Robert Warden here. I wrote the paper back in 2006. I just wanted to say how impressed I am with your reverse engineering! Your graphics and description are very well done. Back then, we didn't have easy access to 3D printers, so I built the first model out of Legos, which is still on my desk! Wishing you all the best - Bob
Hi! Thanks for stopping by, never thought the author of the paper would leave a comment! I wanted to convey my gratitude to you and your colleagues for working on the mechanism and publishing details about it, it was really fun to work through the paper and figure out how all the little pieces interacted. I can't imagine trying to design something like that from scratch though! It's a really remarkable piece of engineering. That's awesome that the first model was done in Legos. Maybe we can petition Lego to make a kit for it 😀 Cheers!
I somehow stumbled upon your paper a few months ago and I found it rather informative. I learnt a lot reading it. And suddenly this video popped up with a working replica. Amazing
Wow really interesting, thanks for making this video. I had wondered how they did this! I've been thinking "how do they handle the backlash on such an actuator?" Turns out..... they LEAN INTO IT. Well done!
Thanks Destin, glad to hear you enjoyed it! It's such a clever and smart mechanism, makes me happy that other folks found it just as interesting. too Cheers for your JWST episodes btw, really enjoyed seeing all the behind-the-scenes engineering details!
I'm one of 20 flight systems engineers who will be responsible for the JWST spacecraft over the life of the mission. I basically understood how this mechanism worked but seeing it in action is hugely informative. Thank you for taking the time to learn, reverse engineer, and share this with the world. This video will become a standard part of our training program.
Wow!... Does a 10 billion dollar space program really need to use a youtube creator's content for its operational training without any sort of commercial arrangement in place?..... As a fellow engineer I know you're well paid for your work. I also bet that if a 10 billion dollar project helped themselves to your work product that you created with your own resources on your own time without any compensation other than a thankyou note you might have something to say in a court of law about it, I know I would.... how is helping yourself to this creator's content for commercial purposes any different?
@@GRDwashere I appreciate your concern for @Breaking Taps intellectual property for I too wish to see this channel succeed, but one of us is missing how RU-vid works. My understanding is that creators put their content on RU-vid for fair use. We watch, like, subscribe, and leave reviews of the work... which in this case merits glowing admiration.
I led the extremely talented team that developed the Telescope Elements at Ball (Primary Mirror Assemblies, Secondary Mirror Assembly and Aft Optics Subsystem). I must say the reverse engineering of Bob's actuator shown in this video is impressive. What will really blow your mind is how we designed the hexapod and other support structure to not deform the mirror surface during hexapod motion at 50 kelvin, while having the structural integrity to survive launch. Also mind bending is how we polished the mirrors at ambient after taking them cold to create a deformation map at cryo and then polishing the inverse at ambient. crazy tech. BTW, these same actuators are used for the ROC (Radius of Curvature) actuator in the center of each primary mirror segment as well as the six secondary mirror hexapod actuators. Good video.
Do you know Bob Warden who developed the actuators? That had been one thing I has been wondering, how such cold temperatures might cause deformation to parts. Bob indicated that while space is very cold it is also very stable, but part of me wonders how those kinds of conditions could have been accurately replicated on earth to begin with.
The sophistication of the JWST mirror actuators is especially interesting to me in contrast with ours. I work on the motion control hardware and software systems for the Hobby-Eberly telescope, a 1990s telescope design with 91 hexagonal segments. While our mirror segments have a similar job to do and roughly similar levels of precision, our system is much simpler, with only three degrees of freedom -- tip, tilt, and piston (no translation or rotation, and we can't warp our ~115kg ceramic glass mirror segments). Our system has three separate actuators per segment, with a similar level of gear and lever reduction, but a very different mechanical design. I'm really impressed with the level of thought that went into the JWST actuator design, and just as amazed as anyone else at the engineering effort that has gone into building such a system that must operate for years with no maintenance.
Very cool! I'm more than a little jealous of your job :) That's interesting that the Hobby-Eberly uses 3dof system... I wonder why they decided to go with 7 for JWST? Just to prevent any potential Hubble-esque accidents and be able to really correct anything that goes wrong once in orbit? Any further reading about how Hobby-Eberly actuators work? Would love to take a peek at them! :)
@@BreakingTaps, I think the difference in wavelengths and focal length have something to do with it... JWST is meant for wavelengths of 0.6-28.3 μm with focal length of 131.4m whereas Hobby-Eberly is for wavelengths from 350 nm to 1800 nm with focal length of 13.08m.
@@chronokoks well, it’s not really a question of the age of the companies that determines how well they stick to the budget. Lockheed’s skunkworks division routinely came in under budget
In case you use them on things beyond demos, it's worth noting that those small steppers don't have an exact 1:64 reduction. Instead they have a multistage gear train which results in a 25792/405 ~= 63.684 gear reduction EDIT: I received a follow up question and decided to crack open one of my steppers to confirm and found that it has a different gear train than some other manufacturers! You can find reports of the 25792/405 = (31*32*26*22)/(11*10*9*9) reduction on the arduino forums. Mine contains a (32/9)*(22/11)(27/9)*(24/8) = 64:1 drive train. So it seems it can vary and you need to test which one you have, or open up the case and start counting.
This was amazing to watch, you really did it justice! Dad initially made the actuator out of legos, it’s great to see the final product in space! Thanks for an enjoyable presentation!
@@mitlanderson I had to convince him to write a comment! I am so proud of what my dad has accomplished over the years, and it is so cool to see this many people interested in a mechanism he created over 15 years ago!
@@rondawarden6346 you should be very proud, your father has had a hand in building one of the collections mechanisms for one of our most technologically advanced creations in human history. He's been gracious enough to answer my questions and being a Land Surveyor I've discovered we've got some crossover. So cool!
@@rondawarden6346 Oh wow. I thought it was cool to see that the person who wrote the paper had commented. I think it's somehow even cooler to learn that the reason they did so was that one of their children talked them into it. :) Nice work, Ronda! (And nice work, "Dad"! Love the lego model!!)
This is the first time I see compliant mechanisms being used for their absolute full potential where no other type of system could achieve such simplicity and effectiveness. Truely an amazing design!
Very slick! Being able to drive fine and coarse with a single motor like that is absolutely fantastic. I could kinda see it coming part way through when I saw both the cam and the bevels, but I couldn't figure out how the bottom didn't move all the time... that double disk is hilariously great. Question though - how does it prevent back-driving? you mentioned at the beginning wanting something that didn't need an active motor to station-keep, so I'm assuming it's possible to turn the stepper off and have it stay put., but there's no worm and wheel or anything that I would "expect" to make it one-way. does it have to do with the flexure mechanism, or just a lot of friction in the preliminary planetary?
Oh! That's my fault, I totally forgot to mention it. There's a little friction brake hidden inside the mechanism, down by the "coarse" gear. Before disengaging the coarse stage they set the friction brake which prevents the flexure from backdriving the system. Details are pretty sparse, they don't really talk about how the brake is actuated but it sounds like another flexure and probably something like a solenoid that's held open and closes shut when de-energized, if I had to guess. From the paper: > The coarse drive gear was chosen for the brake location because it is the largest torsional element on the coarse drive shaft thereby requiring the least force to constrain. The brake uses a double cantilevered beam to support and apply force to two Vespel buttons. The buttons slide along a raised surface on the coarse drive gear And yeah, I love the simplicity of the double disk arrangement. So simple and elegant! Another aspect I didn't mention is that the fine stage is contributing to the distance (in a sinusoidal pattern) while the coarse stage is running, so they actually have it all modeled out to know how much the whole thing is actually moving. In addition to an LVDT sensor monitoring progress too.
@@BreakingTaps There are some patents and papers online that discuss it - there's no actuation, it's always providing friction. So it just makes that gear a little stiff to turn, but it stays put when not driven.
Fun fact - that double-disk mechanism is the same mechanism used in safes that use rotating dials. It's a really awesome mechanism because you could essentially add as many disks as you want, to get as many outputs as you need, with just a single stepper motor. Here's a cool video by Jared Owen visualising the same disk mechanism used with three disks in a combination lock: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-sftkP4CjjZs.html
I worked with optical engineers and scientists at Eastman Kodak. So I totally enjoyed the challenge you took up; to create a model of the JWST hexagonal mirror actuators. You really filled in the details I was wondering about. BALL Aerospace or NASA should thank you or recognize you for your effort to science/tech educate others. Awesome! Thanks very much.
When I first saw your channel, I thought it was going to be "just" another machinist channel (which is no bad thing) but your take on "engineering meets science" really does hit the spot. Keep up The Great Work.
Thanks! And yeah, my channel has wandered all over the place since it's inception... the name doesn't quite fit anymore. Oh well! I'm sure I'll break some more taps in the future just to keep the name relevant 😇
@@BreakingTaps Thanks for being more than just machining. It's so much cooler to see what machining can make and how far you can push the limits to create really amazing things. More so than most machining channels ever try to do anyway. TOT is still magic though.
🎙New shop space... new acoustic problems! I was so excited to film this video that I neglected to put up acoustic treatment first. Sadly the reverb ended up being awful (no surprise, it's literally a metal-walled barn). I knocked the reverb down in post but tried to be gentle to avoid artifacts; hopefully it doesn't sound too much like a tin-can or robo-voice. I'll get some acoustic panels on the walls ASAP! Thanks for your patience! 🙏
You've got one of the best produced channels on RU-vid, especially for a technical channel, and we're all watching it on our phones.... Don't sweat it dude, you're doing amazing work in every aspect
I also found that paper and enjoyed the technical detail. However, I definitely had to reread it multiple times to understand what all was involved. I appreciate this video explanation that I'm sure will let more people understand the design.
My significant other, who normally is not at all interested in space or tech stuff, had read about the JWST in some magazine, and was for some reason blown away by the nanometer scale adjustments possible in the mirrors. I thought it was adorable =) She couldn't wrap her mind around how such a thing was possible. And neither could I (the family space nerd). So huge thank you for this enlightening video explaining one of the most remarkable details about the JWST that "no one is talking about".
It absolutely blows my mind that there are minds in our world that are capable of both conceptualisation of what the gadget needs to do, through to the full manufacture of it.
I'm really excited about the rest of the presentation, but the first mind blowing moment was learning that Ball does aerospace engineering. I feel like I need to go learn more about that company's history now.
It would be well worth your time. I worked at Ball for 21 years in Aerospace as a technician. Ball has built a multitude of satellites. They built the COSTAR instrument (eyeglasses) for the Hubble to save it's focus. I certified the cleanrooms/tents that the actuators were attached to the mirror segments (I walked among them) in. After studying Ball, you will never look at a Ball jar or a pop can the same. A lot of the engineers are PHD's.
I was thinking about you in relation to these actuators just the other day! I was reminded of that open source microscope platform that you showed off; similar in that they both move very small amounts, i guess
Using the backlash to do the fine adjustment is pure genius. True out-of-the-box thinking. Normally when trying to achieve high precision, designers try to minimize backlash, not maximize it.
As far as I know, such "a large input motion results in a small output motion" is called the "motion de-amplification effect", there are several examples, "nanoconverter" is one of the earliest designs, also based on flexure hinge.
Minor correction to a small point, the Ball company the does the food packaging stuff like mason jars isn’t actually the same company as the Ball Aerospace company. The packaging division was spun off as its own company which was eventually sold to Newell Brands (who own stuff like Sharpie, Coleman, Rubbermaid, etc). The Ball Aerospace company is its own thing.
Hey this is amazing. Would you be open to doing an informal/educational clubhouse room on Small Steps & Giant Leaps with one of the JWST optics engineers themselves?
Oh my goodness, maybe? I'm not sure I'm qualified to do something like that to be honest! 😅 I'm just some hack in his garage playing around with things he doesn't really understand haha. Shoot me an email (info@breakingtaps.com) and we can chat some more :)
I would think it would be amazing to also have Chris Robinson (posted above) in on this given his work with the Hobby-Eberly. The three of them on one podcast would be very interesting.
@@BreakingTaps Don't worry too much, most of us at SSGL are just unemployed spacenerds haha. Many retired people from NASA and the space industry like to drop by because we ask very technical questions and they get to talk about what problems they've had to go solve instead of worrying about politics.
This was designed by the guy that designed the Antikythera mechanism 2000 years ago. Ball has him stored in a closet where he is in suspended animation. When Ball has a special problem like this they thaw him out and he designs it. It's amazing that they've been able to keep this secret for so long.
Amazing work! I wish NASA did such detailed deep dives into various aspects of the JWST or the Mars rovers. They make plenty of content for the general population and STEM students, but you really have to go digging if you want to find these technical papers with all the juicy details.
Ball Aerospace...I recognized the font when you showed their logo. A quick search confirmed that they ARE, in fact, a subsidiary of the company that has made home canning jars for your Grandma's green beans for about a century or so....Wow, way to diversify!....That's really something!
Thank you for that. I have been trying to think for a while now how you could get actuators with such fine adjustments as 2 nanometres with motors and gears. Now I know, with cams and flexures. Simple once you've been told. Genius if you have to invent it! Can't wait to see the JWT pictures. I think we will be amazed at what we will see.
Glad to see flexures in industry. Kind of reminds of how early micrometers worked. The fine/coarse adjustment is on the border between stupid and genius, but it makes sense given the particulars of the application.
I love a lot of rocket engineering for this. You'd think they are these hyper complex engineering nightmares, but they much prefer as simple as possible. And its all these stupid simple, reliable mechanisms that come together to launch things off the planet.
I am an electronics engineer, and the engineers designed that are all geniuses! (Considering the extreme requirements), and the programming involved controlling all those motors. pretty sure they have some sort of feedback to tell the system the exact location and position of all those motors and mirrors. There are lots of things involved in designing such actuators, like materials being use, (mechanical / thermal / physical, characteristics). Torque of those motors, gears reductions, etc. Electrical, feedback and control, etc. This is just amazing! Like WOW!
Great content, as always! I’ve been very curious about those actuators for some time, but hadn’t dug in very deep. Thank you for digging up the info, and applying it so we all can have a better understanding and maybe apply it forward. I’ve often thought if we can make such actuators simple enough, they could be made massively parallel to turn large optical flats into focusing mirrors, with the potential to make truly enormous astronomical optics.
You would have to bend the mirror far too much and an optical flat is even more effort to produce than a parabolic mirror. Segments might of course be even more difficult to make&test.
Amazing video! Yesterday I found the paper about these actuators and I thought it would be great if someone demonstrated how they worked. As an upgrade for the coarse positioning screw I would suggest a dry leadscrew. The igus dryspin leadscrews for example should work a lot better than a standard M8 thread.
Will take a look, thanks! I was hunting around for a small leadscrew but couldn't find something appropriate, and didn't feel like taking an angle grinder to one of the larger ones I had sitting around. I'll check out what Igus has for future projects!
I am an engineer myself and I asked myself: How could they move these mirrors so accurate? Thank you very much for your very good work. Best regards from Germany
Any idea of the MTBF of that flexing mechanism? I'm sure they did their calculations and that thing should last for a long time, but I always get worried about fatigue failures when I see designs like that.
I didn't see any data on the flexure itself, but I did find this paper [1] which summarized failure testing of the actuator design. The conclusion was that bearings in the gearmotor always died before any other component in the actuator. It seems each motor has been allocated a certain number of revolutions (60,000) based on failure testing and statistical extrapolation from there. The older paper I based the design on mentioned a similar thing ("The gearmotor has been noted as a life limited item and motor revolutions must be recorded.") but didn't go into details. [1] www.spiedigitallibrary.org/conference-proceedings-of-spie/8442/84422I/Actuator-usage-and-fault-tolerance-of-the-James-Webb-Space/10.1117/12.924596.short?SSO=1 [1b] Scihub link: sci-hub.se/10.1117/12.924596
Another straightforward demonstration of the mechanical principles used in this motor is the surface forces apparatus, or SFA, designed by Jacob Israelachvili. The SFA uses a shaft attached to a stepper to push a weak helical spring against a stiff double-cantilever spring. The flexure of the double-cantilever springo produces step sizes of a couple angstroms. Looking at a diagram of the displacement mechanisms for the old Mk 2 SFA design, it is fascinating how clear it is that a 2-spring system can be used almost exactly like gear ratios. Also, spring-based systems are great for minimizing backlash... don't ask me why though
This mechanism actually looks much simpler than I would have imagined. I thought these mechanisms would consist of blocks of massive machined titanium with nanometer precision wheels and screws all over the place, made out of some funky ceramics with diamond ball bearings and radioactive batteries
I've been wondering how they can possibly be so precise since I heard about the calibration procedure. This is super cool! And now my questions are answered
In 1972, Philips Consumer Electronics, in the Netherlands, developed and implemented a piezo electric actuator which changed the angle of a 6 micron ferrite video head moving at a scanning rate of approximately 64uS per seconds IN REAL TIME!!! Frankly, the mechanical actuators on the JWST are a very course and rough cousin compared to the ultra fine precision achieved by Philips in its Video 2000 dynamic track following system (DTS).
I had a calculus teacher named Ira Becker who worked at ball aerospace while Hubble was being developed. He made several contributions. I can’t say he was the best teacher, but I have the upmost respect for his contributions to science.
This is amazing, crazy cool tech and great job recreating it! And an early congrats on the 1M+ views this video will no doubt have within in a few weeks :D
For getting multiple finished products print a set of parts get them right, then silicon mold the working parts. I have used Alumilite in the past, it's easy to work with and decently strong. For higher strength light weight parts print molds and use the forged carbon fiber method.
Love the model, the voiceover while talking through the functionality as it happened was extremely helpful in sorting out the fine/course stages and how they interact.
Ingenious! Thanks for the super-clear explanation - I'd always wondered how the hell they got such crazy precision - and now I know! I particularly like the idea of the flexure because the joints have no slop or wobble.
I've been wondering this whole time how they have such a reliable way to make these fine adjustments. It's still mindbogglingly beyond my know-how but I love this demonstration. Nothing makes me happier than simple tools for a complex job. The engineers at ball have to be damn proud of this machine. I'm really loving this channel - thanks for sharing!
That is an absolutely INCREDIBLE solution to having coarse and fine movement with a single driver. Absolutely beautiful. Wonderful work, and even better presentation!
What a fantastic video. I love this kind of practical design where you read about something and then go about reverse engineering it in CAD. I do this all the time and it really helps me understand how something works and even why some design choices were made.
Finally! I had been wondering what the mechanism for mirror adjustment was, and even tried searching for it several times. ....might actually try the whole assembly + electronics.
Similar but distinct terminology appears in aviation and marine use. It's worth noting these terms in order to make the similarities familiar. The three rotational degrees of freedom are pitch, roll, and yaw. The less familiar translational parameters are heave (up-down), surge (fore-aft), and sway (port-starboard). All of these, at sea, tend to be periodic due to wave activity, at much less than 1Hz. I have the impression that the periodicity of motion in devices in space occurs at acoustic frequencies, where there can be material resonance and any number of component issues. At sea, the most hazardous conditions are when the wavelength of sea motion is close to the vessel length. It's the sort of resonance that can break a ship in half after a few iterations.
I don't usually comment on videos but I have to say I'm super impressed with your explanation; in-depth but not over the top and also very clear. Great work!
Working on the mechatronics/optomechanics behind semiconductor manufacturing optics (EUV etc.) at my day job, the first thing i was interested in about the JWST how the actuation is designed. I found the paper aswell and found this beautiful mechanism explained. Your take and model is very nicely done! Thank you for that. We have a lot of (beautiful and complex) flexures aswell, which sadly very few people probably will ever see because of NDAs...
What amazed me is the company that makes aluminum cans and mason jars.... Also does stuff in space. Wow, who would have known. Great video btw. These model exploitations are super interesting.
I loved your video. Like you, I think the actuators are among the more fascinating pieces of tech on JWST. I stumbled upon the Warden paper some months ago (and it's still on my computer desktop, lol), and I became convinced after studying it that while the coarse stage is being driven, the fine stage actually repeatedly "oscillates" through its range, as the fine stage cam is always driven. I thought I must be wrong about this, but your video has finally proven it to me! This seemed odd to me until I delved into the science of flexures and how the material choice and design of these interesting actuators avoids mechanical fatigue. Having the fine stage always driven is an elegant solution that reduces the complexity dramatically. The Rube Goldberg in my brain would've designed a way to disengage the fine stage while the coarse stage was moving. And of course, that would be wrong! Many kudos to your ingenuity and creativity in making this working model!
Followed JWST since Christmas (thrilled enough to link on my business website home). Outside of the instrument and project itself, I find this is the most useful video I found to date, to explain the mechanics behind mirror alignment. Brilliantly presented and an inspiration. Thank you for taking the time to do that.
Thank you for this mirror actuator video. The local New said the fine adjustment came first, then course adjustment. They couldn’t understand and I have been trying my best how or why fine adjustments would ever come first. The fine stage was a lot easier to drive first for testing. Thanks.
I am excited to see you excited. Thanks! The heritage of those Hexapods and the actuators are also going to be used for the focal plane of Nancy Grace Roman Space Telescope (NASA’s next gen large telescope) . Ball is starting to test it
If I remember correctly, there was a type of rotational password lock often seen in old movies where you enter the three digits by rotating clockwise, then counterclockwise, and then clockwise again. The internal structure works exactly the same way, not sure if they took inspiration from that.
I've been wondering how these actuators work, because they're amazingly precise and most content about JWST never gives you a good look at them. Thanks for the excellent breakdown!
Woah, that's awesome! Do you happen to remember what metal(s) the bits were made from? The paper mentioned titanium being used for the gears, and a few other vague references to titanium, but nothing definitive for the main frame and flexures. I sorta assumed Ti for those as well but couldn't track it down
Amazing video, plenty of detail and a model to see the mechanism more clearly, the actuator appeal to simplicity to address a complex problem, really cool