fence for use in something like a Sketchup model to show off a concept, I’m afraid this does not look like a fence.

Since you have the finer resolution. There is one setting that I was recently playing around with for a completely different project. Why not try out thickening the wall lines? It will be interesting if the affect is different when the slicer has more dwell time over an extruded measure.

The setting I am specifically referring to is here:

This is somewhat of an absurd example but it visually gives you and idea of what the slicer attempts to do. It would be very interesting to see what affect that has with a 0.20 nozzle.