I like the ball magnet mounts. A slight variation on the design I’m using on one of my setups is to stack a washer on a disk magnet of similar diameter and press fit that into a print washer side out. Then on the bolt side each blot passes through the heat set nut then gets capped with an acorn nut. The acorn dome to washer ring interface is self centering, 3 of those on a fixture and translation is locked down while still allowing the same fine adjustment.
Set-screws tipped with a ball-bearing and a bag of dowel pins would let you make quite solid 3D printed kinematic mounts. You can also buy the same screws and threaded bushings used in the commercial mounts. Makes it easy to design high quality custom mounts. Like a X-Y translation kinematic mount.
Thank you for sharing. I do a lot of 3D printing of mechanical parts. I use a PRUSA MK3S w/ MMU2 w/ 0.4 mm nozzle. Some tips that may or may not work depending on the slicer software you use. In advanced settings the width of the extrusion can be controlled to some degree. Make external perimeters 0.35 - 0.4 mm with a slower feed speed. Then internal perimeters 0.6 - 0.65 mm with a high feed speed and layer height 0.15 mm. infill at 0.7 mm. In the infill settings combine infill every 4 layers. This will give smoother exterior to your parts and better tolerance while maintaining fast print speeds. Also I use cubic infill with 30 - 40% density. Your magnetic ball idea is a good one, however to greatly decrease unwanted movement make the receiving side concave to fit the magnet. Magnetic PLA is quite effective for magnetic attractive surfaces and with a multi material setup you can use it sparingly where needed. Putting a 3D printed fence around the movable part at 0.15 - 0.2 mm backlash can also further constrain against unwanted movement. Use printed gussets and struts of various types everywhere with access holes for tools and fasteners to increase rigidity. It is always worth the extra print time to eliminate as many potential problems as possible. It is also much faster to print your own custom supports in the part design for overhangs as you can speed print time and reduce material waste with good design and no generated supports rather than automatically generated supports. This also gives you better control of interface gaps and tailor them to the material you are printing with. For very low density infills you can force the slicer to add perimeters where you want solid infill by making very thin voids at 0.05 mm. This takes a bit of practice as it depends on the printer, slicer and filament type but you can make voids that are so small that they are practically solid and fused but mathematically the slicer sees a perimeter. This will force the low density infill around these voids with a perimeter shell. For locations with a high chance of breakage like a low diameter post on a plate, make a hole for a flat head machine screw to self tap snugly and insert the screw as metal layer reinforcement. Using woodworking joinery techniques and captive nuts with machine screws is a very effective way to join 3D printed parts together with minimal fasteners. Hope this helps. Keep up the great content.
Check out this video by Lost in Tech showing that you can go with an even larger width based on the chosen nozzle: ua-cam.com/video/0DAP5Zm1jvk/v-deo.html
@@nidavis Very interesting. Thank you for sharing. The main point I was driving home was using a thin 0.35 - 0.4mm outer perimeter to reduce layer lines and increase accuracy and surface quality, while using fat lines to fill the wall thickness fast. Skipping 3 infill layers then filling them with a fat layer. You need to tailor the concept to fit your part.
I wonder if he could just ask e&f if he has any insights. Since he often talks about being close to finishing his PHD working in a high laser power and optics lab.
use print in place flexures! probably the best solution for a rigid, very accurate and possibly very fine adjust mount (as you can use leverage to reduce travel)
There is a 3D printed microscope stage that uses flextures floating around the net somewhere. Afik it drifted 10-15um a day which was slow enough to track cells in a culture dish over several days, but would still mess up a lot of optical setups.
Kinematic mounts don't just touch in 3 places like you have. To fully kinematically define the location of the component, you need 6 points of contact. The thorlabs one uses a ball-in-cone (which provides 3 points of contact), a ball in V (to provide 2 more points of contact), and a ball on plane (1 point of contact). Together these 6 points fully constrain the 6 DOF so you don't get sliding around of the component.
Very nice video on these optomechanical mounts youve got there. I just want to leave one comment about your kinematic mounts here - the 3D printed ones you printed have the 3 set screws in a triangle shape around your object. If you think about it in general it is perfectly fine, but if you for example just want to shift your mirror in horizontal direction you need to use two screws - and this is quite the hassle. Try arranging them like the ThroLabs Mount in an L shape to cover horizontal and vertical movement with just one screw. Same goes for your LED mount. These additional moves to achieve a certain directionality in your optical setup get pretty frustrating over time. Also - when mounting an LED/Laser/Light source it's usually easier to mount it fixxed and use a mirror or the following setup to guide it instead of adjusting the LEDs position and direction.
Hi, I love your videos. I think I know why you had problems with your kinematic mounts. They aren't kinematic mounts as understood in mechanical engineering. They don't remove all degrees of freedom. When three balls contact three planes ( tips of the screws ) you are left with 2 degrees of freedom. If the balls would touch three V grooves you would have very repeatable and fully constrained connection. You could easily achieve this with three pairs of shoulder bolts in the plane of the mount that would create v grooves. I admire your work to popularize science and engineering, good luck.
@@gedr7664 It's probably easier to just think about them as three pairs of parallel pins that provide a track for the balls to contact. If you search in the Thorlabs catalog for "Kinematic Bases" you will find examples of this.
@@gedr7664 I forgot that youtube shadowbans comments with links :/ . In the video here on youtube titled " 3D Printed Kinematic Coupling Proof of Concept " by Joshua Vasquez you can see the described idea.
I came here to say this as well! +1 for Josh. He's a great guy to talk to as well. He has a video showing the concept, I think it would have helped a lot here ua-cam.com/video/a9irK9rOUHY/v-deo.html
I feel like flextures should be the name of the game for fine adjustments in 3d printed mounts rather than sliding contacts. they're just much less wobbly. i also wonder if it's possible to design a robust magnetic mount in a 3d print that's repeatable at least to a degree.
Would flexures not suffer from creep issues similar to the cantilever mounts? Could this be mitigated by switching to a different filament like nylon? (or one of the nylon+filler composites that print better but require abrasive-resistant printer parts)
When i need rigidity i usually put round channels (or holes if you like) in the design that then i fill with a metal rod. Normally i use 3,4 or 5 mm rods but depending on the part you can use any size and it is very easy to find the right setting in the printing so the fitting is right to hammer in the rods. You can also print the holes slightly oversized and then glue in the rods but i found much better to hammer in them. The rigidity is astounding.
3:58 "functionally it's the same as this kinematic mount" Yeah, no. It's missing preload on the screws so it can wiggle around, and the printed bracket could be (should be) reinforced with a cross member. I also wonder how much the prints will bend and flex around when the temperature changes. Probably they can be tempered by baking / cooking in hot water?
For the kinematic mounts, you could buy ball-point set screws and then print a conical housing for the magnets. The conical housing would go above the magnet and allow the ball on the set screw to sit in the cone. If the housing is thin enough, the set screw would still be pulled into contact, though I'm not sure how well it would work compared to direct magnet-on-screw action. This would help with the translation issues you were having and would not be particularly expensive to try I think.
Been hoping for a video like this for years. Ever since I started collecting and assembling a real photonics lab… i’ve been on the hunt and interested in this type of stuff. Since I’m already around the stuff in my day-to-day business I decided to put the word out and look extra carefully when visiting labs in “academia” and in research labs/small manufacturing facilities… where I could buy wholesale optical tables and breadboards and stuff like. I’ve been really lucky to get a few super nice optics tables and things like that for pennies on the dollar… You know how it goes with these huge manufacturers and labs when they upgrade or whatever… they kind of just auction it off or toss it. I’ve seen folks fill up entire parking lot full of dumpsters with high-end test equipment that were still in calibration. Sickening to say the least. But a wonderful opportunity for folks in need of stuff like that., Although I’ve been very lucky and have a arsenal of mounts for my tables and breadboards, I’m always needing something weird… I’ve designed and printed a few myself but that’s not some thing I excel with and I usually struggle.. So I’ve been hoping someone would seriously tackle 3-D printed photonics mounts and interfacing accessories. “Thought emporium“ tinkered around with the idea in a video years ago… and there have been a few random posts throughout the Internet and laser forms and stuff. But nothing to this level. Your attention to detail and design goes above and beyond… As usual. Super stoked you’ve taken it this far and I hope you show us any updates or things that you add to it in the future.
Awesome DIY project! I work as an optical engineer with these kind of breadboard optics . I really appreciate the effort! but if Im honest I see some cool ideas but also a lot of problems. Some tips and comments: - Press fitting optics is not a good idea. Big chance to ruïne the coating on the optical element. I would prefer a nylon set screw.And especially dichroic mirrors should be mounted stress free. - Objectives are very sensitive to alignment and it is worth the effort to decouple translation and tip/tilt alignment. - I think the flexure joins are awesome. - When designing/building it is super helpfull the place alignment aids and pinholes. Perhaps that can speed up the alignment. Keep up the great work!
optics bits are crazy expensive. the companies I worked for shelled out a lot of money for my projects. 3D printing is interesting for some parts but as you found out, not all.
I got a few prototype Germanium lenses a long time ago. My 3D printer was still working back then. And I designed and printed really really bad lens holders which were really bad. They almost look like the extrusion stands you showed - but just stand on the table, with lens elements just hold on by gravity. There are better solutions if you look through ebay a bit, or even find modern kits. Which I still don't have. It was enough for me to figure out what combination of lens elements I needed and how they were spaced and rotated. I managed to print a lens housing that was 4 parts with printed threads. They even allowed "focusing". And it was a lot of fun. Really low resolution thermal camera (80x60) is Soo much better with a narrow lens than the wide lens it usually comes with. I sadly dropped and shattered one of the elements, so I never tried to use that lens again with my 160x120 camera or even the 384x288 and 640x480 cameras I have now as well as their giant lenses.
Two things I've had some success with too: 1. Using micro-steppers meant for mirror adjustments in camera's. They cost only a dollar each on aliexpress and can be useful for hard to reach screws, fine adjustments, or anything you'd rather have computerized. At the cost of having to add more electronics, of course. 2. mounting the too-be-adjusted plane between two dissimilar springs on each screw (so you have a screw going trough [top base-spring k1-adjustment plane-spring k2-bottom base]). It works as a small shock-mount (sometimes good sometimes bad) and linear step down gear as the springs won't compress the same amount if you tighten the screw.
For the red diode, unthreaded sleeves could be put on top of more deeply embedded magnets to allow for the bolts to be accurately located while still allowed to travel vertically. This would limit the maximum tilt in any direction, but seems like a good compromise for fine adjustments.
Also take the bolts, put a nut on them till the bottom of the thread is just above the nut, and use a file to make the bottom of the bolt flat, so that the curled end of the rolled thread does not make it wobble. You can use each nut twice, once per face, to do the filing, and then use them as the inserts in the print, heat sealing them into the plastic, or throw them away as M3 nuts are very cheap in steel.
Or perhaps, if you have an available lathe, cut into the tips of the bolts to make the end of the bolt concave or conical (the negative of a cone that is) so that the bolt centers itself on the ball magnet
You may also be able to use a larger non-magnetic ball pressed into the base, and a cup (even just a hole) in the center of the tripod to improve the tripod alignment repeatability. The magnets will still pull down, and the ball/cup will provide the centering.
I'd like to apologize for the horribly janky and imprecise mechanisms I build on this channel, compared the the beautiful air bearing work over on yours! 😅 Found your channel the other day and have been working through the back-catalog! Really great stuff!
You might be interested to google kinematic coupling. I do not know much about them but they are used to get precise tool changing in 3d-printers. Could be useful for making more precise attachments of your different components.
For spacing elements a certain distance you might look into using gage blocks. They are relatively inexpensive and you can stack them together to get a specific size. It might work better than trying to manually shift around and you can hold your spacing steady while tightening down the screws.
Very cool, I've really enjoy following this project. I'm a machinist/designer and have spent some time designing and making optical components both at home (3d printed) and at work (machined). This has been almost surreal to follow you on some very similar paths. Some random comments. I've found springs in 3d printed kinematic mounts give really good feedback, perhaps a bit more positive than magnets. The awesome technical papers at Bal-Tec cover some topics of magnets in kinematic structures. Off the shelf contact points (flat on ball or ball in socket) might make alignment a bit more predictable on some of these kinematic mounts. Also - ThorLabs has CAD files for all their products. It makes reverse engineering some designs really quick. It's also fun to get into some of their design intent. Thanks for sharing!
You should use balls and rollers from bearings to joint parts. The three balls create an extremely precise and repeatable connection of moving parts. The alternative is flexure.In your video there are examples of using flexure.
For the unstable mountings, instead of screw-insert in moving part-magnet, why not to use screw-moving part-spring-insert in base? (Same system as the manual leveling of an old 3d printer)
if you have a lathe handy, then drilling a conical recess (even with the point of a regular twist drill) into the end of the set screws might help them mate better against the surface of the spherical magnets and could constrain them a bit better in translation. Ball-in-cone is one of the classical kinematic design elements AFAIK
Thanks for sharing your successes and failures so we can all learn, I didn't know about UC2, that's interesting. For the 3-screw tilt adjustment, you could try using cup-point grub screws, which have a semi-spherical hole in the end that would centre themselves onto the spherical magnets. The screw and magnet size need to be carefully selected (magnet diameter slightly larger than screw hole sphere diameter). You may need some lateral float in one of the magnets or screws to allow for tolerances. Instead of using magnets, you can use a ball bearing sandwiched between two cone shaped holes, then use 3 screws to tilt the cup and ball joint. If you don't want a ball bearing, 3 sets of springs and screws can be used. Springs push 2 plates apart at the corners and the screws tighten to pull the plates closer together. There are also flexure arrangements, where you bend the material at a designed weak point, to tilt the lens. A good option is a virtual ball joint, where a large diameter 3D printed (partial) ball, whose centre would be in the middle of the optical axis, is mounted on one side of the lens/laser, and is pulled tight using a screw, nut and spring into a ball cup. The ball cup has a large hole to allow the screw to move, and there would be a larger cup with small hole for the screw head to sit on top of, outside of that. This gives the ideal centre of rotation. Getting the ball and cup smooth is tricky but can be done with a fine print layer height and some sanding and vapour polishing. As before, the ball wants to be slightly larger than the cup so that it can't move laterally. I noticed your laser was on an arm, away from the adjustment platform. This amplifies the adjustment and moves the laser in an arc rather than tilting on axis like a gimbal. Keeping the adjustment plane as close to the lens or laser optical axis as possible gives much better control.
the flexure mounts you showed at the end gave me and idea. instead of three point magnetized mounts, how about two 90 deg offset stacked hinges with one adjustment screw per hinge. that should eliminate side to side movement
i guess those only works if yaw and pitch are all you need. this wouldn't allow for depth changes like the three point system... i guess you could stack three hinges at 120 deg offset?
Have you tried drilling sockets into the thread end of the bolts to cup your balls? Maybe as a proof of concept flip a few cap head bolts upside down so the ball partly fits into the cap. It seems like they would snap into the low energy state which would be centered over the balls.
Nice.. For the laser diode and microscope objective holders i can suggest adding a spring .. the part will move laterally on adjustment but will go back to its place
I wonder how printed element compare to eg aluminium one, even made from screwed parts in terms of stiffness and reaction to temperature changes and vibrations after longer period of time. Like for example those tilted 45* mirrors. 3d printing is amazing but I think with delicate devices like optic stuff even cheap aluminium parts crudely made without even CNC cutting and that are just screws together would be better... If you can't keep your project on one flat surface where you could just stack them like those cubes.
The 3-screw adjustment for optics and lasers could be improved to resolve away the shift by pressing the magnets into the back side of the 3D prints and the screws must reach down through some "guide" holes, maybe with some rubber O-rings pressed into the screw-side to hug around the screws for stabilizing and retaining position better. It wouldn't be completely ideal but it would be a cheap simple way to greatly improve on the concept and lock things down more to where the screws are pretty much the only adjustment that can be effected upon the optical devices.
5:40 with the stands could you use screws shaped to a cone at the bottom, so the sit well located, before you use the screw action to adjust the precise, repeatable location? - like 4gwVucFaPw but smaller
I use a Ball ans sockets for my tripod mounts. You could just find a large enough ball bearing for the center pivot.
2 роки тому
About the screw on top of the magnet arrangement, one idea for you: Drill a small hole on top of the magnet and make the screw end sharp, just like a nail.
I was thinking about spot drilling a cone into the screws. The potential problem I see with a system like this is that it over constrains the part. Maybe that would lead to unexpected shifting when adjusting it.
Thanks for sharing your experience. I would argue that a lot of the things where you said 3D printing is not a good idea - it is actually the specific design you show for the mount, not the fact that they are 3D printed. Using those magnets, for example, and the simplistic right-angled joints on your brackets. With better design considerations you can do probably all this thing with 3D printed mounts and ordinary nuts and bolts. Nice work though.
I’ve been toying with the idea of 3D printing flexture based optics equipment based loosely on OpenFlexture. Maybe something like that could get you the resolution your looking for!
Yes, you can! I recently put together a low-cost stepper and inspection microscope doing something just like this. I'm going to try and start documenting it soon (hopefully!)
What has pushed me away from 3D printing opto-mechanical components is the poor CTE performance of thermoplastics used in FDM printers. Still super impressed with what you've been able to demonstrate on your channel with 3D printed optical parts.
@@AndrewGillard Plastics in general have 2X to 10X higher CTE than metals. Also, the poor thermal conductivity of plastics leads to larger temperature gradients across the material, which causes warping. In optical applications, warping is typically worse than just linear expansion and contraction because it results in angular displacement of critical datums.
VERY COOL! ...And wow! @1:58 where is that optical table photo from? The most dense setup I have ever seen. I am an optical science guy and I appreciate your mounts and parts. Nice work. But you also showed some of those designs are not really 'kinematic.' That said, I 3d printed parts to build a LiDAR in 2020 and it worked a lot better than expected. You might have a product at least for education with the magnetic system.
This is a wealth of great information, amazing work! If it's helpful, I suggest using the right "drop-in" t-nuts when possible instead of those finnicky "hammer head" t-nuts. They're far less likely to move around. Also it looks like you may be using PLA for much of this? The 'printability' is great for prototyping but I've learned going with ABS for parts like this avoids so many headaches later. The inevitable creep with PLA, PETG (and even Nylons) can get so frustrating. I rarely even prototype with PLA now since decent ABS can be very cheap. For finalized revisions, my go-to is usually CF-ASA, or cheaper CF-ABS, for further improved stiffness and shrinkage resistance. You may have a different school of thought on this but I just wanted to throw it out there.
I'd love to work with ABS, but my printer really struggles (not enclosed, poorly tuned). At some point I'll stop being lazy and build an enclosure though, would be great to work with ASA or ABS!
@@BreakingTaps I had a cheap ikea table with foam board walls sitting over a slightly modded Ender 3 running Klipper and it basically printed nothing but ABS parts for the past year. PETG can be more frustrating to print than ABS IMO. Considering what I've seen you do on this channel I think you could figure it out before finishing breakfast ;)
I'm no expert, but I would assume that annealing the 3d printed parts would make them more rigid and increase the effective TG of the part futher to make it less likely to suffer temperature dependent error when in use. Its a whole 'nother project, but it might be worth checking out. EDIT: Holding brackets for kinematic mounts, @ 10:28 are $18 each on Aliexpress.
Annealed printed parts are much more rigid, however their dimensions change. A good alternative is probably to remelt it in powdered salt: You get the strength of annealed parts with the original dimensional accuracy. They have to be printed with 100% infill though.
I have no clue about anything here, but from the looks of it the UC2 system suffers from the same problem as your laser diode mount? Or am I missing a concept that eliminates that?
The older system does suffer from the "sliding around" phenomenon a bit (it's one of the reasons I didn't end up using it for the confocal). They have a newer version that snaps into place which I think eliminates that problem but I haven't played with it yet. I found that recessing the magnets more than I have in that example helps too. That said, the design of most UC2 components doesn't require adjusting kinematics "inside" the cube, so any misalignment can be tweaked by just fidgeting with the cube. That ends up being pretty easy and you aren't fighting adjustment screws _and_ the magnets like in my designs.
For the laser issue, wouldn’t it be a good idea to print a ‘3 sided pyramid’ into the laser base and the same feature into the base with the three magnets…. Side the 3 sided pyramid hole to accept a cheap ball bearing… The steel ball is sandwiched between and still allows pitch and tilt while constraining is lateral movement.
Still hoping you are going to make a video about making telescope mirrors.....I am pretty much stuck on my project and hoping you are going to inspire me or point me in a new direction.
Maybe some day! Looks like a lot of work to get setup (grinding machine and measuring setup, etc) so I've always stayed away, but it would be great to have an optical griding setup to use on various projects 🤔
Sort-of on the subject of optics.... absolutely love the lighting in the talk segments of this video. And secondly.... where did you get that fantastic hoodie from?.. I want one.... I want one NOW.
YES! I can't believe I caught this 6 minutes after release. I know I have the little bell with all notifications but somehow I still miss it. I wonder if the ball + screw adjustment system could be fixed with a registration point or two
I was wondering if some lubrication of the magnets against the bolts would improve things at all - i.e. if it was (at least partly) a case of the magnets not being strong enough to overcome the friction. Probably still wouldn't be perfect - registration points sounds like a better plan - it was just a thought I had while watching :) I'm much more of a software engineer than mechanical, and I've not been able to sink as much time as I'd like into designing parts for my 3D printer, so I'm still pretty new at this sort of thing!
Sorry to bother you with another comment. Regarding the James Webb telescope “golden mirror” in your other video… Was that just a piece of cardboard spray painted gold? A piece of gold anodized aluminum? Im Just curious as I’ve been toying with the idea of re-creating one while I have access to the equipment. Just for fun and as a display piece, And I would only do it in a batch of one or maybe a few at a time because I think they are gorgeous and the way they made them is absolutely incredible. The technology in those mirrors is insane … Not to mention the interface and adjustment mechanisms. Just breathtakingly awesome
It was CGI actually 😇 I printed a replica to hold in my hand, and then replaced it with a 3d model while editing. I agree it'd be super cool to have a replica as a desk toy!
3D printed bearings... Taper bearings? maybe externally pressurised hydrostatic bearings? 3D printer could not directly print tapers accurately... But assemble and Keep increasing preload when driven to rotate by say a drill, and high spots should bottom out and might be able to achieve a tollerance similar to the ground glass couplings of chem lab glass wear.
A silly, low effort suggestion: just turn the m3 screws around backwards and the spherical magnets will sit in the hex and maybe give some more precision. and maybe cut a slot in the bottom of the screw to turn with a tiny screwdriver? dunno. sweet video! good luck!
yunno... that would probably work well! Some other folks have suggested cutting/grinding a sphere on the ends, but just flipping them around would probably work just as well and be a lot easier! Will give it a shot :)
Did anyone else wince a little inside every time he moved a part around? I had to keep reminding myself that he's already got to realign all the fiddly bits again anyway, and it's okay. 😅
Your videos are great but the camera shots have a somewhat lower framerate. Maybe something to do with editing, but it really stands out. Other than that I really enjoy them :)
A very interesting topic. I've been thinking about a designer for optomechanics for a long time. The main parts were supposed to be printed. I will be very pleased if you look at them and try them in your work? I will also upload them to GrabCAD. I have blocked other convenient sites. The time is strange.
Honestly, beside the postholders, everything else would have been actually cheaper if you bought the "nice stuff" directly from thorlabs, once you factor in all the time, efforts and materials spent to 3D print stuff that doesn't actually work. Even the postholders, they are actually worth only if you already have a 3D printer, the files correctly sliced and you just have to push the "start" button to get the final pieces: if you have to spend more than an hour and you have a paying job, you would have been better just buying them... The problem with 3D printing is that it "feels" free, since you don't have to pay someone else the moment you get the finished piece, but it is actually way, WAY more expensive than mass produced items, especially if you value your time in any form...
I agree, but also sorta disagree :) For a business use-case, absolutely. No excuse not to just go buy the real thing and get on with whatever is making money. For hobby, I think it's a little more difficult to say. Each base/post/lens or mirror assembly would be $60-100 depending on style (fixed element or kinematic). I have maybe 10 different elements, although some of that could change to help decrease cost, but still looking at $500-1000 without blinking. Throw in several redesigns where a few element dimensions change and you're looking at a pretty expensive hobby project. Now of course, once you purchase the components you have them for future projects so it's not a lost cost. And if you make a good salary the "time cost" is definitely something to consider. But for prototyping I think 3D printing is hard to beat. I used less than a roll of filament and probably no more than a few hours designing everything, so salary + material time it still came out ahead. And lets you make customizations very quickly, instead of waiting 2-3d for the parts to arrive by mail to work on the next iteration (and hope you don't need to re-order again and restart the waiting period).
It's a trick to get everyone to keep watching this video over and over again! :D (And it's an interesting video that I've bookmarked for future reference, so it's worth watching again in a way :p)
So awesome and super helpful, just as always!!! Actually, regarding the magnet bolt problem, couldn't you just add a cone so that the bolts are somewhat restricted? Like this: _ || bolt \ / |●| magnet | | I know, this wouldn't add uber precision, but it could potentially help :)
For tip and tilt you only need two adjustment screws. You can fix the 3rd one with a spacer. You will need all 3 in case you want to refocus. Some simple (wooden) clothes pegs can hold lenses and other stuff without 3d printing or clamping screws. Allways preload adjusting screws for fine adjustment.
Hi I am designing a DIY Polarimeter with 0.003° precision right now this is really helpful
Happy to help! Goodluck with your project, sounds neat!
Would love to see some videos or something about it
I like the ball magnet mounts. A slight variation on the design I’m using on one of my setups is to stack a washer on a disk magnet of similar diameter and press fit that into a print washer side out. Then on the bolt side each blot passes through the heat set nut then gets capped with an acorn nut. The acorn dome to washer ring interface is self centering, 3 of those on a fixture and translation is locked down while still allowing the same fine adjustment.
Set-screws tipped with a ball-bearing and a bag of dowel pins would let you make quite solid 3D printed kinematic mounts. You can also buy the same screws and threaded bushings used in the commercial mounts. Makes it easy to design high quality custom mounts. Like a X-Y translation kinematic mount.
I can attest to the accuracy of 3d printed kinematic mounts designed like this. They work great and are very repeatable
do you have a link to sth like that? i cant really imagine how you mean it
I don't do anything with optics but still found this insightful and useful. Thanks for taking the time to share your experiences with us!
Thank you for sharing. I do a lot of 3D printing of mechanical parts. I use a PRUSA MK3S w/ MMU2 w/ 0.4 mm nozzle. Some tips that may or may not work depending on the slicer software you use. In advanced settings the width of the extrusion can be controlled to some degree. Make external perimeters 0.35 - 0.4 mm with a slower feed speed. Then internal perimeters 0.6 - 0.65 mm with a high feed speed and layer height 0.15 mm. infill at 0.7 mm. In the infill settings combine infill every 4 layers. This will give smoother exterior to your parts and better tolerance while maintaining fast print speeds. Also I use cubic infill with 30 - 40% density. Your magnetic ball idea is a good one, however to greatly decrease unwanted movement make the receiving side concave to fit the magnet. Magnetic PLA is quite effective for magnetic attractive surfaces and with a multi material setup you can use it sparingly where needed. Putting a 3D printed fence around the movable part at 0.15 - 0.2 mm backlash can also further constrain against unwanted movement. Use printed gussets and struts of various types everywhere with access holes for tools and fasteners to increase rigidity. It is always worth the extra print time to eliminate as many potential problems as possible. It is also much faster to print your own custom supports in the part design for overhangs as you can speed print time and reduce material waste with good design and no generated supports rather than automatically generated supports. This also gives you better control of interface gaps and tailor them to the material you are printing with. For very low density infills you can force the slicer to add perimeters where you want solid infill by making very thin voids at 0.05 mm. This takes a bit of practice as it depends on the printer, slicer and filament type but you can make voids that are so small that they are practically solid and fused but mathematically the slicer sees a perimeter. This will force the low density infill around these voids with a perimeter shell. For locations with a high chance of breakage like a low diameter post on a plate, make a hole for a flat head machine screw to self tap snugly and insert the screw as metal layer reinforcement. Using woodworking joinery techniques and captive nuts with machine screws is a very effective way to join 3D printed parts together with minimal fasteners. Hope this helps. Keep up the great content.
Check out this video by Lost in Tech showing that you can go with an even larger width based on the chosen nozzle: ua-cam.com/video/0DAP5Zm1jvk/v-deo.html
@@nidavis Very interesting. Thank you for sharing. The main point I was driving home was using a thin 0.35 - 0.4mm outer perimeter to reduce layer lines and increase accuracy and surface quality, while using fat lines to fill the wall thickness fast. Skipping 3 infill layers then filling them with a fat layer. You need to tailor the concept to fit your part.
Lol, love your sweatshirt!! I have the tie-dye t-shirt version! :D
I wonder if he could just ask e&f if he has any insights. Since he often talks about being close to finishing his PHD working in a high laser power and optics lab.
use print in place flexures! probably the best solution for a rigid, very accurate and possibly very fine adjust mount (as you can use leverage to reduce travel)
There is a 3D printed microscope stage that uses flextures floating around the net somewhere. Afik it drifted 10-15um a day which was slow enough to track cells in a culture dish over several days, but would still mess up a lot of optical setups.
creep is an issue for plastic flexures.
Could you get the creep under control with relatively hard resin (SLA) prints?
@@floriandaler5327 "get the creep under control" brought other things to mind.
@@floriandaler5327 it will certainly help but won't completely eliminate creep. Crosslinked plastics like SLA resin or epoxy are also better vs creep.
Kinematic mounts don't just touch in 3 places like you have. To fully kinematically define the location of the component, you need 6 points of contact. The thorlabs one uses a ball-in-cone (which provides 3 points of contact), a ball in V (to provide 2 more points of contact), and a ball on plane (1 point of contact). Together these 6 points fully constrain the 6 DOF so you don't get sliding around of the component.
Very nice video on these optomechanical mounts youve got there. I just want to leave one comment about your kinematic mounts here - the 3D printed ones you printed have the 3 set screws in a triangle shape around your object. If you think about it in general it is perfectly fine, but if you for example just want to shift your mirror in horizontal direction you need to use two screws - and this is quite the hassle. Try arranging them like the ThroLabs Mount in an L shape to cover horizontal and vertical movement with just one screw. Same goes for your LED mount. These additional moves to achieve a certain directionality in your optical setup get pretty frustrating over time.
Also - when mounting an LED/Laser/Light source it's usually easier to mount it fixxed and use a mirror or the following setup to guide it instead of adjusting the LEDs position and direction.
Hi, I love your videos. I think I know why you had problems with your kinematic mounts. They aren't kinematic mounts as understood in mechanical engineering. They don't remove all degrees of freedom. When three balls contact three planes ( tips of the screws ) you are left with 2 degrees of freedom. If the balls would touch three V grooves you would have very repeatable and fully constrained connection. You could easily achieve this with three pairs of shoulder bolts in the plane of the mount that would create v grooves. I admire your work to popularize science and engineering, good luck.
i am struggling to visualise creating a v groove with a shoulder boult -- could you please provide a sketch or drawing of this?
@@gedr7664 Here is this idea shown by Joshua ua-cam.com/video/cfSHss5j5KU/v-deo.html&ab_channel=JoshuaVasquez
@@gedr7664 It's probably easier to just think about them as three pairs of parallel pins that provide a track for the balls to contact. If you search in the Thorlabs catalog for "Kinematic Bases" you will find examples of this.
@@gedr7664 I forgot that youtube shadowbans comments with links :/ . In the video here on youtube titled " 3D Printed Kinematic Coupling Proof of Concept " by Joshua Vasquez you can see the described idea.
I came here to say this as well! +1 for Josh. He's a great guy to talk to as well. He has a video showing the concept, I think it would have helped a lot here ua-cam.com/video/a9irK9rOUHY/v-deo.html
I feel like flextures should be the name of the game for fine adjustments in 3d printed mounts rather than sliding contacts. they're just much less wobbly.
i also wonder if it's possible to design a robust magnetic mount in a 3d print that's repeatable at least to a degree.
Would flexures not suffer from creep issues similar to the cantilever mounts? Could this be mitigated by switching to a different filament like nylon? (or one of the nylon+filler composites that print better but require abrasive-resistant printer parts)
When i need rigidity i usually put round channels (or holes if you like) in the design that then i fill with a metal rod. Normally i use 3,4 or 5 mm rods but depending on the part you can use any size and it is very easy to find the right setting in the printing so the fitting is right to hammer in the rods. You can also print the holes slightly oversized and then glue in the rods but i found much better to hammer in them. The rigidity is astounding.
3:58 "functionally it's the same as this kinematic mount"
Yeah, no. It's missing preload on the screws so it can wiggle around, and the printed bracket could be (should be) reinforced with a cross member.
I also wonder how much the prints will bend and flex around when the temperature changes. Probably they can be tempered by baking / cooking in hot water?
Awesome!!! Being able to print tubes, lens barrels and mirror holders were a main reason for me to get into 3D printing!
For the kinematic mounts, you could buy ball-point set screws and then print a conical housing for the magnets. The conical housing would go above the magnet and allow the ball on the set screw to sit in the cone. If the housing is thin enough, the set screw would still be pulled into contact, though I'm not sure how well it would work compared to direct magnet-on-screw action. This would help with the translation issues you were having and would not be particularly expensive to try I think.
Been hoping for a video like this for years.
Ever since I started collecting and assembling a real photonics lab… i’ve been on the hunt and interested in this type of stuff. Since I’m already around the stuff in my day-to-day business I decided to put the word out and look extra carefully when visiting labs in “academia” and in research labs/small manufacturing facilities… where I could buy wholesale optical tables and breadboards and stuff like. I’ve been really lucky to get a few super nice optics tables and things like that for pennies on the dollar… You know how it goes with these huge manufacturers and labs when they upgrade or whatever… they kind of just auction it off or toss it. I’ve seen folks fill up entire parking lot full of dumpsters with high-end test equipment that were still in calibration. Sickening to say the least. But a wonderful opportunity for folks in need of stuff like that.,
Although I’ve been very lucky and have a arsenal of mounts for my tables and breadboards, I’m always needing something weird… I’ve designed and printed a few myself but that’s not some thing I excel with and I usually struggle..
So I’ve been hoping someone would seriously tackle 3-D printed photonics mounts and interfacing accessories.
“Thought emporium“ tinkered around with the idea in a video years ago… and there have been a few random posts throughout the Internet and laser forms and stuff. But nothing to this level. Your attention to detail and design goes above and beyond… As usual.
Super stoked you’ve taken it this far and I hope you show us any updates or things that you add to it in the future.
Great video, love the sweatshirt ❤️
Awesome DIY project!
I work as an optical engineer with these kind of breadboard optics . I really appreciate the effort! but if Im honest I see some cool ideas but also a lot of problems. Some tips and comments:
- Press fitting optics is not a good idea. Big chance to ruïne the coating on the optical element. I would prefer a nylon set screw.And especially dichroic mirrors should be mounted stress free.
- Objectives are very sensitive to alignment and it is worth the effort to decouple translation and tip/tilt alignment.
- I think the flexure joins are awesome.
- When designing/building it is super helpfull the place alignment aids and pinholes. Perhaps that can speed up the alignment.
Keep up the great work!
Cheers for the tips, appreciate it! Will keep these in mind for future projects
Good points !
optics bits are crazy expensive. the companies I worked for shelled out a lot of money for my projects. 3D printing is interesting for some parts but as you found out, not all.
I got a few prototype Germanium lenses a long time ago. My 3D printer was still working back then. And I designed and printed really really bad lens holders which were really bad. They almost look like the extrusion stands you showed - but just stand on the table, with lens elements just hold on by gravity. There are better solutions if you look through ebay a bit, or even find modern kits. Which I still don't have.
It was enough for me to figure out what combination of lens elements I needed and how they were spaced and rotated. I managed to print a lens housing that was 4 parts with printed threads. They even allowed "focusing". And it was a lot of fun. Really low resolution thermal camera (80x60) is Soo much better with a narrow lens than the wide lens it usually comes with.
I sadly dropped and shattered one of the elements, so I never tried to use that lens again with my 160x120 camera or even the 384x288 and 640x480 cameras I have now as well as their giant lenses.
Love the Carbon Tet shirt so much.
Two things I've had some success with too:
1. Using micro-steppers meant for mirror adjustments in camera's. They cost only a dollar each on aliexpress and can be useful for hard to reach screws, fine adjustments, or anything you'd rather have computerized. At the cost of having to add more electronics, of course.
2. mounting the too-be-adjusted plane between two dissimilar springs on each screw (so you have a screw going trough [top base-spring k1-adjustment plane-spring k2-bottom base]). It works as a small shock-mount (sometimes good sometimes bad) and linear step down gear as the springs won't compress the same amount if you tighten the screw.
Optical rail tends to use dovetail faces for mounting instead of slot-and-T-nut, which is more resistant to slip over time.
For the red diode, unthreaded sleeves could be put on top of more deeply embedded magnets to allow for the bolts to be accurately located while still allowed to travel vertically. This would limit the maximum tilt in any direction, but seems like a good compromise for fine adjustments.
Also take the bolts, put a nut on them till the bottom of the thread is just above the nut, and use a file to make the bottom of the bolt flat, so that the curled end of the rolled thread does not make it wobble. You can use each nut twice, once per face, to do the filing, and then use them as the inserts in the print, heat sealing them into the plastic, or throw them away as M3 nuts are very cheap in steel.
Or perhaps, if you have an available lathe, cut into the tips of the bolts to make the end of the bolt concave or conical (the negative of a cone that is) so that the bolt centers itself on the ball magnet
You may also be able to use a larger non-magnetic ball pressed into the base, and a cup (even just a hole) in the center of the tripod to improve the tripod alignment repeatability. The magnets will still pull down, and the ball/cup will provide the centering.
Lego is great for optical breadboards, the accuracy and tolerance is incredible.
I want to believe this, but what Lego parts hold the lenses and mirrors?
You gave me a ton of great ideas for a Raman Spectrometer project I've been at for over a year. Thanks!
1:56 holy CRAP!
New breaking taps yeet
I'd like to apologize for the horribly janky and imprecise mechanisms I build on this channel, compared the the beautiful air bearing work over on yours! 😅 Found your channel the other day and have been working through the back-catalog! Really great stuff!
1:55 How many Optomechanical Components would you like to mount?
*Yes*
Bedankt
Thank you! ♥
You might be interested to google kinematic coupling. I do not know much about them but they are used to get precise tool changing in 3d-printers. Could be useful for making more precise attachments of your different components.
Excellent! Exactly the kind of information I needed.
For spacing elements a certain distance you might look into using gage blocks. They are relatively inexpensive and you can stack them together to get a specific size. It might work better than trying to manually shift around and you can hold your spacing steady while tightening down the screws.
Very cool, I've really enjoy following this project. I'm a machinist/designer and have spent some time designing and making optical components both at home (3d printed) and at work (machined). This has been almost surreal to follow you on some very similar paths.
Some random comments. I've found springs in 3d printed kinematic mounts give really good feedback, perhaps a bit more positive than magnets. The awesome technical papers at Bal-Tec cover some topics of magnets in kinematic structures. Off the shelf contact points (flat on ball or ball in socket) might make alignment a bit more predictable on some of these kinematic mounts. Also - ThorLabs has CAD files for all their products. It makes reverse engineering some designs really quick. It's also fun to get into some of their design intent.
Thanks for sharing!
Oh neat, I didn't realize Thor shared their CAD models. Cheers for the tip, off to download a bunch of models!
You should use balls and rollers from bearings to joint parts. The three balls create an extremely precise and repeatable connection of moving parts. The alternative is flexure.In your video there are examples of using flexure.
For the unstable mountings, instead of screw-insert in moving part-magnet, why not to use screw-moving part-spring-insert in base? (Same system as the manual leveling of an old 3d printer)
Yeah that would work great! That's basically how the real ones do it too, a spring preload against the adjustment screw
if you have a lathe handy, then drilling a conical recess (even with the point of a regular twist drill) into the end of the set screws might help them mate better against the surface of the spherical magnets and could constrain them a bit better in translation. Ball-in-cone is one of the classical kinematic design elements AFAIK
Carbon tet gang represent
yellow trash
Yee yee carbon tet!
Ps: fuck thorlabs
Thanks for sharing your successes and failures so we can all learn, I didn't know about UC2, that's interesting.
For the 3-screw tilt adjustment, you could try using cup-point grub screws, which have a semi-spherical hole in the end that would centre themselves onto the spherical magnets. The screw and magnet size need to be carefully selected (magnet diameter slightly larger than screw hole sphere diameter). You may need some lateral float in one of the magnets or screws to allow for tolerances.
Instead of using magnets, you can use a ball bearing sandwiched between two cone shaped holes, then use 3 screws to tilt the cup and ball joint. If you don't want a ball bearing, 3 sets of springs and screws can be used. Springs push 2 plates apart at the corners and the screws tighten to pull the plates closer together. There are also flexure arrangements, where you bend the material at a designed weak point, to tilt the lens.
A good option is a virtual ball joint, where a large diameter 3D printed (partial) ball, whose centre would be in the middle of the optical axis, is mounted on one side of the lens/laser, and is pulled tight using a screw, nut and spring into a ball cup. The ball cup has a large hole to allow the screw to move, and there would be a larger cup with small hole for the screw head to sit on top of, outside of that. This gives the ideal centre of rotation. Getting the ball and cup smooth is tricky but can be done with a fine print layer height and some sanding and vapour polishing. As before, the ball wants to be slightly larger than the cup so that it can't move laterally.
I noticed your laser was on an arm, away from the adjustment platform. This amplifies the adjustment and moves the laser in an arc rather than tilting on axis like a gimbal. Keeping the adjustment plane as close to the lens or laser optical axis as possible gives much better control.
the flexure mounts you showed at the end gave me and idea. instead of three point magnetized mounts, how about two 90 deg offset stacked hinges with one adjustment screw per hinge. that should eliminate side to side movement
i guess those only works if yaw and pitch are all you need. this wouldn't allow for depth changes like the three point system... i guess you could stack three hinges at 120 deg offset?
Have you tried drilling sockets into the thread end of the bolts to cup your balls? Maybe as a proof of concept flip a few cap head bolts upside down so the ball partly fits into the cap. It seems like they would snap into the low energy state which would be centered over the balls.
from where is the image at 1:59? what does it show?
Nice..
For the laser diode and microscope objective holders i can suggest adding a spring .. the part will move laterally on adjustment but will go back to its place
I wonder how printed element compare to eg aluminium one, even made from screwed parts in terms of stiffness and reaction to temperature changes and vibrations after longer period of time. Like for example those tilted 45* mirrors.
3d printing is amazing but I think with delicate devices like optic stuff even cheap aluminium parts crudely made without even CNC cutting and that are just screws together would be better... If you can't keep your project on one flat surface where you could just stack them like those cubes.
What's the deal with phosphorous pentoxide??
Nice hoodie :)
The 3-screw adjustment for optics and lasers could be improved to resolve away the shift by pressing the magnets into the back side of the 3D prints and the screws must reach down through some "guide" holes, maybe with some rubber O-rings pressed into the screw-side to hug around the screws for stabilizing and retaining position better. It wouldn't be completely ideal but it would be a cheap simple way to greatly improve on the concept and lock things down more to where the screws are pretty much the only adjustment that can be effected upon the optical devices.
Thank you, great video but where is the repository?
5:40 with the stands could you use screws shaped to a cone at the bottom, so the sit well located, before you use the screw action to adjust the precise, repeatable location? - like 4gwVucFaPw but smaller
I use a Ball ans sockets for my tripod mounts. You could just find a large enough ball bearing for the center pivot.
About the screw on top of the magnet arrangement, one idea for you:
Drill a small hole on top of the magnet and make the screw end sharp, just like a nail.
I was thinking about spot drilling a cone into the screws. The potential problem I see with a system like this is that it over constrains the part. Maybe that would lead to unexpected shifting when adjusting it.
Thanks for sharing your experience. I would argue that a lot of the things where you said 3D printing is not a good idea - it is actually the specific design you show for the mount, not the fact that they are 3D printed. Using those magnets, for example, and the simplistic right-angled joints on your brackets. With better design considerations you can do probably all this thing with 3D printed mounts and ordinary nuts and bolts. Nice work though.
For diode and microscope, switch position of screws and balls. That way you will not disturb whole diode/microscope when adjusting single screw.
Would be nice to test it in a optical simulation software like 3doptix before building everything.
I’ve been toying with the idea of 3D printing flexture based optics equipment based loosely on OpenFlexture. Maybe something like that could get you the resolution your looking for!
Could you replace the galvo with a DLP projector mirror to pattern a whole 2d array of pixels at once?
AOM (acoustic Optic Modulator) is beter than galvo no inertia !
I hope one day Breaking taps gonna tried a DIY for AOM 😹😹😹
Yes, you can! I recently put together a low-cost stepper and inspection microscope doing something just like this. I'm going to try and start documenting it soon (hopefully!)
Have you tried using 3d printed parts in a high vacuum chamber? Probably when the plastic outgas the moisture it will make for some nices HV things...
What has pushed me away from 3D printing opto-mechanical components is the poor CTE performance of thermoplastics used in FDM printers. Still super impressed with what you've been able to demonstrate on your channel with 3D printed optical parts.
Any idea if resin is any better in that regard?
Just curious; I don't even have a resin printer and have little desire for one just yet :)
@@AndrewGillard Plastics in general have 2X to 10X higher CTE than metals. Also, the poor thermal conductivity of plastics leads to larger temperature gradients across the material, which causes warping. In optical applications, warping is typically worse than just linear expansion and contraction because it results in angular displacement of critical datums.
So we need metal 3D printed optical part holders. Need more money.
Scientific America did a small book on the manufacture of optical mounts it’s quite good.
VERY COOL! ...And wow! @1:58 where is that optical table photo from? The most dense setup I have ever seen. I am an optical science guy and I appreciate your mounts and parts. Nice work. But you also showed some of those designs are not really 'kinematic.' That said, I 3d printed parts to build a LiDAR in 2020 and it worked a lot better than expected. You might have a product at least for education with the magnetic system.
This is a wealth of great information, amazing work! If it's helpful, I suggest using the right "drop-in" t-nuts when possible instead of those finnicky "hammer head" t-nuts. They're far less likely to move around.
Also it looks like you may be using PLA for much of this? The 'printability' is great for prototyping but I've learned going with ABS for parts like this avoids so many headaches later. The inevitable creep with PLA, PETG (and even Nylons) can get so frustrating. I rarely even prototype with PLA now since decent ABS can be very cheap. For finalized revisions, my go-to is usually CF-ASA, or cheaper CF-ABS, for further improved stiffness and shrinkage resistance. You may have a different school of thought on this but I just wanted to throw it out there.
I'd love to work with ABS, but my printer really struggles (not enclosed, poorly tuned). At some point I'll stop being lazy and build an enclosure though, would be great to work with ASA or ABS!
@@BreakingTaps I had a cheap ikea table with foam board walls sitting over a slightly modded Ender 3 running Klipper and it basically printed nothing but ABS parts for the past year. PETG can be more frustrating to print than ABS IMO. Considering what I've seen you do on this channel I think you could figure it out before finishing breakfast ;)
Hot inserts should be on adjecent plate and use springs between...
Would machining a cone into the end of your adjustment screws help with them locating better on the magnets?
This is MY JAM! :D
@1:56 is the structure real?! if, whats the plan? xDDD
Haha, yep! Was from some kind of quantum mechanics experiment: www.mpq.mpg.de/4860026/12_07_26
@@BreakingTaps un-freakin-believable
I'm no expert, but I would assume that annealing the 3d printed parts would make them more rigid and increase the effective TG of the part futher to make it less likely to suffer temperature dependent error when in use. Its a whole 'nother project, but it might be worth checking out. EDIT: Holding brackets for kinematic mounts, @ 10:28 are $18 each on Aliexpress.
Annealed printed parts are much more rigid, however their dimensions change. A good alternative is probably to remelt it in powdered salt: You get the strength of annealed parts with the original dimensional accuracy. They have to be printed with 100% infill though.
You could print molds for forged carbon fibre parts. Carbon fibre has a very low thermal expansion to my knowledge.
I think 3D printed DIY kinematic mounts for mirror (not only) could be optimised with flexture.
I have no clue about anything here, but from the looks of it the UC2 system suffers from the same problem as your laser diode mount? Or am I missing a concept that eliminates that?
The older system does suffer from the "sliding around" phenomenon a bit (it's one of the reasons I didn't end up using it for the confocal). They have a newer version that snaps into place which I think eliminates that problem but I haven't played with it yet. I found that recessing the magnets more than I have in that example helps too. That said, the design of most UC2 components doesn't require adjusting kinematics "inside" the cube, so any misalignment can be tweaked by just fidgeting with the cube. That ends up being pretty easy and you aren't fighting adjustment screws _and_ the magnets like in my designs.
Did you run into issues with temperature dependency due to the PLA changing dimensions over temperature?
For the laser issue, wouldn’t it be a good idea to print a ‘3 sided pyramid’ into the laser base and the same feature into the base with the three magnets…. Side the 3 sided pyramid hole to accept a cheap ball bearing…
The steel ball is sandwiched between and still allows pitch and tilt while constraining is lateral movement.
I like the hoodie
Still hoping you are going to make a video about making telescope mirrors.....I am pretty much stuck on my project and hoping you are going to inspire me or point me in a new direction.
Maybe some day! Looks like a lot of work to get setup (grinding machine and measuring setup, etc) so I've always stayed away, but it would be great to have an optical griding setup to use on various projects 🤔
Could you machine a concave in the end of the screws so they always align with the center of the ball magnets?
still not great as the screws can still wobble in the threaded inserts. I had a long fight with this effect with the print bed of my 3d printer.
Do the heat set inserts have much backlash at the tolerances you're using them at?
For the optical mounts I built, a bit of preload was sufficient to get rid of any backlash.
Sort-of on the subject of optics.... absolutely love the lighting in the talk segments of this video.
And secondly.... where did you get that fantastic hoodie from?.. I want one.... I want one NOW.
I might be mistaken but it is merch from tom from the explosions&fire channel or extractions&ire
Correct! Explosions&Fire merch: explosionsandfire.myspreadshop.com/
@@BreakingTaps would have to be !
YES! I can't believe I caught this 6 minutes after release. I know I have the little bell with all notifications but somehow I still miss it. I wonder if the ball + screw adjustment system could be fixed with a registration point or two
🤗 Apologies you caught it on more of a "useful info but not super interesting topic" sort of video! More of a braindump than anything. :)
@@BreakingTaps I think they are just as good as your "main" ones. A little different in style but in a good sense
I was wondering if some lubrication of the magnets against the bolts would improve things at all - i.e. if it was (at least partly) a case of the magnets not being strong enough to overcome the friction.
Probably still wouldn't be perfect - registration points sounds like a better plan - it was just a thought I had while watching :)
I'm much more of a software engineer than mechanical, and I've not been able to sink as much time as I'd like into designing parts for my 3D printer, so I'm still pretty new at this sort of thing!
Sorry to bother you with another comment.
Regarding the James Webb telescope “golden mirror” in your other video… Was that just a piece of cardboard spray painted gold? A piece of gold anodized aluminum?
Im Just curious as I’ve been toying with the idea of re-creating one while I have access to the equipment. Just for fun and as a display piece, And I would only do it in a batch of one or maybe a few at a time because I think they are gorgeous and the way they made them is absolutely incredible. The technology in those mirrors is insane … Not to mention the interface and adjustment mechanisms. Just breathtakingly awesome
It was CGI actually 😇 I printed a replica to hold in my hand, and then replaced it with a 3d model while editing. I agree it'd be super cool to have a replica as a desk toy!
3D printed bearings...
Taper bearings? maybe externally pressurised hydrostatic bearings?
3D printer could not directly print tapers accurately... But assemble and Keep increasing preload when driven to rotate by say a drill, and high spots should bottom out and might be able to achieve a tollerance similar to the ground glass couplings of chem lab glass wear.
A silly, low effort suggestion: just turn the m3 screws around backwards and the spherical magnets will sit in the hex and maybe give some more precision.
and maybe cut a slot in the bottom of the screw to turn with a tiny screwdriver? dunno. sweet video! good luck!
yunno... that would probably work well! Some other folks have suggested cutting/grinding a sphere on the ends, but just flipping them around would probably work just as well and be a lot easier! Will give it a shot :)
Great stuff...cheers.
Oh my lord what the heck is even happening at 1:57... hard to believe that's even real!
Seriously
Some kind of quantum mechanics Higgs experiment. Don't know if I'm more impressed or horrified for whatever poor grad student has to work on that!
You probably already know about the Chinese built quantum computer using optics, any thoughts on that?
nice hoodie :)
why dont constrain it laterally?
Nice nice content 👍👍👍👍
get a ballend mill and cup the ends of your screws for the magnetic balls. :)
It seems I am not the only one looking for cheaper ways to build optical systems
Did anyone else wince a little inside every time he moved a part around? I had to keep reminding myself that he's already got to realign all the fiddly bits again anyway, and it's okay. 😅
Haha, it made me wince too 😄 Only filmed this after I decided to start working on a v2 and didn't mind messing up the alignment :)
I could tell that he planned to start over by the way he was randomly moving everything, but I was still wincing the whole time.
I'm still wincing reading the comments 😬
Super fan here...
Hi 👋
👋🙂
Your videos are great but the camera shots have a somewhat lower framerate. Maybe something to do with editing, but it really stands out. Other than that I really enjoy them :)
A very interesting topic. I've been thinking about a designer for optomechanics for a long time. The main parts were supposed to be printed. I will be very pleased if you look at them and try them in your work? I will also upload them to GrabCAD. I have blocked other convenient sites. The time is strange.
Honestly, beside the postholders, everything else would have been actually cheaper if you bought the "nice stuff" directly from thorlabs, once you factor in all the time, efforts and materials spent to 3D print stuff that doesn't actually work. Even the postholders, they are actually worth only if you already have a 3D printer, the files correctly sliced and you just have to push the "start" button to get the final pieces: if you have to spend more than an hour and you have a paying job, you would have been better just buying them...
The problem with 3D printing is that it "feels" free, since you don't have to pay someone else the moment you get the finished piece, but it is actually way, WAY more expensive than mass produced items, especially if you value your time in any form...
I agree, but also sorta disagree :) For a business use-case, absolutely. No excuse not to just go buy the real thing and get on with whatever is making money.
For hobby, I think it's a little more difficult to say. Each base/post/lens or mirror assembly would be $60-100 depending on style (fixed element or kinematic). I have maybe 10 different elements, although some of that could change to help decrease cost, but still looking at $500-1000 without blinking. Throw in several redesigns where a few element dimensions change and you're looking at a pretty expensive hobby project.
Now of course, once you purchase the components you have them for future projects so it's not a lost cost. And if you make a good salary the "time cost" is definitely something to consider. But for prototyping I think 3D printing is hard to beat. I used less than a roll of filament and probably no more than a few hours designing everything, so salary + material time it still came out ahead. And lets you make customizations very quickly, instead of waiting 2-3d for the parts to arrive by mail to work on the next iteration (and hope you don't need to re-order again and restart the waiting period).
Tôi giờ mới biết đến nó 🤝👌
Your link to the 'companion video' links right back to itself here instead fyi, it's recursive at the moment.
Argh, I am bad at youtube. Will fix, thanks!
It's a trick to get everyone to keep watching this video over and over again! :D
(And it's an interesting video that I've bookmarked for future reference, so it's worth watching again in a way :p)
where are the printable files??
So awesome and super helpful, just as always!!!
Actually, regarding the magnet bolt problem, couldn't you just add a cone so that the bolts are somewhat restricted? Like this:
_
|| bolt
\ /
|●| magnet
| |
I know, this wouldn't add uber precision, but it could potentially help :)
You know this guys youtube channel : HuygensOptics
He makes crisp stuff....
I bought a $20 green laser on Amazon, now can someone show me how to make my own GPU chip with it? Do I need a microscope and some silicone?
Did you get the design from "Real genius with Val Kilmer"?
If so, you better get your optic clean isn't...
For tip and tilt you only need two adjustment screws. You can fix the 3rd one with a spacer. You will need all 3 in case you want to refocus.
Some simple (wooden) clothes pegs can hold lenses and other stuff without 3d printing or clamping screws. Allways preload adjusting screws for fine adjustment.