I am about to do something stupid again, but I want to get some feedback before I commit to it. (who am I lying to? I may do it anyways)
I always wanted to get more range of motion. In the past, I tried the all TPU unbreakable openflexure microscope and a hybrid PLA/TPU one with questionable performance Indestructible body -BUT DONT TRY THIS AT HOME- - Contributions - OpenFlexure Forum
Since then, new filament formulations, improved printers, and slicers have changed the game. I am thinking of retrying the hybrid approach by using a flexible filament on the flexures and PLA in the rest. However, I am thinking that even if I find the right filament combination, I may encounter some features in the design that may block the movement of the columns. This image from my TPU model probably explains it better.
My questions are: How those crash points can be modified? will the motors have enough torque to move that far? do I need longer screws? what else might need to be changed?
1 Like
There are more collision points than just that. We have been working on this quite a bit recently as we noticed that the actual range is not set correctly #444. I made changes in !512 to address this.
The range of motion overall is set by a parameter flex_a. However in some places the angle of 6 degrees is hard coded - that wall is one of them I think.
The biggest problem will be the space for the stage to move towards the illumination dovetail. That probably needs to move back, but that causes problems with the rest of the z-axis shape.
You probably don’t need longer screws.
1 Like
It feels like a big redesign. You are right, lots of other collision points I did not consider like the dovetail and the cable channels. I also imagine the stage will move down a bit as it moves further out. This will need a lot of Z-axis compensation to keep everything in focus.
z should be OK, the change is h - hcos(theta). We have ±2mm travel in z, and the legs are 60mm high. So that should allow nearly 15 degrees of motion, over twice the standard 6 degrees. And of course with more angular range in the flexures, z would also be allowed to move further than 2mm…
There would be a lot to change to make it work fully, but not so much for a first test whether the idea works at all. The worry is that it would be wobbly. You can check that, even if it does not actually move any further than normal.
My thought would be to isolate the flexures in the SCAD as a separate part, and print on a multi-material printer so that only the flexures are in the flexible filament, not a whole layer.
That is a great idea. I will try just the hinges and not the whole layer. This may work better. I am not familiar with openscad at all (too much math for me). Using the slicer paint function may work as well.
At the top of the legs you will need the whole long bridge in one material. That might fail in TPU, but it could be possible to have a first layer of PLA for the bridging and then flexible layers. Or you need to put in a few custom supports on that long bridge under the stage.
On other flexures it would help for them to be a shape that the slicer will lay as parallel strands - so long and thin across the flexure.
the print failed. the flexures are too small to print like this.Too hard to make modifications using Google SketchUp, and Fusion 360 will not convert the model appropriately 
Just pushing us along the road towards mutual madness 
:
Use this body - main_body_with_box.stl (1.7 MB) as the main part, and then use this
flex.stl (33.7 KB) as a modifier in the slicer to change the filament type for the flexures (including extensions of the flexures into the main part). Both STLs have a bounding box so that all being well they will align correctly.
I have tried slicing in Prusa Slicer for the XL with tool changers, and it looks as though there are no major places that it should fail…
This only does the flexures around the xy stage. It does not do the flexure at the base of the actuator column - which is the one under most stress, nor the z-actuator. Still, it would tell you if it will print, and it will tell you something about motion stability over the normal range of motion.
And do remember to bear in mind the title of your thread at all times!
1 Like
I modified the modifier to include all of the flexures 
all_flex.stl (53.6 KB)
Thanks!. but how am I supposed to use this? I need to find a way to subtract this flexures from the main body model and replace with this ones I guess. I was doing this manually. I was planning to use a little hack to make the printer think this are different objects. Essentially I offset the flexure faces by 0.04. This is enough to make the slicer think this is a separate objects and still have a small gap that allows melting the materials together. I did one have a lot more to go
In Prusa Slicer you can select an object and right click to add a modifier. That can be a simple cube or you can load an STL, like the flexures file. The modifier can be subtracted from the shape if you want to cut out the flexures, or the modifier can be used to adjust the printing parameters, in this case to set the print to a different filament.
I don’t know how Bambu works, but it is based on Slic3r like Prusa Slicer and it must have something similar.
2 Likes
I did it! Loaded the main body, right click and then add a modifier loading the flexure STL file.This may work!
2 Likes
Excited to see how this turns out.
If it works well we will need to think about if delamination between layers affects longevity. If it does, we can think of ways to adjust the flexure layer so the “hard” material pins the flexure material in more places.
1 Like
Nope…did not work. the flexures easily detached when trying to remove the object from the bed. Also, really bad bridging failure
1 Like
The bridges were a worry, they are long even in PLA. There are various things that might help both problems, but I think making something reliable this way is seems unlikely. It has been very useful to see the multimaterial print tried.
