Print contracts below Bridge
Dear Forum!
I am developing a device for blind people to detect light. No joke. The device is for e.g. reading the LED status on an internet router etc.
The main body is a cuboid of 80x32x32mm size. The case has a fillet so I print it upside down so that the fillets are on top and the open end downwards. That works great! But when I put the inner parts into the case I notice that the dimensions do not fit any longer. The case has shrunken about 1 mm on the top side, below the fillet. I print with PRUSA PETG utilicing prusaslic3r with default settings.
In all further print the size of the top (roof) is exactly 32mm, as expected. The shrinkage is always below the roof.
I ditched the fillet for further investigation:
This is simply an empty box with a flat roof. One can clearly see the inward shrining below the roof. The thickness of the walls/roof is 2mm.
If I get rid of the roof I get a perfect box with 80x32x32mm. No shrinking at all! So it has nothing to do with printer mechanics, scaling, axes, servos, etc.
At first I thought about oveheating inside the closed box. The termal energie produced by the headbed inside the box can no longer dissipate upwards. So I tweaked the gcode to shut down the headbed at 16mm height (midprint).At the time the roof is printed the headbed was at 30 degrees. Exactly the same outcome - so heat jam is not the problem. Printer is in a thermocontrolled enclosement, so cold drafts can also be neglected.
I digged a bit deeper and focussed on the bridging.
These are frames of 32x6x(8,16,24,32)mm. Left side without support material right side with support. Utilizing a macro lens things become clearer:
8mm height:
16mm height:
24mm height:
32mm height:
The effect is already to guess at 8mm then it increases linearily with height. What stunns me most is that the support does seem to have no effect!
My guess is that the bridging which drags the filamanent across the gap puts enought tension on the anchor points that the are contracted. With increasing height the lever becomes longer and therefore the contraction more pronounced. The measured contraction rate is 3% which is close to the shrinking rate of PETG.
How to solve this? Any help appreciated.
Here the STL-Files:
And the gcode_files:
Cheers,
Volker
RE: Print contracts below Bridge
It's also just as likely your test is exposing the strength of the material: with equal forces, a 24 mm pedestal will bend less than a 32 mm pedestal.
The step in a wall is often due to print speed changes and cooling times the slicers uses. When it prints the walls, it's a very small area compared to the first bridged floor layer. You can examine this in slicer using the feature speed view. In filament there are adjustments for cooling that can help reduce these steps. There was a long thread discussion these "bulges" at layers where interior features change. The OP misspelled "buldges" - so it is easy to search for.
RE: Print contracts below Bridge
Just to eliminate the simple things:
- Did you measure and average your filament diameter and set that in your slicer?
- Did you calibrate your extrusion multiplier for that specific filament?
- Did you calibrate Linear Advance for that specific filament? LA can result in variations in extrusion widths if not calibrated.
- Does dialing back speeds with the front knob affect anything? I print exterior perimeters at 25mm/s for consistency.
Before going too far down the rabbit hole chasing issues, make sure the slicer is working with the actual filament values. I find surface quality varies a lot between filaments if I don't calibrate. In my experience, shifts at solid layers and layers with sparse infill are often due to variation between filament. Thinner walls can add to the challenge, but 2mm is thick enough to avoid those sorts of issues.
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
RE: Print contracts below Bridge
Other thoughts:
- Have you tried the same print with PLA just to compare material characteristics?
- Any chance you could upload the STL or a project file (3MF) with your settings?
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
RE: Print contracts below Bridge
Other thoughts:
- Have you tried the same print with PLA just to compare material characteristics?
I will try that. Stay tuned.
- Any chance you could upload the STL or a project file (3MF) with your settings?
The STL files are attached in the original post.
Thanks for the lightning fast response
Volker
RE: Print contracts below Bridge
Here's a simple view of the problem with the test part:
RE: Print contracts below Bridge
By the way - how are you handling the infamous RGB LED's? Simple monochrome photo detectors won't do much if green/yellow/red/blue all come from the same LED. Us color blind people would like to know... lol.
RE: Print contracts below Bridge
By the way - how are you handling the infamous RGB LED's? Simple monochrome photo detectors won't do much if green/yellow/red/blue all come from the same LED. Us color blind people would like to know... lol.
We will first go for the intensity. That is tricky enough. People tend to have a device that just does the right thing. They like to have a device that the can be put in the direction of the sun to get an information how far sunset is, or if it is cloudy. On the other hand they like to use the same device to check if there is the light on in the cellar, or if a LED is blinking on a device. There are several orders of magnitude in light intensity between these scenarios. We are yet working on the OpAmps to get this possible. Probably we will need differential measurements or we will need a second sensor for ambient light detection. All of this will make for the performance of the measurement but on the other side the device becomes more complicated.
A device to deal with colors is also on our list. With the new µProcessor Boards from e.g. NVIDIA featuring several hundreds of GPUs this looks like a simple task. Take a live video from a decent camera and push it through a neuronal net. We can do this, no problem.
But:
* The power consumption of these devices is enourmous. Maybe it would be better to include such a functionallity in a cell phone.
* And not to forget, you need to get training data for your KI. And only the big players have the data. If the maker community likes to exist in the future we should start to do not only open hardware but also open KI, open images, open measurements, open ciruits etc. Also we should work more closely to other sources of free data like the goverment and the universities.
Sorry for so much off topic info here.
Cheers,
Volker
RE: Print contracts below Bridge
@tim-m30
Sorry, but I am really sure that your argument is not correct.
The specific weight of PETG is so much lower than its elastic modules. Make an experiment. Print a part of 32x80x2 mm. Then put a weight on the middle of the supported far ends. You will have to put several gramms - I suppose >100 g to bend it about the 0.5 mm we are talking here. But the weight of the roof is wide below one gram.
Cheers,
Volker
RE: Print contracts below Bridge
Other thoughts:
- Have you tried the same print with PLA just to compare material characteristics?
I am currently printing the PLA equivalent. But I think this is futile.
PETG and PLA have more than an order of mangitude (25) different elastic properties.
What works for PLA will in most cases fail on PETG. What works for PETG will in most cases work on PLA.
My prints with PETG are really good. Additionally I adjusted today the bridge-parameters (not related to current problem!!!) to 100% FAN and 0.7 bridge-flow with really good bridging results up to the 40 mm gaps I am targeting. These parameters did not effect the shrinking behavior.
Cheers,
Volker
RE: Print contracts below Bridge
Sometimes theory is beaten by reality.
Here the PLA result - the same as the PETG.
bridges_0.15mm_PLA_MK3S_1h5m.gcode
Now I am really puzzled...???
Cheers,
Volker
RE: Print contracts below Bridge
You are correct: fact is always as good as theory. Note the DROOP in the bridged layer. If you straighten that bridged strand you will find the length is equal to the very top strand above it. As a result, you will also find the width of the walls when pulled in by that strand will "push out" that same amount and will be equal to the width of the top layer.
Warm plastic, that which has just been printed, is less strong than cold plastic. It bends more easily as weight is applied.
Place something rigid between the posts for this test and you'll find the walls no longer bend and the wall will be flush as expected.
Also -.gz is not compatible with most of the computers out there, whereas .zip is.
RE: Print contracts below Bridge
Try this little change: add something that prevents the sidewalls from moving and the surface will be even to the top.
What really hilarious is you can SEE the bend in the wall posts.
RE: Print contracts below Bridge
@tim-m30
So your are suggesting to build up some kind of stabilisator material that is later removed.
Really good idea. Will go into it. Stay tuned.
Cheers,
Volker
RE: Print contracts below Bridge
@tim-m30
I do not see any bending in the posts in your picture ??
RE: Print contracts below Bridge
RE: Print contracts below Bridge
Yep. This is the bend from my picture. It is obvious to see the bend there.
And now?
RE: Print contracts below Bridge
If your printing this upside down (open end downwards), then cant you pause the print and insert something to hold the sides out, then continue the print. once finished and the print is removed, the spacer can be removed from inside.
Obviously if your not staying with the printing, then this will not work, and you will have to use Tim's idea and have to waste filament.
Have you tried printing the first overhang layer slower and with cooling at 100% so that it has already shrunk some as it is being pulled across, It's a fine line between breaking and getting sucessfull first bridges though.
Normal people believe that if it ain’t broke, don’t fix it. Engineers believe that if it ain’t broke, it doesn’t have enough features yet.
RE: Print contracts below Bridge
@chocki
I will try this. But my changes in Fan-Speed (100%) alone did not have any influence on the effect. Will try to adjust print speed as well.
Cheers,
Volker
RE: Print contracts below Bridge
@volker
I'm at work now, but will have a play with this when I get home, what bridging speed do you use?, I noticed the material is PETG.
To be honest I have never printed anything like this where this could happen, but it will be good practice for when I do.
Normal people believe that if it ain’t broke, don’t fix it. Engineers believe that if it ain’t broke, it doesn’t have enough features yet.