Linear Actuators

[Robotman]

Well, after getting a politely worded semi-kicking from both Bones and Dingo (no hard feelings!) in my "Starting out" post, I followed Bones' link to www.xrobots.co.uk/modules/news/ and discovered that the guy there has built his own linear actuators.

Now, these look good. And Bones is too right about how things come along just when you need them. Cordless screwdriver (motor + battery) is available in Dick Smith for $9.98 a pop (cat. T-4831). Cheaper than buying motors separately - and you get a battery too! And that's similar to what this guy at the website used in his actuators.

I'm also thinking to connect all the "free" 3V6 batteries in parallel to provide power for all the actuators, plus use Altronics' DC-DC convertor kit ($9.95, cat. K-6330) for stepping up to +5V for logic, and bingo, its time to have a look at just how deep that deep end is

And now, my question to the viewers is: Has anyone out there used/built linear actuators? How fast/slow are they? Yep, that's very general - so, what extension length? - what RPM? - what screw thickness? Should they be "series elastic" (spring dampened), or is this overkill? Unforeseen pitfalls? Maybe I should just go and build one, test it, and report back? Yes. That's a very good idea.

Until then...

[Bones]

Hi,
No affence taken.
See what I mean by thing just appear when you need them. The little screw drivers are great and cheap. Having a speed controller and a high speed motor so you can adjust the rate of travel. Let us know how you go.
Bones

[Robotman]

The preliminary results are in:

First off, the engineering effort was a lot for someone with limited tools and almost no experience in metal work. I worked mostly from the article given at the above URL, using just about the same techniques.

I purchased 1/4" rod, cut 4 lengths, and threaded them. Using two mudguard (big) washers, a long 8mm bolt with nut, and a small piece of 5mm plyboard I proceeded to make the "traveling nut". A third mudguard washer and I had the "end-cap" that secured the "running rods" that the "traveling nut" with its rods would slide along. An aluminium bracket was fashioned to hold the motor and secure the other ends of the "running rods". The overall result is ugly and required filing extra clearances in holes because I wasn't precise enough. This is low tolerance engineering going on here!

The extension (travel) length is around 50mm and the unextended length (including motor in bracket) is around 280mm. I won't give correct dimensions because this is all prototype stuff and it doesn't really matter if it is big or small - the lengths are just what I ended up with after all of my cutting, filing, bending and threading. The weight is unknown, but the rods slide smoothly enough and anyways, the practical weight by feel of hand is fairly low.

I discovered that the Tricky Dicky (that's DSE for those much younger than me!) screwdriver has an RS380S motor in it. That can run at 7V2 nominal, but they are running it at 3V6. The actuator motion is slow at this lower voltage. I tested it a 5V and the thing really started to whir. At 7V2 it should literally belt along, but I'm concerned about the plastic planetary gearbox that is attached to the motor (the one that comes with the screwdriver). It might strip itself internally, even though there isn't much output load (weight).

Still, I think the overall motion speed of the actuator is insufficient for long travel lengths, and so any robot joint using the actuator will have to be designed to translate a small distance into a large angle. This will also mean a fine thread to allow for relatively precision positioning - although this shouldn't matter in a "dynamic learning feedback system" for the robot in which it learns the limits of its own body The other issue is how to secure the threaded bolt from slipping out of the hex shaft on the output of the gearbox. I might have to drill a hole for a grub screw (more threading )), or simply drill a hole right through and insert a cotter pin or one of those springy clip type thingies (name?).

I'm going to develop another linear actuator but this time using smaller sized hardware (e.g. brass rods and tubing, smaller gauge aluminium, etc.) with finer screw thread pitch, lighter overall weight, and then any weighty joints should be supported using "axles" and other such things so the load stress on the actuator is minimal.

I took some pics of this first prototype, but I think it would be best to try the smaller version and then write up a construction article - unless I find the whole exercise just won't work for walking speed joint movement control and I end up reverting to servo motors! Of course, I could try building some air muscles, but Bones is doing this - aren't you? - and he should have a nice report for us all very soon! ;D

Cheers!

[Bones]

Nice work on the actuator. Can't wait to see some pics.
And LOL yes a report is coming on the air muscles.
And video etc.
It's on hold for a little bit. I'm building a table mill at the moment. www.rcmodels.net/cnc/hobby/classic.html
I you look at the next link how he does the Z axis (second pic down)
He pre tensions his nuts to stop play would be good for your actuator.
Using a stepper motor out of a printer would be good for the motor drive too.
www.rcmodels.net/cnc/hobby/setup/adjust.html
We got Acme threaded rod and teflon nuts for the mill.
Acme is a square thread so its not trying to expand the nut while your trying to turn it like a standard thread does.
You could use ball screw. It's made for this application but not cheap.
www.barnesballscrew.com/ball.htm

Bones