09-17-2014, 10:53 AM
Hi all, this is just a thought I want to throw out, trying to find ways around a requirement for something I'm planning to build (at the Crap-O-CAD stage at the moment)...
I'd like to build a linear actuator but with some constraints - it needs to be compact, have about a 3/4" linear movement and do it with about 90-degrees of shaft rotation - yes, that's about a 3" lead if it were threads! Also, I want to make it "bidirectional", so that it can be actuated by the shaft rotation, or the shaft can be rotated by linear action. I guess that for this I'd ideally want a helix angle close to 45 degrees?
A regular multi-start thread would work, but the 3" pitch is going to be hard to machine conventionally by screwcutting (my lathe only goes down to 2tpi), and I'm pretty convinced that friction is going to make it a non-starter. I could thread mill it, by messing with change gears and mounting a milling spindle on the carriage (at a suitable angle for the thread), but I'd still be worried by the friction in the actuator...
An alternative that I thought of was a short ballscrew, minimal friction, multi-start and it should be able to accommodate the needed pitch and helix angle, but I've not found any online so far, which brings me to a Daft Idea:
If I were to thread mill the shaft, using for example a 1/4" endmill tangential to the shaft at 45 degrees, I could make the shaft fairly easily (for certain values of easy involving indexing the starts, fiddling with change gears) with half-circular ball tracks;
a number of precision rods the same diameter as the intended bearing balls could be tightly wrapped around the shaft in the ball tracks, ends turned out at 90 degrees to mimic the path of the recirculating balls and brought out through an outer shell;
the ends of the shell could be blanked off (with the shaft passing through at each end) and a Suitable Plastic or resin poured/injected in to fill the remaining space;
the shaft could then be unscrewed, the rods removed carefully into the space formerly occupied by it, and I should be left with a ball-nut, just needing balls and return channels to be added to complete it.
I've left out things like boring (relieving) the nut for clearance, mould release, choice of plastic/resin (DIY Moglice-alike?), but so far I can't see any impossibilities in the process - so I must have missed something! This doesn't need leadscrew accuracy, it's simply an actuator moving the end of a lever through 3/4" or so remotely via the shaft (and the shaft has to rotate, I can't just use a pushrod, I need axial movement to do something else!), repeatability matters but it doesn't need to be to better than maybe a 16th or 32nd of an inch.
OK, what have I missed?
I'd like to build a linear actuator but with some constraints - it needs to be compact, have about a 3/4" linear movement and do it with about 90-degrees of shaft rotation - yes, that's about a 3" lead if it were threads! Also, I want to make it "bidirectional", so that it can be actuated by the shaft rotation, or the shaft can be rotated by linear action. I guess that for this I'd ideally want a helix angle close to 45 degrees?
A regular multi-start thread would work, but the 3" pitch is going to be hard to machine conventionally by screwcutting (my lathe only goes down to 2tpi), and I'm pretty convinced that friction is going to make it a non-starter. I could thread mill it, by messing with change gears and mounting a milling spindle on the carriage (at a suitable angle for the thread), but I'd still be worried by the friction in the actuator...
An alternative that I thought of was a short ballscrew, minimal friction, multi-start and it should be able to accommodate the needed pitch and helix angle, but I've not found any online so far, which brings me to a Daft Idea:
If I were to thread mill the shaft, using for example a 1/4" endmill tangential to the shaft at 45 degrees, I could make the shaft fairly easily (for certain values of easy involving indexing the starts, fiddling with change gears) with half-circular ball tracks;
a number of precision rods the same diameter as the intended bearing balls could be tightly wrapped around the shaft in the ball tracks, ends turned out at 90 degrees to mimic the path of the recirculating balls and brought out through an outer shell;
the ends of the shell could be blanked off (with the shaft passing through at each end) and a Suitable Plastic or resin poured/injected in to fill the remaining space;
the shaft could then be unscrewed, the rods removed carefully into the space formerly occupied by it, and I should be left with a ball-nut, just needing balls and return channels to be added to complete it.
I've left out things like boring (relieving) the nut for clearance, mould release, choice of plastic/resin (DIY Moglice-alike?), but so far I can't see any impossibilities in the process - so I must have missed something! This doesn't need leadscrew accuracy, it's simply an actuator moving the end of a lever through 3/4" or so remotely via the shaft (and the shaft has to rotate, I can't just use a pushrod, I need axial movement to do something else!), repeatability matters but it doesn't need to be to better than maybe a 16th or 32nd of an inch.
OK, what have I missed?
Rules are for the obedience of fools, and the guidance of wise men...
(Douglas Bader)
(Douglas Bader)