Feedback wanted! Experimental new illumination mount

General
#1

Hi everyone,

I have been experimenting with a new way to mount the illumination. The downside is it requires an Allen key to adjust, upside is that it should be more sturdy.

Like all good OpenFlexure designs, there are of course flexures in it! In this design the illumination mount has a big slot in it:

image
image793Ă—691 4.64 KB

and then the condenser has a matching shape on the back. The matching shape is slightly undersized but can be tensioned inside the slot with bolts pulling on nuts inside the condenser:
image
image793Ă—691 7.6 KB
.

I have have made a little video to explain my thinking, it also shows the evolution of this design and how it failed! I would be really interested in feedback. Either here or on the merge request thread on GitLab.

I am also thinking that we can block the bottom of the channel to make the microscope more robust for travel. It can also hold more weight if we want to build bulkier, more advanced illumination schemes. I have also played with designs for holding extra optics between the condenser and the sample:

image
image793Ă—445 5.88 KB

Let me know what you think!

2 Likes
#2

I like it.

“The downside is it requires an Allen key to adjust” we can print knurled knob heads to go on the cap-head screws. Like these https://uk.rs-online.com/web/p/clamping-knobs/0771651/

The other upside is that the condenser is narrower, so the stage clamps will fit alongside it and you can put a slide in more easily from the unobstructed side. With the current wide v-groove the condenser is too wide to allow the clamps to fit beside it, so you end up with clamps from one side and the condenser coming over from the other side, and no way in for a slide :frowning:

The neck in the middle of the double dovetail looks quite thin. Is it stiff enough to prevent the part attached to it from bending sideways? It probably is.

1 Like
#3

It is if only because the flexures will bend first. But the flexures are under some tension already so the bending is not much. I would estimate from simple poking that the force to move it about 1mm is similar to the force to detach the current condenser entirely.

I like the idea of the knurled knobs. I think they will be more reliable if we make ones which which accept a hex head bolt like the the gears do.

#4

I have used the ones that go into a cap head. They work really well, though the ones I have are not 3D printed. I have them on blackened steel cap-heads, which tend to have a slight knurling on the outside of the head which may be helping. Keeping to cap-heads would keep the part count down, unless they are already a different length from what are already used, or even 30mm long like the current hexagonal screws.

I’ll try designing something.

#5

Yeah I am worried about tolerance and about saying that the caps must be knurled. Due to the depth of the dovetail I am using M3x16 which is a new bolt anyway. We can try to make the knobs stand off enough that they take up the extra 9mm and we can use the m3x25 hexes from the columns.

Edited to say: It would be really cool if we can do thumb grippers which are less than 8mm tall so we can have thin optics right under the illumination all using the same ones.

1 Like
#6

I like this idea, and I second the point with the clamps as I’ve run them into the current condenser.

However, from a practical POV I think I’d like a design with a single fixing screw more; If you were to put a V groove into the back of the mounting tower,the condenser mount could be pulled into it, aligning it vertically, and also requiring less hands to adjust it.

Something along the lines of this sketch, viewed from the top:

mount

There’s also more meat for a slot at the back of the mounting tower, avoiding your issue with cracking when tightening the screw.

regards,

cm.

#7

Yeah, I thought about this. What I worry about is that once the screw is loose the thing just wants to fall. My plan was some pre-tension so it can sit firm while fairly loose. It would be interesting to compare how they are to use. I started thinking about springs for pre-tensioning, but we have generally found springs are really hard to spec for distributed manufacturing.

#8

What would be nice would be a screw that switches from left hand thread to right hand in the middle so you can drive both nuts by turning one thing. Totally possible, beautifully simple, but also very non-standard and so the wrong solutions for OF. :crying_cat_face:

#9

Yes, I’ve tought about this too. One solution might be a combination, where a flexure provides the tension while the V slot does the clamping.

Another thing might be that it doesn’t drop freely if the screw is loose – the layer lines should provide quite some friction.

I’m also thinking about whether your design shown above could just be changed to be screwed from the back instead of from the sides, providing a combination of both approaches. The center square between the flexures might need to be wider to allow for the screw.

regards,

cm.

#10

That’s an idea. Think about your previous drawings I prefer the idea of a dovetail this way:

image
image1219Ă—1008 10.8 KB

With the bumps being flexures. The thing is a screw from behind would loosen not tighten. I don’t want to rely on the plastic as it will creep, or on just having a screw press into it as that mechanism wont slide well.

#11

Agree that something better is needed! I like the idea of this, and it certainly looks more study, but the part I can’t work out from the design is how easy it is to keep the condenser central. If I tighten one side up and then the other side, will it remain in the center, or would I need to tighten both sides at the same time?

#12

There is less than 1/2 mm of movement. But as you are pulling both sides into contact, how central it is will depend on the relative stiffness of the two sets of flexures, not how tight you do the bolts. It won’t be perfect to the nearest micron, but nor is the current system, this can then be adjusted at the back.

#13

ah great, sounds good!

#14

@samuelmcdermott Worth noting that this is how it should work. We will have to test if it actually is a problem in reality.

@rabbit I have been thinking a bit more about one vs two screws. The illumination adjustment is done on initial setup, and perhaps very occasionally realigned, but it is not done during normal use. So having 2 screws may make it more stable to fully lock it in position. We then probably want some sort of cap to go over the top and stop things getting into the dovetail and to guide the wiring. This can come down over the side and cover the screws. This will be especially important when we make a circuit board for the illumination.

1 Like
#15

For x-y adjustment, the whole illumination riser has the slots. The adjustment here is only z, so if it goes out in x-y when you tighten that is a small pain, but not a disaster.

#16

I started, ages ago, a version of the illumination mount that used the same fitting as the optics module, i.e. a single screw from the back that pulls a convex V-shaped part into a concave fitting. This should have the same benefits as the optics module (self-centering, narrow, sturdy, single screw), the downside is that loosening the screw enough to adjust it also results in quite a lot of wobble. My ideal solution would be one where the locking screw doesn’t move the dovetail at all (I realise that is unlikely in practice, but minimising the misalignment would be great).

#17

I just had a look at my collection of dead branches: my previous attempt is in illumination.scad on the branch bolted_illumination. I’ve pasted a couple of pictures below.

image
image

This looks to me like it’s just missing the nut trap on the condenser, I don’t ever remember printing it so I definitely haven’t tested (or finished) it! If anyone fancies finishing it and giving it a go, though, it might be a sensible option.

#18

Do you have an image of the back? Is that slot for the screw head and a washer?

#19

It might be worth figuring out what we want from the dovetail/optics rail and thinking about it for both the optics module and the condenser together. This is particularly relevant for e.g. the delta stage, where it might be possible to print the rail for the optics module and the condenser as part of the same print, so they’re co-aligned, and then put a small amount of tweakability into the condenser so it can be fine tuned (maybe with some flexures?).

1 Like
#20

@j.stirling here’s the back, looks to me like a “keyhole slot” similar to that on the objective mount, just for a screw head. I agree having a washer would be better (though that’s more annoying when you’re using it keyhole-style). Maybe a keyhole slot, with the hole big enough for a screw head, and enough space for a washer. That way, if you are using a washer (and putting the screw in from the back) you won’t lose it accidentally, but you can still use the keyhole slot if you want to?

image