Every time I see something interesting, Applied Science has been there before. There needs to be a directory of videos Applied Science has commented at...
For anyone trying to replicate Dan Gelbart’s lathe, Pyramid Granite in Escondido California makes custom machine bases from granite, much like the pieces Dan Gelbart used in his lathe. Additionally, Granitan epoxy and stone composite are an excellent machine base as well.
Fantastic looking spindle would have loved to see it completed and running. Would it be possible to get links to the papers going through the tuning or impregnation process ?
Very nice progress...would it not be better to Grind the OD and ends of the shaft? That would result in a much tighter tolerances and better finish. Ya, almost $300 plus freight for that hunk of Cast Iron for the housing...need to order one for me now. Thanks for the fine Documentation and sharing this with the community!!!
The shaft is ground from the factory better than I ever could. The thrust faces will be ground and then lapped. Thanks for the support! Documenting this is the least I can do to give back to the community, it’s given me so much
Excelent work, It was 3 years ago but I want to ask you question about spindle building, I also making my air bearing spindle for presicion grinding machine but ist there any difference between using radial air bearing or can we just use air busing with thrust face? I will share my machine picture to give you an idea My regards :)
Such a cool project. Great to see so much progress! Have you considered maybe making an air bearing stage to go with it as a followup to this? Don't mean to create more work for you! But, it would be fun to watch. Kudos!
@@cylosgarage so cool! Im working on a linear motor, no air bearings for now though. Take a look at my prototype. I managed to integrate water cooling channels into it, but haven't perfected the process yet to clear them after casting epoxy. Take a look; designs are completely open source. Link to github is in the description: th-cam.com/video/bwdEMA3n0Z4/w-d-xo.html
@@primozlord awesome! i was just thinking about making my own the other week after seeing how expensive they are. Are yours iron core/permanent magnet? Those are the type I'd like because i can use the magnets to preload the bearing stage. Edit: i see they're not lol, no problem. Still very interested in your driving strategy though
@@cylosgarage no, it's an ironless motor. Two permanent magnet tracks with the forcer moving inside the magnetic field. An iron core design has the downside of introducing cogging, which you have to deal with somehow. On a side note; today I discovered that Philips patented a design which allows for 2D movement inside the magnetic field of an linear motor. Brilliant idea! (N-forcer) Edit: Since creating my own driver for this motor was not something I wanted to do, I purchased a bldc driver from Trinamic. With this driver I have the freedom to implement my own driving strategy. There are a few options to get a linear motor initialized. The most basic is to power the coils and let the motor settle into it's magnetic field. Alternatively you can perform an open loop movement towards a home sensor or something. In both cases you know from that moment that it's synched to the magnetic field and you can depend on the encoder afterwards. In the video I linked I enabled the motor and let the motor settle in, after which I tell it to zero it's position and perform some movements in a loop(currently all in position control). I'm still having a hard time getting the motor tuned correctly in very high speeds and acceleration.
Hi, really nice project and thanks for sharing. Hope you can help with one question. How are you going to lap the thrust face of the shaft flat and square? Thanks
Thanks! Once it’s mounted (heat shrunk/press fit) it will be turned between centers. This will be pretty much guarantee the face to be square, but not necessarily flat. It will then be rotary surface ground and lapped using a lap that is a perfect slip fit over the shaft
Have you considered using differential threads on your radial bearing adjustment screws? If you combine an 18 tpi lefthand thread with the existing 20 tpi screws, you would have effectively a 180 tpi adjustment. One turn at 20 tpi moves forward 0.050", while the 18 tpi moves 0.0555" back for a net movement backwards of 0.0055" or 180 tpi.
If I had a nickel ever time I got a comment suggesting differential threads I’d have at least a dollar or two lol. I considered it, yes, but it’s really just a convince thing. 20 TPI works and it’s not really worth the time designing and potential loss of rigidity just for a little increase in user friendliness. Especially since I’m the only one who has to live with my decision.
There is no way you're getting sub-micron backlash out of lead screws. I highly doubt the ways are even that accurate. Not to mention thermal expansion, rigidity and vibration, bearing load and tool deflection. A standard Aloris tool post is routinely repeatable to .0001" or less. Their new indexable version is Published at 0.00005" repeatability in their sales literature. Don't flatter yourself.
Oh my god how I messed with the head of a PhD. who develops a model for simulating roller bearings - he knows quite a bit about air bearings and I told him "Man, but what about those EUV litography machines, that operate under vacuum. How do you have those moving platforms with the waffers moving inside without physical material on material contact. If it touches, it creates particles, and you don't want to have particles inside a EUV machine, so you have to use some kind of an hydrostatic-hydrodynamic bearing :D He stoped talking for 3 minutes and then he said he needs to look at it :D
Wow I’ve never considered this. Many other semiconductor processing equipment operate under vacuum too. Wonder what kind of bearings they use? I see roller bearings in SEMs stages usually, although I have no idea how or even if they are lubricated.
@@andrewphillip8432 it moves with 15g acceleration :) th-cam.com/video/hQrtAJsJ-U0/w-d-xo.html .. so I guess magnetically levitated + linear motors for movement as they write on the website
I'm hoping to do this in the near future hopefully for a co-op owned maker space. Do you have a patreon I could support? Would you be open to letting me send an email with some questions?
@@cylosgarage Differential screws can be simplified if you hold the nut steady. To reduce complexity you could bore and internal/external thread a simple sleeve/bolt to the diameter of your brass locking nut. Internal thread to 1/2-20 and external thread with a slightly different pitch. Put some wrench flats on this sleeve. Your adjusting screw(B2) could stay the same but to adjust, hold B2 to avoid rotation and turn B1. Might be a little uglier on the outside of the spindle but it gets you fine adjustment. Your casting is plenty thick so the sleeve(B1) diameter can be large enough to single point thread with a tight fit to avoid rigidity problems. You could also lengthen B2 to extend out of B1 and add locking nuts to both but it would add some clutter. Enjoying your video and progress!
a few reasons, yes (in order of importance): 1. Thermal stability. Aluminum is far too sensitive to changes in temperature. A few degree temperature increase could cause the housing to grow so much that the air bearings are no longer adequately preloaded. Thermal stability is darn near the #1 priority in ultraprecision 2. Strength. Preloading the 4 bearings creates a fair amount of force that wants to push outwards on the housing and distort it. A distorted housing that is no longer round will not mate well with the surfaces on the spindle mount and hence be less rigid. I did the FEA and determined cast iron would distort far less than aluminum in this application. 3. Cast iron also has a unique vibration dampening property that aluminum does not. This is not of huge importance but is always nice when building machine tools. These three main reasons along a bunch of other smaller factors like cost and machinability led me to decide on cast iron.
Amazing work! How come you have the 3 flat faces on the kinematic mounts, rather than a tapered hole (cone shaped)? In my mind this provides more surface area for rigidity.
This is true, but I also was out to minimize static friction by minimizing contact area. The stiffness of the hardened ball on the hardened mount is adequate
@@cylosgarage I see, fair enough. On another tip, what are you hoping to turn on the lathe once finished? I'm guessing it won't be your average engine lathe jobs
@@lawmatenope, that would end disastrously. diamond turning works on almost all non ferrous metals and plastics such as acrylic. I plan to use nickel or silver plated base metals to make my mirrors. Not sure if I'll be making any lenses, if that's what you were thinking.
He also mentioned that he bought used surplus of "semiconductor machine bearing motors". What kind of machine did he get the air bearings from for $500? BTW your mic is really low, you need to either talk louder or increase your mic's volume before uploading to youtube.
Hi, phenomenal work so far, really enjoying watching the project develop. You mentioned a paper at 9:15. I'd love to know the name of it or if you've got a link to it that'd be even better. Thanks for the updates, keep up the good work!
@@cylosgarage Well, they say 'Art is never completed, only abandoned eventually', so I thought these kinds of projects benefits from other tinkerers. I might have asked because I wanted the design files for myself to try and build on it :P Anyways, do it when you feel comfortable with your designs, thanks to your videos, I am starting to build it for myself in the mean time :)
Wow, seeing the whole thing assembled at the end of the video is really impressive. Very nice work!
Thanks Ben, means a lot coming from you! And thanks for the original inspiration!
Every time I see something interesting, Applied Science has been there before. There needs to be a directory of videos Applied Science has commented at...
Its been 8 years since I came across Dan Gelbart's videos. Seeing your machine take shape is euphoric! Fantastic work, hats off to your perseverance!
For anyone trying to replicate Dan Gelbart’s lathe, Pyramid Granite in Escondido California makes custom machine bases from granite, much like the pieces Dan Gelbart used in his lathe. Additionally, Granitan epoxy and stone composite are an excellent machine base as well.
Beautiful work! cant wait to see more.
I gotchu
It's coming all together, amazed by the perfection of your work, keep it up !
Fantastic looking spindle would have loved to see it completed and running. Would it be possible to get links to the papers going through the tuning or impregnation process ?
Awesome project. Thanks for sharing!
Very nice progress...would it not be better to Grind the OD and ends of the shaft? That would result in a much tighter tolerances and better finish. Ya, almost $300 plus freight for that hunk of Cast Iron for the housing...need to order one for me now. Thanks for the fine Documentation and sharing this with the community!!!
The shaft is ground from the factory better than I ever could. The thrust faces will be ground and then lapped. Thanks for the support! Documenting this is the least I can do to give back to the community, it’s given me so much
Fantastic work!
Excelent work, It was 3 years ago but I want to ask you question about spindle building, I also making my air bearing spindle for presicion grinding machine but ist there any difference between using radial air bearing or can we just use air busing with thrust face? I will share my machine picture to give you an idea
My regards :)
yeah that Y12 plan paper is a gold mine look foward for yours impregnation video
honestly embarrassing how long it took for me to find it. completely missed it in laynemachinetech's description...
@@cylosgarage Adam needed to told me directly so I can go read it...
Are you going by plans from Dan Gelbart, or what’s going on? His work is fascinating.
Such a cool project. Great to see so much progress! Have you considered maybe making an air bearing stage to go with it as a followup to this? Don't mean to create more work for you! But, it would be fun to watch. Kudos!
absolutely. Mk2 is going to include linear motors and air bearing axes.
@@cylosgarage so cool! Im working on a linear motor, no air bearings for now though. Take a look at my prototype. I managed to integrate water cooling channels into it, but haven't perfected the process yet to clear them after casting epoxy. Take a look; designs are completely open source. Link to github is in the description: th-cam.com/video/bwdEMA3n0Z4/w-d-xo.html
@@primozlord awesome! i was just thinking about making my own the other week after seeing how expensive they are. Are yours iron core/permanent magnet? Those are the type I'd like because i can use the magnets to preload the bearing stage.
Edit: i see they're not lol, no problem. Still very interested in your driving strategy though
@@cylosgarage no, it's an ironless motor. Two permanent magnet tracks with the forcer moving inside the magnetic field. An iron core design has the downside of introducing cogging, which you have to deal with somehow.
On a side note; today I discovered that Philips patented a design which allows for 2D movement inside the magnetic field of an linear motor. Brilliant idea! (N-forcer)
Edit: Since creating my own driver for this motor was not something I wanted to do, I purchased a bldc driver from Trinamic. With this driver I have the freedom to implement my own driving strategy. There are a few options to get a linear motor initialized. The most basic is to power the coils and let the motor settle into it's magnetic field. Alternatively you can perform an open loop movement towards a home sensor or something. In both cases you know from that moment that it's synched to the magnetic field and you can depend on the encoder afterwards. In the video I linked I enabled the motor and let the motor settle in, after which I tell it to zero it's position and perform some movements in a loop(currently all in position control). I'm still having a hard time getting the motor tuned correctly in very high speeds and acceleration.
@@cylosgarage Cool. Looking forward to the next installment.
you may consider differential screws for adjustment
Hi, really nice project and thanks for sharing. Hope you can help with one question. How are you going to lap the thrust face of the shaft flat and square? Thanks
Thanks! Once it’s mounted (heat shrunk/press fit) it will be turned between centers. This will be pretty much guarantee the face to be square, but not necessarily flat. It will then be rotary surface ground and lapped using a lap that is a perfect slip fit over the shaft
@@cylosgarage Thanks! Looking forward to the next videos
Have you considered using differential threads on your radial bearing adjustment screws? If you combine an 18 tpi lefthand thread with the existing 20 tpi screws, you would have effectively a 180 tpi adjustment. One turn at 20 tpi moves forward 0.050", while the 18 tpi moves 0.0555" back for a net movement backwards of 0.0055" or 180 tpi.
If I had a nickel ever time I got a comment suggesting differential threads I’d have at least a dollar or two lol. I considered it, yes, but it’s really just a convince thing. 20 TPI works and it’s not really worth the time designing and potential loss of rigidity just for a little increase in user friendliness. Especially since I’m the only one who has to live with my decision.
There is no way you're getting sub-micron backlash out of lead screws. I highly doubt the ways are even that accurate. Not to mention thermal expansion, rigidity and vibration, bearing load and tool deflection.
A standard Aloris tool post is routinely repeatable to .0001" or less.
Their new indexable version is Published at 0.00005" repeatability in their sales literature. Don't flatter yourself.
Oh my god how I messed with the head of a PhD. who develops a model for simulating roller bearings - he knows quite a bit about air bearings and I told him "Man, but what about those EUV litography machines, that operate under vacuum. How do you have those moving platforms with the waffers moving inside without physical material on material contact. If it touches, it creates particles, and you don't want to have particles inside a EUV machine, so you have to use some kind of an hydrostatic-hydrodynamic bearing :D He stoped talking for 3 minutes and then he said he needs to look at it :D
Wow I’ve never considered this. Many other semiconductor processing equipment operate under vacuum too. Wonder what kind of bearings they use? I see roller bearings in SEMs stages usually, although I have no idea how or even if they are lubricated.
@@andrewphillip8432 it moves with 15g acceleration :) th-cam.com/video/hQrtAJsJ-U0/w-d-xo.html .. so I guess magnetically levitated + linear motors for movement as they write on the website
@@chronokoks Ahh that makes sense. That is quite an awesome feat of engineering
@@andrewphillip8432 It's fu**ing nuts really :) :) :)
Those cast iron artifacts look like it was poured over multiple times, if so, does it still have the true strength and properties as it should?
Who knows. That stock certainly left a bad taste in my mouth for Durabar
OK skipper how much to build me a gelbart ultra-precision granite bed lathe with cnc package and grinder??? 1 single micron
100k final offer 😁
I'm hoping to do this in the near future hopefully for a co-op owned maker space. Do you have a patreon I could support? Would you be open to letting me send an email with some questions?
No patreon but I appreciate it. Email is cyruslloyd4@gmail , shoot me a line
Fine thread is actually stronger than coarse thread typ.
why did you chose steel over aluminum for the ball-bearing housing?
Do you mean the spindle housing?
Have you considered using a differential screw (instead of really fine threads) for the adjusting screws?
Yep, decided against it due to complexity of implementation/potential rigidity problems
@@cylosgarage
Differential screws can be simplified if you hold the nut steady.
To reduce complexity you could bore and internal/external thread a simple sleeve/bolt to the diameter of your brass locking nut. Internal thread to 1/2-20 and external thread with a slightly different pitch. Put some wrench flats on this sleeve. Your adjusting screw(B2) could stay the same but to adjust, hold B2 to avoid rotation and turn B1. Might be a little uglier on the outside of the spindle but it gets you fine adjustment. Your casting is plenty thick so the sleeve(B1) diameter can be large enough to single point thread with a tight fit to avoid rigidity problems. You could also lengthen B2 to extend out of B1 and add locking nuts to both but it would add some clutter.
Enjoying your video and progress!
see different thread fine adjust
th-cam.com/video/7RC8WL2ngfA/w-d-xo.html
I’ve explained in multiple videos and countless comments why I’ve chosen not to use differential threads. I was aware of them going into the design
I just found this channel yesterday - too bad for me! But why cruddy graphite when BAM has published work stating its superiority?
BAM?
@@cylosgarage Doesn't Boron Aluminum Mag (+TiB2) sound like a great shorcut bearing material for electric-related machinery?
Super cool project. I also dream of making a "Gelbart" lathe but I doubt it will happen.
Very cool. Why do you have to make graphite segments instead of just boring a solid piece of graphite, if I may ask?
This is working out to be a very interesting project. Thanks for sharing it.
THAT'S SO COOOOL.
Is there a reason for using steel as the air bearing housing vs aluminium for example?
a few reasons, yes (in order of importance):
1. Thermal stability. Aluminum is far too sensitive to changes in temperature. A few degree temperature increase could cause the housing to grow so much that the air bearings are no longer adequately preloaded. Thermal stability is darn near the #1 priority in ultraprecision
2. Strength. Preloading the 4 bearings creates a fair amount of force that wants to push outwards on the housing and distort it. A distorted housing that is no longer round will not mate well with the surfaces on the spindle mount and hence be less rigid. I did the FEA and determined cast iron would distort far less than aluminum in this application.
3. Cast iron also has a unique vibration dampening property that aluminum does not. This is not of huge importance but is always nice when building machine tools.
These three main reasons along a bunch of other smaller factors like cost and machinability led me to decide on cast iron.
Amazing work! How come you have the 3 flat faces on the kinematic mounts, rather than a tapered hole (cone shaped)? In my mind this provides more surface area for rigidity.
This is true, but I also was out to minimize static friction by minimizing contact area. The stiffness of the hardened ball on the hardened mount is adequate
@@cylosgarage I see, fair enough. On another tip, what are you hoping to turn on the lathe once finished? I'm guessing it won't be your average engine lathe jobs
@@lawmate yea definitely not. Mostly optics and rocket injector elements. Hoping to make my own parabolic mirrors for building telescopes
@@cylosgarage cool, sounds nice. Can you turn glass directly with the diamond tooling?
@@lawmatenope, that would end disastrously. diamond turning works on almost all non ferrous metals and plastics such as acrylic. I plan to use nickel or silver plated base metals to make my mirrors. Not sure if I'll be making any lenses, if that's what you were thinking.
Offfff
He also mentioned that he bought used surplus of "semiconductor machine bearing motors". What kind of machine did he get the air bearings from for $500? BTW your mic is really low, you need to either talk louder or increase your mic's volume before uploading to youtube.
Literally search air bearing spindle on eBay and there’s a bunch for not too much money
Hi, phenomenal work so far, really enjoying watching the project develop. You mentioned a paper at 9:15. I'd love to know the name of it or if you've got a link to it that'd be even better. Thanks for the updates, keep up the good work!
digital.library.unt.edu/ark:/67531/metadc1055862/
This paper is like the Bible when it comes to making air bearings. Thanks for the support.
@@cylosgarage Thanks so much, really appreciate it!
Try putting the CAD designs up for download buddy, might inspire a few to build on top and improve upon the work :)
I don’t want to release anything until it’s all finalized and proven to work well. I still make tiny tweaks every now and again.
@@cylosgarage Well, they say 'Art is never completed, only abandoned eventually', so I thought these kinds of projects benefits from other tinkerers. I might have asked because I wanted the design files for myself to try and build on it :P Anyways, do it when you feel comfortable with your designs, thanks to your videos, I am starting to build it for myself in the mean time :)
Can you post a link to the air bearing tuning video?