Spot quiz! Who can work out the mechanism that allows this thing to move about? (apart from you @Dale_Dunn )
http://www.youtube.com/watch?v=B0-Qv0-Y8Fw
Does that mean I already figured it out?
Dave posted a tinkercad model explaining it on the thread :). I expected you to have worked it out anyway tbh.
Is this something more than just a cable-driven H-bot? The cable layout looks a little odd?
Oh hangon.
The Cross-bar is shifted along in a fairly conventional manner.
The carriage running across the cross-bar is driven by a cable running in a sort of dual-looped fashion (fixed at two ends) that means that movement of the carriage can be induced by an off-crossbar-motor. The motion of the crossbar carriage is entirely isolated from the motion of the crossbar itself.
This eliminates the need for H-Bot kinematics. Really nice! I might have to steal this.
(Edited to remove my silly zig-zag theory.)
and @Sanjay_Mortimer gets the cuddly toy! I’d drawn the same conclusion as you’d have to have a closed loop for the carriage movement if you were going to pass the cable across the spool like that and the only way to do that would be to have a figure of 8 cable loop with the carriage in the middle.
Tinkercad link here: https://tinkercad.com/things/k6YSG0TWwac-cablebot-y-axis He’s not finished it but you get the idea.
The guy who came up with it is @dmoorhou or ‘acrux’ on the reprap forums.
I think i’m a bit in love with this simple but excellent cable arrangement.
All the positives of an H-bot system, but none of the drawbacks I can see?
Beats the ultimaker type system in my eyes too, because you’re getting rid of that rotating rod/bushing issue. Now you can use a whole array of linear motion systems - like that awesome hiwin rail in the video.
Is there a catch?
Shit, no I take that back. It’s not a figure of 8 on the carriage it’s a fixed end system, the linear rail is the closed loop: I was trying work out why it doubles back like it does, but it’s because you need to have the top cable attached to the rail on one side and the bottom cable fixed on the other side. It’s very pulley intensive!
Yeah the number of pulleys needed is quite high.
- 8 Idlers, 1 Drive to run the crossbar carriage.
- 6 Idlers, 1 Drive to drive the crossbar itself.
But bearings are cheap. There would be quite a lot of belt used too, which could be expensive in GT2 or similar.
If your bed is a n-metres*n-metres square, you can do a standard mendel arrangment in a “perfectly belt efficient arrangement” in 4n of belt. (Across the bed and back again, in X, same in Y) -
Big-B notation: B(4n)
In this, you need 5n for the carriage drive, and 6n for the bar drive.
B(11n)
Now i’m going to need to find some synchromesh in the UK…
@Sanjay_Mortimer Dave says he’s having some binding issues due to parts being too tight but seems positive enough about it. Posts here: http://forums.reprap.org/read.php?279,222917,page=4
It is rather clever.
Very clever. I want to build something like this. The Z axis is probably more conventional.
@Sanjay_Mortimer SDP-SI do good overseas shipping rates IME. They sell so much useful stuff too! https://sdp-si.com/eStore/Catalog/Group/768. GT2 belt is cheap from china! http://www.aliexpress.com/wholesale?SearchText=GT2+belt&catId=0&initiative_id=SB_20131009163146
Is it an implementation of this?
@Billy Yes it is! 
@Billy good eye! This is quite an elegant design… But not to be a Debbie downer, but I have some concerns. In the link Billy posted, the x axis looks great, but the y axis looks like it might suggest from the same racing as the standard h-bot. Since the 2 ends of the wire are fixed to the stationary frame, when moving the y, one side of the belt is in tension while the other is slacked. If there is flexing in the belt, this can lead to chattering movement. The beauty of the core xy is that both belt loops are continuous. The toil holder completes the structural loop of each loop, so the entire belt is at a constant tension. Perhaps this won’t be an issue is done correctly, as many have gotten traditional h-bots to work fine. Regardless, very cool design!
Happened to have visio open so I thought I’s sketch out the mechanism in 2d. https://lh5.googleusercontent.com/-753_E-QeVmo/UlYKlAfRqGI/AAAAAAAAIEg/Ki-oSosqbEU/w471-h514-no/Unnamed+mechanism.jpg
why not use timing belt?
Yes Tim that is correct just been looking at your diagram. I have yet to design or build a z axis so don’t really know what the issues will be yet, until I start printing. The binding is very minor as the brackets at the ends of the moving rail are not super tight and the cable pulls on the rail not the brackets. It does look kewl moving, I have pu a few marks on the cables with marker pen, figuring out what’s happening by watching the cables does your head in
No-one yet has commented on the truss design, this is so strong and very movement, I used it often in large telescope designs in the past. Just print eight brackets, cut a few bits of 12mm tube and bolt it to any cheap boards top and bottom, then cut a hole in the top board!
Dave aka. Acrux
I could have used a timing belt, one axis needs about 1 meter the other about 1.5 M the linear rails are 400mm to give an idea of scale. However the pulleys would have to be a bit bigger these are only 13mm OD think that is to tight for a belt? Also the cable I have running ummm vertically I would need very different placements to use a belt as the cable turns corners both vertically and horizontally in this version.
I loved the Tantillus design, using cables, XY movement of the head with all motors not moving is great. This just an alternative idea that I was keen to try out, no idea if it will be great or a flop yet