This is purely equipment porn and never likely to be in the hands of mere mortal macro photographers, but it's still something to drool over...
Video:
https://www.youtube.com/watch?v=B756Novxejo
Some interesting numbers:
Travel ranges to 34 mm / 42°
Load capacity to 5 kg
Actuator resolution 40 nm
Min. incremental motion to 0.2 µm
Repeatability to ±0.1 µm
It can also move at 10mm/s which has to be good for something.
Goniometers are so 2014
Moderators: rjlittlefield, ChrisR, Chris S., Pau
-
MaxRockbin
- Posts: 188
- Joined: Sat Nov 08, 2014 11:12 pm
- Location: Portland, OR
Goniometers are so 2014
If your pictures aren't good enough, you're not close enough. - Robert Capa
I've been looking at this sort of platform (quite a few varieties exist) for several years. Have been itching to build one for the Bratcam, and program an Arduino to run it. But there are a bunch of things higher on the to do list.
As you say, they are expensive--at least, all the ones I've priced are. This is a large part of why I'd like to build one and launch a how-to into the public domain. I don't actually think it would be all that difficult.
--Chris
As you say, they are expensive--at least, all the ones I've priced are. This is a large part of why I'd like to build one and launch a how-to into the public domain. I don't actually think it would be all that difficult.
--Chris
-
MaxRockbin
- Posts: 188
- Joined: Sat Nov 08, 2014 11:12 pm
- Location: Portland, OR
In a way, the PI platform is a two-fer. Given the accuracy, it could be used for stacking (horizontal or vertical), as well as making very versatile stage. Assuming your subject doesn't mind going for a ride.
Plus, would you rather tell people that you mount your subjects on a goniometer, which sounds like something left over from a pelvic exam, or on a shiny new robot?
Plus, would you rather tell people that you mount your subjects on a goniometer, which sounds like something left over from a pelvic exam, or on a shiny new robot?
If your pictures aren't good enough, you're not close enough. - Robert Capa
-
RobertOToole
- Posts: 2627
- Joined: Thu Jan 17, 2013 9:34 pm
- Location: United States
- Contact:
-
MaxRockbin
- Posts: 188
- Joined: Sat Nov 08, 2014 11:12 pm
- Location: Portland, OR
Pretty neat!
This delta design though seems kind of limited on the rotations (the hexapod can rotate on 3 axes as well as x,y,z translations).
I did see a design for a do-it-yourself hexapod, but it didn't seem nearly precise enough for macro work. Only $150 in parts though. Companies like this one sell some interesting parts for do-it-yourself roboticists for cheap:
https://www.pololu.com/product/88
Also there's the issue of the driver. There are what are called "G Code" parsers out there (a language for industrial motion control), but I think that's really on a way to specify a path. You still need to tell the hardware what to do with the info.
It's kind of interesting to ponder - especially when all legs can lengthen/contract independently like with the hexapod - exactly how you would program how much to adjust each leg to move and rotate the platform in a specific way. My high school trigonometry book had a final chapter on spherical trigonometry and the only application suggested was celestial navigation. Just hasn't come up a lot for me. But I think that's exactly what you'd need for one of these gizmos....
This delta design though seems kind of limited on the rotations (the hexapod can rotate on 3 axes as well as x,y,z translations).
I did see a design for a do-it-yourself hexapod, but it didn't seem nearly precise enough for macro work. Only $150 in parts though. Companies like this one sell some interesting parts for do-it-yourself roboticists for cheap:
https://www.pololu.com/product/88
Also there's the issue of the driver. There are what are called "G Code" parsers out there (a language for industrial motion control), but I think that's really on a way to specify a path. You still need to tell the hardware what to do with the info.
It's kind of interesting to ponder - especially when all legs can lengthen/contract independently like with the hexapod - exactly how you would program how much to adjust each leg to move and rotate the platform in a specific way. My high school trigonometry book had a final chapter on spherical trigonometry and the only application suggested was celestial navigation. Just hasn't come up a lot for me. But I think that's exactly what you'd need for one of these gizmos....
If your pictures aren't good enough, you're not close enough. - Robert Capa
-
rjlittlefield
- Site Admin
- Posts: 24055
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Interesting devices!
These devices are simpler than that because all the parts are either straight or planar. So, you can start by figuring out where you want the top platform to be, using your favorite parameterization. From that, use ordinary trig to figure out the 3D coordinates of the centers of the ball joints. Then use the Pythagorean theorem in 3D to figure out the actuator lengths.
With a hexapod, control should be especially simple because there's no redundancy. Within some envelope, any combination of actuator lengths is valid. You don't have to worry about breaking anything, even if you adjust the actuators one at a time.
With the Delta Robot v1 that g4lab pointed to, things seem much more difficult. It looks to me like in that design, with 8 actuators, you would have to move multiple actuators in careful synchronization to avoid stressing the joints.
It's sort of like the difference between adjusting a 4-legged chair and a 3-legged stool, while keeping all legs in contact with the floor.
--Rik
I guess it depends on what you mean by "spherical trigonometry". As defined by Wikipedia, "Spherical trigonometry is the branch of spherical geometry that deals with the relationships between trigonometric functions of the sides and angles of the spherical polygons (especially spherical triangles) defined by a number of intersecting great circles on the sphere." In other words, it relates to what happens when the edges of triangles are forced to lie on the surface of a sphere. Lots of interesting things happen there, like the sum of vertex angles is always greater than 180 degrees.It's kind of interesting to ponder - especially when all legs can lengthen/contract independently like with the hexapod - exactly how you would program how much to adjust each leg to move and rotate the platform in a specific way. My high school trigonometry book had a final chapter on spherical trigonometry and the only application suggested was celestial navigation. Just hasn't come up a lot for me. But I think that's exactly what you'd need for one of these gizmos....
These devices are simpler than that because all the parts are either straight or planar. So, you can start by figuring out where you want the top platform to be, using your favorite parameterization. From that, use ordinary trig to figure out the 3D coordinates of the centers of the ball joints. Then use the Pythagorean theorem in 3D to figure out the actuator lengths.
With a hexapod, control should be especially simple because there's no redundancy. Within some envelope, any combination of actuator lengths is valid. You don't have to worry about breaking anything, even if you adjust the actuators one at a time.
With the Delta Robot v1 that g4lab pointed to, things seem much more difficult. It looks to me like in that design, with 8 actuators, you would have to move multiple actuators in careful synchronization to avoid stressing the joints.
It's sort of like the difference between adjusting a 4-legged chair and a 3-legged stool, while keeping all legs in contact with the floor.
--Rik
-
MaxRockbin
- Posts: 188
- Joined: Sat Nov 08, 2014 11:12 pm
- Location: Portland, OR
Good point about spherical trigonometry. There is no need to calculate the angles on the surface of a sphere or arc lengths.
Just deltas of x,y,z for the vertices after rotation in 3 axes (which corresponds to a 3 new positions on a great circle centered at the center of the platform) and deltas for simple xyz translations. So you can use regular Cartesian/Pythagorian/Euclidean distance and piece things together - with a little plane trigonometry.
I looked up the Delta Platform on Wikipedia:
https://en.wikipedia.org/wiki/Delta_robot and the object of it seems to be to maintain the orientation of the platform while it moves around. It doesn't rotate.
So that reduces binding issues. They contrast it with the Stewart platform - aka hexapod.
But, unfortunately, not too exciting for use as a macro platform.
Just deltas of x,y,z for the vertices after rotation in 3 axes (which corresponds to a 3 new positions on a great circle centered at the center of the platform) and deltas for simple xyz translations. So you can use regular Cartesian/Pythagorian/Euclidean distance and piece things together - with a little plane trigonometry.
I looked up the Delta Platform on Wikipedia:
https://en.wikipedia.org/wiki/Delta_robot and the object of it seems to be to maintain the orientation of the platform while it moves around. It doesn't rotate.
So that reduces binding issues. They contrast it with the Stewart platform - aka hexapod.
But, unfortunately, not too exciting for use as a macro platform.
If your pictures aren't good enough, you're not close enough. - Robert Capa
-
rjlittlefield
- Site Admin
- Posts: 24055
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Resurrecting this thread from last year...
I noticed this morning that the creator of the Delta Robot linked earlier by g4lab is now selling a complete kit for a positioning device described as follows:
--Rik
I noticed this morning that the creator of the Delta Robot linked earlier by g4lab is now selling a complete kit for a positioning device described as follows:
See https://www.marginallyclever.com/produc ... latform-v2 for more information.Rotary Stewart Platform v2
CAD $597.92
This is a six degree of freedom motion control platform. It can heave and tilt (up/down), sway and roll (left/right), as well as surge and pitch (forward/back). This is a scale model of the machine found under flight simulators, multi-axis milling machines...
--Rik
And if you want to build one, check this link: http://www.instructables.com/id/Arduino ... -Platform/
I would probably build it more like my 3d printer with magnetic ball joints and carbon fiber arms.
I would probably build it more like my 3d printer with magnetic ball joints and carbon fiber arms.