Corrections, additions, and FAQs: 1) Thanks to Steve for the idea of a saddlepoint mirror, and giving me the excuse to bust out the silvering solutions again! ua-cam.com/video/sP0cwLuEwsw/v-deo.html 2) Yes, since I know people will ask, I have seen the HEN3DIRK channel that has some incredible videos of electroplated SLA prints. I haven't embarked on that challenge quite yet… ua-cam.com/video/ywPi1osy7o8/v-deo.html 3) "use a resin printer!" yes, you're totally right! I have an old one in a box that I've never ventured to use. it's on my list! In this case, I still doubt I'd have achieved an optical surface, although if you read that paper I flashed up on the screen, they claimed to get something approaching diffraction limited with an SLA printer, which just feels impossible. I think their smoothing resin was also very different from what you can buy on amazon. 3b) Someone in the discussion below about resin printing said that you could buy one with 50nm resolution, and I highly doubt this. you CAN easily get one with sub-50um precision, but unfortunately 1 micron = 1000 nm. light is crazy small 4) z-axis wobble (or any inconsistency on the rails, belts, and screw) ABSOLUTELY could have been at fault for some of the wobblyness. I can't stress how small these changes are. It was difficult to film it in a way where you could tell there was a difference from layer to layer. For every purpose that isn't reflecting light, it was perfect lol. 5) I DID actually sand down my best-performing mirror prior to coating. you can see the sanded part in the "coating" close up, and you can briefly see the green jig with blue tape bubbles chucked into the lathe. for some reason I didn't actually film that?? I put the mirror on the lathe and use a few sandpapers. It made it look better to the eye, but didn't fix everything apparently. 6) I'm doing this crazy silvering process instead of some kind of chrome paint because I was initially trying to make optics, and for that the surface needs to be absolutely perfect. That's also why I didn't do vapor smoothing, which I think can leave a slightly porous surface.
Hi Brian, the way you think about and explain mechanics and physics is great. You said you like reading about science, so I have a link to a science article that I am certain you will enjoy reading. Are you interested?
The striations you saw in your 3D printed mirror are most likely caused by the infill pattern that you used. At 10:00 in the video you can see the deformation caused by the infill in the surface. From testing with PLA it takes about 8~10 mm of wall thickness to get rid of this bleed through tension. For small parts like this the easiest solution is to just print it solid.
Have you considered using a resin 3d printer? They can be bought rather affordably and have much higher accuracy. Given all the work you do with chemicals, a resin printer should not be an issue. All the best and keep up the good work!
Getting a resin printer at work dramatically improved the precision and details of what we could print. Even though it was an expensive early formlabs printer, it more than paid itself off. Also, the UV cured resin is basically just printer resin. :P
@@jormam69 Yes, but it would likely severely reduce the inter-layer inaccuracy that the filament printer was suffering in the video. They'll never be perfect optics, but for something made in a garage where spray bottles, a heat gun, and a flashlight are standard tools, resin printing ought to make for a much higher quality result with not too much change to the process.
Oh thank god…I was worried this was one of those video pairs things that happen quite often where similar UA-camrs post similar videos within a few hours or days of one another and people accuse one or the other of copying the other. I love when y’all relate it to one another. Thank you for this awesome video and I hope y’all two can collab again sometime!
The artifacts you were seeing on the Z is due to your lead screw not being 100% straight, MirageC has a few good videos on the subject and something you can print to help with it. Not sure if there is a solution for the Prusa mini but i would be supersized if there isn't.
I love how these guys kinda regularly work in tandem, basically all the time. Straight up working together though, just made my miracle! of a 3! day weekend!
You can electroplate the mirror with nickel on top and keep it from getting rubbed off! And in general you can plate any other plastic thing after you seed it with silver. There are also silvering solutions based on citric acid, ascorbic acid (yes, vitamin C), EDTA and silver nitrate that plate directly onto PLA. I don't have a 3D printer but I plate stuff, a neighbor gave me some stuff to try, and I couldn't get the silver to stick with Tollens-based alkaline solutions. I realized that NaOH and bases in general slowly de-polymerize PLA. I'm thinking that while it takes quite some time to dissolve a part completely, it's acting fast enough that the surface is getting dissolved too fast for the silver to stick. So I cooked up the silver lemonade I mentioned before. No respirator needed, no messy spray bottles, just soak the thing in the solution. If you wanna try it I can give you more deets :P
@@arthurmoore9488 I just keep stock solutions of every one:. 10mg/mL for the AgNO3, EDTA, and ascorbic acid. 100mg/mL for citric acid. Mix 1 part of nitrate, citric acid and EDTA solutions each. Add 2 parts ascorbic acid soln. to start plating. To activate parts for plating soak in dilute SnCl2 acidified with HCl, then rinse, then soak in dilute AgNO3 (like 10x less than the stock). Concentrations are not super important for this step, but it seeds silver nanoparticles. This will help your part preferentially plate. The bath is truly electroless for a few minutes and only plates activated parts, then it starts silvering the container and eventually crashes out silver powder.
You should try non-planar 3d printing, so that you are able to create the mirror surface as the last layer, without having layer steps or placing the part in a weird way. Same sanding may be required to obtain a completely smooth surface though.
Haha that video by Freya is FANTASTIC! It brought a tear to my eye :) Awesome work on the mirrors!
Рік тому+1
With classic FDM printer there is also a possibility of non-planar printing, creating G-code which moves printing head simultaneously in all three directions. How far one can put this technique depends, as I have read / watched videos about, on the clearance around the printing head. Much better surface for surfaces that are close to being flat. With saddle shape, as mentioned in Steve Mould video, you should be able to create a custom G-code that takes advantage of the fact that the $z = x*y$ shape can be composed entirely out of straight lines at different 3d angles.
polystyrene mirror sheets can be heatshrinked to form this saddle mirror with first surface. You could 3d print the saddle with small holes for the shrink bed. The polystyrene isn't viscous enough to mold to layer lines, so they disappear. Great stuff, thanks for all the vids!
Two suggestions. One is to smooth the surface before silvering it. Sandpaper, glass beads, stuff like that. Or even hitting it with a heat gun to melt it so the surface flows together into a single surface. Also, the rotating technique makes a lot of sense, so why not keep doing that when using the spray bottles?
You could print a negative of your shape, heat acrylic until it becomes soft, and pressform it between the positive and negatives of your forms. As a matter of fact, you could do that with thin aluminum foil and silver that. Perhaps glue the foil to acrylic or pvc sheets and pressform it. Or if you prefer, you could vacuum form it. So many options!
It is amazing what things you know with such passion. I love it. The whole 15 minutes are filled with non-stop non-redundant information. The intuitive genius vibes I am getting might be why it is so good.
I've seen so much about saddles but no one ever goes past the 2-fold "Pringle" to the 3-fold "monkey saddle". Years ago I sanded and polished a dollar coin and bent it into shape using 2 3d printed 3-fold saddle surface forms. The result was quite good, especially after electroplating it with copper to give it a more uniform surface.
Did you consider printing with ABS instead and them smoothing it? I've never done it myself but that should remove the layer lines on top, and then either the UV cure resin might work better, or you could just skip the resin entirely and put the silver straight on top of the ABS.
To get layer artifacts out you could try filler and sanding. If you want to avoid coating the 3D printed part, you could try slumping the acryllic sheet over the 3D printed part in a temperature controlled oven. Acryllic has a glass transition temperature of around 100°C. Polycarbonate 3D print filament has a glass transition temperature of 147°C. Alternatively make a plaster cast and then slump over that.
Should look at nonplanar 3d printing teaching tech has some good videos on that. Would have help a tun with the surface finish on the printed ones, and it seems like an ideal use case for the tech. Might also look into vacuforming for your reflectors. Should be able to make a form on the lathe and just form a sheet to it to get the ideal surface.
I had a very similar problem a while ago and realized that you can get surprisingly good optical quality double-curved mirrors by 3d-printing molds and then thermoforming a aluminium-coated polystyrene sheet. The melted/softened 1mm thick PS sheet does average out the layer steps of FDM prints while the mold is curving the AL layer in the sandwiched polystyrene sheet nicely. I used that for enlarging and redirecting pico laser projectors with an okay-ish quality. There is a video on my channel where I discuss the process and show some results.
That layer inconsistency was something I had to chase in my 3d printing journey for a while. I tracked it down to my v groove wheels (I had an ender 3 at the time), changing them from the TPU wheels to the firmer polycarbonate wheels almost immediately solved my issue. My stock TPU wheels also developed flat spots from preheating in my heated enclosure. With your one-sided prusa, I would blame the wobble and inconsistency on the lack of support.
Would love to see you try this again with a modified slicer for the printer. They have some that can modulate the z-axis at the same time so you have a nice smooth top surface.
Subbed to both of you for a fair while and thought that it was odd that you both had saddle shaped mirrors come out at the same time. Turns out that it wasn't a coincidence.
I would have rigged some variety of centrifuge to "smooth" the resin, with the "top" of the part facing the hub. You could literally make one with some twine or thin rope and swing it in a circle. The effect would be like putting the part under extra Gs, causing the resin to flow more quickly and strengthening surface tension. You could even do it in a sealed container for a "clean" environment.
Oooh, or, alternately, you could use this method just to get a conductive surface, then electroplate the surface until you have a very (relatively) thick layer built up, then just polish to a high gloss.
If you wanted to do this better, you could use the 3dprinter to make abrasive forms, basically make a form to hold your piece and a form to hold sand paper or a guide for some sort of rotary tool and use that to polish away all large imperfections using a higher grit sandpaper, and from there you can coat it with the uv resin and have a much less wavy surface.
Have you ever heard about vapor smoothing of 3D parts? You basically 3D print a part out of ABS or ASA and put it in a air tight jar with acetone on the floor. (heating up a bit gives quicker results) The finish is almost mirror like. The key is to let the part sit for a couple of days to reharden the surface for further coating with your silver applying method.
The filament can vary enough to do that. Also heating fluctuations in the bed and hot end. There is also this cog teeth on the extruder that cause this. I think a post process would be the easiest way to remedy the issue. Unless you want to buy a resin printer or a Bambu labs X1C. Jamie Mantzel did a video on making a clay dome mold from a pendulum. I think this same sort of method could be used to smooth or iron the surface out to a perfect flatness. I haven't put much though into the mechanism for the other shapes of mirrors. The con cave pendulum could use a ball that is heated at the effecting end. With a nonstick sheet between it and the surface of the 3d print. The joint would be the key to making it smooth. There are high quality ball joints for RC cars out there that might fit the bill. You could also smooth something else out on the surface. Like body filler. Jamie's video - ua-cam.com/video/5zgTcK3Ri44/v-deo.html
Aww, I was kinda hoping you'd go into glass making like Huygens Optics... aaand I only mention it case I'd love it as a fallow-up / how-to thing. Neat stuffs.
Silver tarnishing: Mirror stereoscopes - for 3D viewing of aerial photography, have first- reflection silver mirrors on glass, and if kept in their case last for many years without tarnishing. Most degradation is from students fingerprints when the pick them up by grabbing the mirrors, rather than the handles
have you looked into non-planar fdm printing? It's been used for creating smooth model aircraft wings to reduce drag from layer lines. I would guess that it would either get rid of the banding you saw or change it's orientation away from the reflective surface.
There are pre-coated acrylic first surface mirror sheets available that I've used for this kind of stuff before (making curved mirrors for ldgraphy). The surface is pretty sturdy and even survived putting into boiling water to get it easier to bend.
Maybe try SLA 3D prints next time? They have a much finer layer resolution! Either that, or vacuum forming acrylic sheets over a buck! - (You can even buy mirrored acrylic, granted it is second surface!) Awesome video as usual by the way, really enjoyed watching it! 🤓😎
have you heard of non-planar 3d printing? you would basically have the surface layer be just one layer of filament. with the slopes and hills built up with layers beneath. Not sure quite how it would help in this project, but I would imagine it would make this project easier.... somehow.
I saw an anti bird device that was to be hung from the branch of a fruit tree that was a 10 turns helix made of a strip of acrylic. Part of that helix looked like a saddle, my reflection in it was turned 90°. I assume the difference of focal distance between orthogonal directions is controlled by the pitch of the twist. Search “Swivel Bar Bird Repellent Reflective Scare Sticks Spiral Deterrent” to find the thing. The scale is wrong for Steve’s application of course. That could be an inspiration to produce a larger saddle mirror from a larger twisted strip of acrylic. Making a longer strip and doing turns would smooth the holding clamp irregularities.
I have used Angel Gilding products for more than a decade and I can say that the reason that the silver is coming off the sheet is because of the greater surface tension of the sheet so the wetting agent is not as effective as it is with other coatings. The solution for that is to use a Flame Treatment with a propane torch. Another point is that if you really want to print mirrors, just get rid of the FDM Printer and get yourself a SLA Printer, 4K resolution 3D SLA Printers are not even expensive any more(they are now pushing 8/12K). Silver mirrors will tarnish if not protected with a clear coat, but that will make them second surface mirrors. If you really want to 3D print and coat your own mirrors I would suggest investigating on to Electroless Nickel plating
UV cure resin is also used in resin printers in a similar way to the process you described, a mask is put on the LCD and then a UV backlight cured the resin through the mask. More advanced printers don’t use LCDs or anything like that though.
Surely SLA / resin / whatever we're calling it now is the way to go for 3d printing mirrors. Saturn 2 can manage ~20um layer heights and you can use the same resin to smooth it.
From my own personal experience with 3D printing, I'd say your best bet is probably using ASA, ABS or something like that and acetone smoothing it. If done correctly acetone smoothing can give you a surprisingly good mirror finish. Perhaps printing the part with non planar 3D printing, would also help making the part a little smoother to begin with. In that case you'd also be printing it flat, so the z wobble wouldn't have such a pronounced effect. Obviously you also want the smallest layer height possible, the Prusa Mini should be able to go down to 0.05mm, which is probably worth the quadruple print time in this case...
When you pulled out the cast acrylic, i thought you were gonna use one of the 3D-printed saddles as a vacuform buck! That'd potentially be a good way to transfer the macroscale surface profile to the piece of acrylic while retaining the microscale smoothness of the acrylic sheet and mitigating the mid-scale lumpiness of the form.
I'm wondering if you could even heat some glass sheets to warp them into a custom shape, that way the silvering solution is being used as intended and would hopefully be more stable
You may want to try vapor smoothing the 3d printed part with acetone before using your resin. Vapor smoothing, and perhaps pre-sanding, could help at the macro level, now that you have the micro level seemingly more figured out with the uv resin.
A comment I didn’t see scrolling through: Other common causes of those sorts of banding is subpar filament. Either differences in melt rate, though that causes more subtle changes, or more seriously, inconsistent diameters. 3D printers meter their filaments via distance - how many millimeters to advance the filament per non-travel move. If the diameter of the filament is varying even as little as +/- 0.03 mm, that’s still a 3.3% difference in diameter, which is as much as a 10% swing in volumetric extrusion, if my math is right. There’s ways to make the Klipper and other firmwares for 3D printers use a filament width sensor for exactly this reason, but an easier option is to just buy high end, tighter tolerance filament. Edit: I don’t suspect the Z screw simply because the banding doesn’t look very consistent, and you showed a Prusa printer. They’re not very “high spec” as far as features but they’re well built so I doubt that’s the issue. A bent/loose Z screw usually results in a regular pattern.
Dont know if it resin based, but the stuff they fix holes in teeth (other than the older quicksilver thingy) is also UV-hardened. It's awesome doctor fills the whole, puts on the UV light 10 seconds later its as hard as teeth.
Tap water will have an effect because we used to use tap water to test our silver nitrate solutions. The chlorine ions in it will react and precipitate as silver chloride and turn the solution milky with a hint of blue.
Nifty as always ! YT's damn notification system has always been known to have issues but this is at least the 4th time I'm aware of that it failed me for this channel even after re-selecting the bell on like every visit. God forbid they just hire some more humans.
If it doesn’t stick very well to acrylic then you could maybe try polycarbonate as well. It is clear too and you can get it in sheets just like the acrylic.
Maybe you could get better surface quality via some kind of lapping between two or more prints? The prints are decently dimensionally accurate at a large scale, but they're wildly inconsistent at small details. This makes individual prints really bad, but it's very unlikely that two different prints would have the same errors. If you took the average of the two prints, it'd be closer to the shape that you want. Average out enough prints, and you've got a perfect surface. But because these would both be concave surfaces, you can't just lap one against the other, so you'd need to print out some matching convex versions to use as an in-between. It's like a weird version of the Three Plates Method for making flat surfaces. While this method would hopefully/probably/maybe work for telescope mirrors that are axially symmetric, I'm not sure if it would work for the saddlepoint mirrors. I guess while you couldn't rotate them all the way around, you could jostle them back and forth a little? Still a lapping action, but with movement measured in millimeters.
i don't think you need the in-between plate as long as you don't do relative rotation, you can get a quite nice hyperbolic surface as demonstrated here by ROBRENZ ua-cam.com/video/Va4TGDnQqDs/v-deo.html
My guess on the issues with the 3-D printed mirror are because you used an STL. I don’t believe PrusaSlicer allows the use of STEP files, but some thing like Bambu or Orca Slicer does. That would likely get you less faceting and a smoother surface finish overall. I’m still not fully up to speed on this, but you might need a 32 bit printer to print step files with arcs in the G code instead of line segments. Maybe someone else who is experienced printing STEP files can chime in.
Next on someone's to do list, if they haven't done it already... is to make a 6 way actuator to rotate the axis up-down-sideways and around to balance and keep a ball bearing as close to the center point as possible.
You need to do a collab with Applied Science or tech ingredients. They could definitely help you perfect this technique. Both of them have done similar things.
Part 2: Make a perfect saddle shape out of wood then vacuum form some plastic around it and finally silver it. That should give you a perfectly flat surface and the wooden jig is much easier to work with than a plastic or resin one.
I think the reason why resin did not form very evenly is that there was just way too much of it. Spin-coat the blank to get a really thin even layer, then slow it down and cure while slowly rotating. If one is not enough, repeat the process again. Don't try to do it all in one go. As for parabolic mirrors in general, spin a petri dish full of resin. It will naturally form a perfect parabola that can then be cured with the UV light while continuing to spin.
Corrections, additions, and FAQs:
1) Thanks to Steve for the idea of a saddlepoint mirror, and giving me the excuse to bust out the silvering solutions again! ua-cam.com/video/sP0cwLuEwsw/v-deo.html
2) Yes, since I know people will ask, I have seen the HEN3DIRK channel that has some incredible videos of electroplated SLA prints. I haven't embarked on that challenge quite yet… ua-cam.com/video/ywPi1osy7o8/v-deo.html
3) "use a resin printer!" yes, you're totally right! I have an old one in a box that I've never ventured to use. it's on my list! In this case, I still doubt I'd have achieved an optical surface, although if you read that paper I flashed up on the screen, they claimed to get something approaching diffraction limited with an SLA printer, which just feels impossible. I think their smoothing resin was also very different from what you can buy on amazon.
3b) Someone in the discussion below about resin printing said that you could buy one with 50nm resolution, and I highly doubt this. you CAN easily get one with sub-50um precision, but unfortunately 1 micron = 1000 nm. light is crazy small
4) z-axis wobble (or any inconsistency on the rails, belts, and screw) ABSOLUTELY could have been at fault for some of the wobblyness. I can't stress how small these changes are. It was difficult to film it in a way where you could tell there was a difference from layer to layer. For every purpose that isn't reflecting light, it was perfect lol.
5) I DID actually sand down my best-performing mirror prior to coating. you can see the sanded part in the "coating" close up, and you can briefly see the green jig with blue tape bubbles chucked into the lathe. for some reason I didn't actually film that?? I put the mirror on the lathe and use a few sandpapers. It made it look better to the eye, but didn't fix everything apparently.
6) I'm doing this crazy silvering process instead of some kind of chrome paint because I was initially trying to make optics, and for that the surface needs to be absolutely perfect. That's also why I didn't do vapor smoothing, which I think can leave a slightly porous surface.
I would have 3D printed only the perimeter and then shrink-wrap it
hey how can i contact you ?
Why not try a resin 3d printer? The layer line issue is almost non existent.
Resin 3D printing is much smoother and much stronger than FDM printing
Hi Brian, the way you think about and explain mechanics and physics is great. You said you like reading about science, so I have a link to a science article that I am certain you will enjoy reading. Are you interested?
Nice. They popped up on my feed next to each other.
Had an Atomic Shrimp video stuck in the middle
K-Mart Blue Light Special Thor and science communicator Kyle Hill somehow snuck in between them on my feed.
"The Continuity of Splines". It's a good video. I find it cool that many of my favorite science channels know and/or watch each other.
I'm so glad you referenced Freya's video on splines. it's truly a remarkable video
The striations you saw in your 3D printed mirror are most likely caused by the infill pattern that you used. At 10:00 in the video you can see the deformation caused by the infill in the surface. From testing with PLA it takes about 8~10 mm of wall thickness to get rid of this bleed through tension. For small parts like this the easiest solution is to just print it solid.
Have you considered using a resin 3d printer? They can be bought rather affordably and have much higher accuracy. Given all the work you do with chemicals, a resin printer should not be an issue. All the best and keep up the good work!
Getting a resin printer at work dramatically improved the precision and details of what we could print. Even though it was an expensive early formlabs printer, it more than paid itself off.
Also, the UV cured resin is basically just printer resin. :P
good optics require precision which is a fraction of a wavelength so no printer would be able to do this
@@jormam69 Yes, but it would likely severely reduce the inter-layer inaccuracy that the filament printer was suffering in the video. They'll never be perfect optics, but for something made in a garage where spray bottles, a heat gun, and a flashlight are standard tools, resin printing ought to make for a much higher quality result with not too much change to the process.
@@PyroNicampt True!
@@jormam69 he did not try to build a space telescope!
Hell yeah! That spline continuity video is amazing. I think about it every time I see reflective surfaces now lol.
eyyy shoutout to the continuity of splines video!
Oh thank god…I was worried this was one of those video pairs things that happen quite often where similar UA-camrs post similar videos within a few hours or days of one another and people accuse one or the other of copying the other. I love when y’all relate it to one another. Thank you for this awesome video and I hope y’all two can collab again sometime!
The artifacts you were seeing on the Z is due to your lead screw not being 100% straight, MirageC has a few good videos on the subject and something you can print to help with it. Not sure if there is a solution for the Prusa mini but i would be supersized if there isn't.
I realized I should have added that to the list but didn’t redo the voiceover. Could have been the screw or the rails!
Great collaboration, wonderful to see you two working together on this cool topic
I love how these guys kinda regularly work in tandem, basically all the time. Straight up working together though, just made my miracle! of a 3! day weekend!
Freya's videos on splines are absolutely stunning.
I really love when my favorite educators interacting with each other
Steve Mould? Man, it's such an honor to get to collaborate with such a high caliber creator, that Mould guy should be really happy.
You can electroplate the mirror with nickel on top and keep it from getting rubbed off!
And in general you can plate any other plastic thing after you seed it with silver.
There are also silvering solutions based on citric acid, ascorbic acid (yes, vitamin C), EDTA and silver nitrate that plate directly onto PLA.
I don't have a 3D printer but I plate stuff, a neighbor gave me some stuff to try, and I couldn't get the silver to stick with Tollens-based alkaline solutions. I realized that NaOH and bases in general slowly de-polymerize PLA. I'm thinking that while it takes quite some time to dissolve a part completely, it's acting fast enough that the surface is getting dissolved too fast for the silver to stick. So I cooked up the silver lemonade I mentioned before.
No respirator needed, no messy spray bottles, just soak the thing in the solution. If you wanna try it I can give you more deets :P
This sounds really interesting. What ratio did you use of each of those?
@@arthurmoore9488 I just keep stock solutions of every one:. 10mg/mL for the AgNO3, EDTA, and ascorbic acid. 100mg/mL for citric acid.
Mix 1 part of nitrate, citric acid and EDTA solutions each. Add 2 parts ascorbic acid soln. to start plating.
To activate parts for plating soak in dilute SnCl2 acidified with HCl, then rinse, then soak in dilute AgNO3 (like 10x less than the stock). Concentrations are not super important for this step, but it seeds silver nanoparticles. This will help your part preferentially plate. The bath is truly electroless for a few minutes and only plates activated parts, then it starts silvering the container and eventually crashes out silver powder.
@@arthurmoore9488 EDTA is the disodium salt btw, Na2EDTA, not the acid itself
I come from steves video, because i was told to watch yours, now you tell me to go watch steves. This is a neverending loop.
I watched this video because Steve's video sent me here.
Then I realized I was already subscribed to your channel.
Nice!
I swear I have never clicked a video this fast, this is truly one of my favorite channels. THANK YOU.
You should try non-planar 3d printing, so that you are able to create the mirror surface as the last layer, without having layer steps or placing the part in a weird way. Same sanding may be required to obtain a completely smooth surface though.
Haha that video by Freya is FANTASTIC! It brought a tear to my eye :) Awesome work on the mirrors!
With classic FDM printer there is also a possibility of non-planar printing, creating G-code which moves printing head simultaneously in all three directions. How far one can put this technique depends, as I have read / watched videos about, on the clearance around the printing head.
Much better surface for surfaces that are close to being flat.
With saddle shape, as mentioned in Steve Mould video, you should be able to create a custom G-code that takes advantage of the fact that the $z = x*y$ shape can be composed entirely out of straight lines at different 3d angles.
polystyrene mirror sheets can be heatshrinked to form this saddle mirror with first surface. You could 3d print the saddle with small holes for the shrink bed. The polystyrene isn't viscous enough to mold to layer lines, so they disappear. Great stuff, thanks for all the vids!
Numberphile also released a video on Parabolic Mirrors! What a Convergence of knowledge.
I'm telling you, it's a conspiracy to make us all smarter, people! Wake u... I mean, go back to bed!!! ;)
Two suggestions. One is to smooth the surface before silvering it. Sandpaper, glass beads, stuff like that. Or even hitting it with a heat gun to melt it so the surface flows together into a single surface. Also, the rotating technique makes a lot of sense, so why not keep doing that when using the spray bottles?
I have been subscribed to both you and Steve Mold for quite awhile. I love your honest approach to doing science.
You could print a negative of your shape, heat acrylic until it becomes soft, and pressform it between the positive and negatives of your forms. As a matter of fact, you could do that with thin aluminum foil and silver that. Perhaps glue the foil to acrylic or pvc sheets and pressform it. Or if you prefer, you could vacuum form it. So many options!
It is amazing what things you know with such passion. I love it. The whole 15 minutes are filled with non-stop non-redundant information. The intuitive genius vibes I am getting might be why it is so good.
I remember my preliminary investigation into lens making and how fast I noped out of that as a thing I reasonably expected I could do 😂 Awesome job!!
Here from Steve's channel. Very glad to be here, especially as someone with my surname.
A collaboration between a couple of my favorite science edutainers.. Yes please.
Great to see you getting collabs from some of the big-time UA-camrs
I was excited for a collab when I saw both AlphaPhoenix and Steve show up in notifications with the same subject. Great job guys!
I've seen so much about saddles but no one ever goes past the 2-fold "Pringle" to the 3-fold "monkey saddle". Years ago I sanded and polished a dollar coin and bent it into shape using 2 3d printed 3-fold saddle surface forms. The result was quite good, especially after electroplating it with copper to give it a more uniform surface.
Did you consider printing with ABS instead and them smoothing it? I've never done it myself but that should remove the layer lines on top, and then either the UV cure resin might work better, or you could just skip the resin entirely and put the silver straight on top of the ABS.
And you could even to the acetone bath while spinning to eliminate any drooping
He said vapor smoothing wasn't done due to the possibility of pores
@@benjaminshields9421 oh, mustve missed it, do you have the timestamp?
I saw both thumbnails and my expectations were met.
Try using a high resolution resin printer, even hobby grade machines can create voxels at 20µm!
How have both Sebastian Lague and Alpha Phoenix released a video in the same day that includes a bit about Freya Holmer? Neat coincidence
Amazing crossover episode!
To get layer artifacts out you could try filler and sanding. If you want to avoid coating the 3D printed part, you could try slumping the acryllic sheet over the 3D printed part in a temperature controlled oven. Acryllic has a glass transition temperature of around 100°C. Polycarbonate 3D print filament has a glass transition temperature of 147°C. Alternatively make a plaster cast and then slump over that.
Should look at nonplanar 3d printing teaching tech has some good videos on that. Would have help a tun with the surface finish on the printed ones, and it seems like an ideal use case for the tech.
Might also look into vacuforming for your reflectors. Should be able to make a form on the lathe and just form a sheet to it to get the ideal surface.
That video on splines is fantastic
Just found your channel through Steve's. Excellent combination of instructive and humourous. Keep up the content and collabs!
I had a very similar problem a while ago and realized that you can get surprisingly good optical quality double-curved mirrors by 3d-printing molds and then thermoforming a aluminium-coated polystyrene sheet. The melted/softened 1mm thick PS sheet does average out the layer steps of FDM prints while the mold is curving the AL layer in the sandwiched polystyrene sheet nicely. I used that for enlarging and redirecting pico laser projectors with an okay-ish quality. There is a video on my channel where I discuss the process and show some results.
That layer inconsistency was something I had to chase in my 3d printing journey for a while. I tracked it down to my v groove wheels (I had an ender 3 at the time), changing them from the TPU wheels to the firmer polycarbonate wheels almost immediately solved my issue. My stock TPU wheels also developed flat spots from preheating in my heated enclosure. With your one-sided prusa, I would blame the wobble and inconsistency on the lack of support.
Would love to see you try this again with a modified slicer for the printer. They have some that can modulate the z-axis at the same time so you have a nice smooth top surface.
This is still a fairly manual process with post processing scripts but will be interesting to see when 3D slicing is more available and automatic
I was going to say "weird that you and Steve mould put out videos about weird mirrors in the same day." Makes more sense that this was a colab.
Subbed to both of you for a fair while and thought that it was odd that you both had saddle shaped mirrors come out at the same time. Turns out that it wasn't a coincidence.
10:17 FREYA!!!! i loved her videos about bezier curves and splines!!
Jumped out of my chair when I saw the splines video's mention lmao I just watched it a couple days ago and it's such a great video
Reasonably effective, affordable, expendable first-surface mirrors? Nice.
I would have rigged some variety of centrifuge to "smooth" the resin, with the "top" of the part facing the hub. You could literally make one with some twine or thin rope and swing it in a circle.
The effect would be like putting the part under extra Gs, causing the resin to flow more quickly and strengthening surface tension. You could even do it in a sealed container for a "clean" environment.
I thought about this! I think it’s a great idea but couldn’t figure out a simple way to build it
Oooh, or, alternately, you could use this method just to get a conductive surface, then electroplate the surface until you have a very (relatively) thick layer built up, then just polish to a high gloss.
If you wanted to do this better, you could use the 3dprinter to make abrasive forms, basically make a form to hold your piece and a form to hold sand paper or a guide for some sort of rotary tool and use that to polish away all large imperfections using a higher grit sandpaper, and from there you can coat it with the uv resin and have a much less wavy surface.
Have you ever heard about vapor smoothing of 3D parts?
You basically 3D print a part out of ABS or ASA and put it in a air tight jar with acetone on the floor. (heating up a bit gives quicker results)
The finish is almost mirror like. The key is to let the part sit for a couple of days to reharden the surface for further coating with your silver applying method.
The filament can vary enough to do that. Also heating fluctuations in the bed and hot end. There is also this cog teeth on the extruder that cause this. I think a post process would be the easiest way to remedy the issue. Unless you want to buy a resin printer or a Bambu labs X1C. Jamie Mantzel did a video on making a clay dome mold from a pendulum. I think this same sort of method could be used to smooth or iron the surface out to a perfect flatness. I haven't put much though into the mechanism for the other shapes of mirrors. The con cave pendulum could use a ball that is heated at the effecting end. With a nonstick sheet between it and the surface of the 3d print. The joint would be the key to making it smooth. There are high quality ball joints for RC cars out there that might fit the bill. You could also smooth something else out on the surface. Like body filler. Jamie's video - ua-cam.com/video/5zgTcK3Ri44/v-deo.html
That's crazy, Steve Mould just posted a video on the same topic!
Oh...
Aww, I was kinda hoping you'd go into glass making like Huygens Optics... aaand I only mention it case I'd love it as a fallow-up / how-to thing. Neat stuffs.
man i love these collabs!
Man . . . i also watched that 1 hour-long spline video around a week ago XD . . . youtube recommendation is surprising XD
now THIS is the most ambitious crossover of all time (and this joke is soo old...)
Silver tarnishing: Mirror stereoscopes - for 3D viewing of aerial photography, have first- reflection silver mirrors on glass, and if kept in their case last for many years without tarnishing. Most degradation is from students fingerprints when the pick them up by grabbing the mirrors, rather than the handles
Awesome video! Such an interesting connection between manufacturing and optics.
Excellent crossover!
I've been subscribed since before 2^12 and while I don't watch every video, you do deserve every view you get. You're amazing.
have you looked into non-planar fdm printing?
It's been used for creating smooth model aircraft wings to reduce drag from layer lines. I would guess that it would either get rid of the banding you saw or change it's orientation away from the reflective surface.
There are pre-coated acrylic first surface mirror sheets available that I've used for this kind of stuff before (making curved mirrors for ldgraphy). The surface is pretty sturdy and even survived putting into boiling water to get it easier to bend.
Maybe try SLA 3D prints next time? They have a much finer layer resolution!
Either that, or vacuum forming acrylic sheets over a buck! - (You can even buy mirrored acrylic, granted it is second surface!)
Awesome video as usual by the way, really enjoyed watching it! 🤓😎
I need to make some elliptical mirror that has 2 focus length so that light passes through 1 will always pass through another. Thanks for the video!
have you heard of non-planar 3d printing? you would basically have the surface layer be just one layer of filament. with the slopes and hills built up with layers beneath. Not sure quite how it would help in this project, but I would imagine it would make this project easier.... somehow.
I saw an anti bird device that was to be hung from the branch of a fruit tree that was a 10 turns helix made of a strip of acrylic. Part of that helix looked like a saddle, my reflection in it was turned 90°. I assume the difference of focal distance between orthogonal directions is controlled by the pitch of the twist. Search “Swivel Bar Bird Repellent Reflective Scare Sticks Spiral Deterrent” to find the thing. The scale is wrong for Steve’s application of course. That could be an inspiration to produce a larger saddle mirror from a larger twisted strip of acrylic. Making a longer strip and doing turns would smooth the holding clamp irregularities.
I have used Angel Gilding products for more than a decade and I can say that the reason that the silver is coming off the sheet is because of the greater surface tension of the sheet so the wetting agent is not as effective as it is with other coatings. The solution for that is to use a Flame Treatment with a propane torch. Another point is that if you really want to print mirrors, just get rid of the FDM Printer and get yourself a SLA Printer, 4K resolution 3D SLA Printers are not even expensive any more(they are now pushing 8/12K). Silver mirrors will tarnish if not protected with a clear coat, but that will make them second surface mirrors. If you really want to 3D print and coat your own mirrors I would suggest investigating on to Electroless Nickel plating
UV cure resin is also used in resin printers in a similar way to the process you described, a mask is put on the LCD and then a UV backlight cured the resin through the mask. More advanced printers don’t use LCDs or anything like that though.
Amazing video bro
Channel growing well, glad to see it, you seem like a nice fella :)
Get ready for 2^20. Lots of Steve Mould viewers coming in.
Surely SLA / resin / whatever we're calling it now is the way to go for 3d printing mirrors. Saturn 2 can manage ~20um layer heights and you can use the same resin to smooth it.
From my own personal experience with 3D printing, I'd say your best bet is probably using ASA, ABS or something like that and acetone smoothing it. If done correctly acetone smoothing can give you a surprisingly good mirror finish. Perhaps printing the part with non planar 3D printing, would also help making the part a little smoother to begin with. In that case you'd also be printing it flat, so the z wobble wouldn't have such a pronounced effect. Obviously you also want the smallest layer height possible, the Prusa Mini should be able to go down to 0.05mm, which is probably worth the quadruple print time in this case...
When you pulled out the cast acrylic, i thought you were gonna use one of the 3D-printed saddles as a vacuform buck! That'd potentially be a good way to transfer the macroscale surface profile to the piece of acrylic while retaining the microscale smoothness of the acrylic sheet and mitigating the mid-scale lumpiness of the form.
I'm wondering if you could even heat some glass sheets to warp them into a custom shape, that way the silvering solution is being used as intended and would hopefully be more stable
You may want to try vapor smoothing the 3d printed part with acetone before using your resin. Vapor smoothing, and perhaps pre-sanding, could help at the macro level, now that you have the micro level seemingly more figured out with the uv resin.
"... a perplexingly patchy patina." Haha! ... Brilliant! This episodes sponsor! ... maybe ... dunno, but brilliant all the same! :)
A comment I didn’t see scrolling through: Other common causes of those sorts of banding is subpar filament. Either differences in melt rate, though that causes more subtle changes, or more seriously, inconsistent diameters. 3D printers meter their filaments via distance - how many millimeters to advance the filament per non-travel move. If the diameter of the filament is varying even as little as +/- 0.03 mm, that’s still a 3.3% difference in diameter, which is as much as a 10% swing in volumetric extrusion, if my math is right. There’s ways to make the Klipper and other firmwares for 3D printers use a filament width sensor for exactly this reason, but an easier option is to just buy high end, tighter tolerance filament.
Edit: I don’t suspect the Z screw simply because the banding doesn’t look very consistent, and you showed a Prusa printer. They’re not very “high spec” as far as features but they’re well built so I doubt that’s the issue. A bent/loose Z screw usually results in a regular pattern.
That's a really great video on splines
Make form of a print, lay the acrylic on top, heat the acrylic and it will soften and shape it self as the form, silvering it and - done.
Dont know if it resin based, but the stuff they fix holes in teeth (other than the older quicksilver thingy) is also UV-hardened. It's awesome doctor fills the whole, puts on the UV light 10 seconds later its as hard as teeth.
You and steve mould are psychicly linked
Tap water will have an effect because we used to use tap water to test our silver nitrate solutions. The chlorine ions in it will react and precipitate as silver chloride and turn the solution milky with a hint of blue.
Nifty as always ! YT's damn notification system has always been known to have issues but this is at least the 4th time I'm aware of that it failed me for this channel even after re-selecting the bell on like every visit. God forbid they just hire some more humans.
Now I feel like I need a alpha phoenix and huygens optics collab on making mirrors
If it doesn’t stick very well to acrylic then you could maybe try polycarbonate as well. It is clear too and you can get it in sheets just like the acrylic.
Maybe you could get better surface quality via some kind of lapping between two or more prints? The prints are decently dimensionally accurate at a large scale, but they're wildly inconsistent at small details. This makes individual prints really bad, but it's very unlikely that two different prints would have the same errors. If you took the average of the two prints, it'd be closer to the shape that you want. Average out enough prints, and you've got a perfect surface. But because these would both be concave surfaces, you can't just lap one against the other, so you'd need to print out some matching convex versions to use as an in-between. It's like a weird version of the Three Plates Method for making flat surfaces.
While this method would hopefully/probably/maybe work for telescope mirrors that are axially symmetric, I'm not sure if it would work for the saddlepoint mirrors. I guess while you couldn't rotate them all the way around, you could jostle them back and forth a little? Still a lapping action, but with movement measured in millimeters.
🤔
i don't think you need the in-between plate
as long as you don't do relative rotation, you can get a quite nice hyperbolic surface
as demonstrated here by ROBRENZ ua-cam.com/video/Va4TGDnQqDs/v-deo.html
wow it was the most extremely interesting vídeo i've see this week...
..perhapps you can mix someway a "fresnel /inspired" device with this designs using also somehow gravity working for you side???....
My guess on the issues with the 3-D printed mirror are because you used an STL. I don’t believe PrusaSlicer allows the use of STEP files, but some thing like Bambu or Orca Slicer does. That would likely get you less faceting and a smoother surface finish overall.
I’m still not fully up to speed on this, but you might need a 32 bit printer to print step files with arcs in the G code instead of line segments. Maybe someone else who is experienced printing STEP files can chime in.
Might be time for you to watch Peter Brown polish regular epoxy with micro mesh. Your printed base shapes with epoxy coatings would've been perfect.
Next on someone's to do list, if they haven't done it already... is to make a 6 way actuator to rotate the axis up-down-sideways and around to balance and keep a ball bearing as close to the center point as possible.
You need to do a collab with Applied Science or tech ingredients. They could definitely help you perfect this technique. Both of them have done similar things.
Part 2: Make a perfect saddle shape out of wood then vacuum form some plastic around it and finally silver it. That should give you a perfectly flat surface and the wooden jig is much easier to work with than a plastic or resin one.
Great video! I love the 3d printed mirror videos!
OMG I totally love how you love what you do.
Nice
I think the reason why resin did not form very evenly is that there was just way too much of it. Spin-coat the blank to get a really thin even layer, then slow it down and cure while slowly rotating. If one is not enough, repeat the process again. Don't try to do it all in one go.
As for parabolic mirrors in general, spin a petri dish full of resin. It will naturally form a perfect parabola that can then be cured with the UV light while continuing to spin.