cable drive

[der_fisherman]

You make some excellent points in your reply, many thanks.
Regards
Der Fisherman

[Mike Everman]

Accepting, then overcoming the problems seem to me to be the way forward, and understanding what the underlying reasons are for those problems might suggest the solution.

Well put, man. If I had a mantra, it would be that!

With a micro thread cut into the surface of the capstan matching the "lay" of the cable, you might even find you had solved the walking problem.

Very nice. I used to work for a small aerospace company, and we made a wire deployer for a spacecraft. The wire was on a drum with a thread that the wire lay in, and the drum itself was a nut that ran on a rod with the same thread pitch. So the effect was that the wire came out a hole in the housing and never wavered. It was walking off the drum thread at the same rate the drum was walking from one end of the housing to another. Seemed like magic.

You can have as many wraps as you want (though considering the poor stiffness of the cable, it doesn't need many), and a 75mm dia drum would only do 8.5 turns for 2m of travel. Now that I think about it, it's totally simple to make this recirculate, and to be anchored so it's not purely traction. whee!

[Mike Everman]

Twer I to do it, it would look like this. The possibilities are many with the design, once you accept the cables' stiffness. You can make several cable runs for really high stiffness, but using a screw thread as a bearing this way will only take so much side load.
The part I didn't feel like drawing is the return pulleys, single large wheel kicked at the proper angle, and the attachment of table to wire. Leave that to your imagination.

Also, there's a pesky feature that the driveshaft between wire drums moves axially a bit while you're trying to drive it.

[BillTodd]

Twer I to do it, it would look like this. The possibilities are many with the design, once you accept the cables' stiffness. You can make several cable runs for really high stiffness, but using a screw thread as a bearing this way will only take so much side load.

Another brilliantly simple Everman design

May I suggest using a roller nut at the stationary end.


When I tried a cable drive on a home brew plotter, the cable walking was not the problem (I arranged it so when the carriage was at the drum end the cable was square to it; Thus at the far end of the carriage travel the cable was at a only slight angle).

The problem was cable resonance, so Mike;

what do you think about pinching* your belt drive idea and using a moving capstan and let the cable rest on a flat (rubber?) surface. This would kill resonances and allow a tighter cable, thus fewer wraps around the capstan.

[* if you're going to steal an idea, steal a good one ]

[Mike Everman]

Hi Bill,
I think it's a grand idea. I'm not sure it follows that you'll be able to get higher tension. The tensile strength doesn't change, after all. Adding runs is the best bet there.

Guitar string modes will certainly be damped, but overall stiffness would not be improved unless you really push the tangent point down hard and get some traction contribution.

I think a roller screw would be great on mine. In fact the spool should have threads and the rollers could run right on the same grooves as the cable..

[greybeard]

.........I think a roller screw would be great on mine. In fact the spool should have threads and the rollers could run right on the same grooves as the cable..

Mike, do you mean that you have turned the sketch in post 23 'inside out' ,
with say three rollers outside the cable drum, a much deeper threaded groove for the cable to run in, with perhaps the rollers pinching the cable into the grooves ?

John
edit - No I've got something wrong - sketch please.

[Mike Everman]

Mike, do you mean that you have turned the sketch in post 23 'inside out' ,
with say three rollers outside the cable drum, a much deeper threaded groove for the cable to run in, with perhaps the rollers pinching the cable into the grooves ?

John

That could be made to work, but no, I was thinking the rollers would run on some bare thread either side of the cable. It's all a faaaaaascinating mental exercise. Enough cables and it could be stiff as hell. I wonder how much some really big threaded stock would be, like 50mm x 2mm pitch... hmmm maybe could buy a stud or giant setcscrew.. Of course, one can single point whatever is right, but I was thinking of a design where one could build with no fancy machine tools.

Hey, I'll tip a pint in your direction in April. I'm going over to the UK for some fun with jet engines and such. I'll be in Oxford for a couple of days, then over to Gloucestershire, and maybe a day trip up to Rotherham to look at some model turbines.

[greybeard]

Large diameter fine screw threads - bicycle hub thingy where the pedals go ?

Look forward to the beer pointing
John

[BillTodd]

I think it's a grand idea. I'm not sure it follows that you'll be able to get higher tension. The tensile strength doesn't change, after all.

I meant higher than one could get with an undamped wire (since higher tension tends to shift resonances up towards the step rate)

[Mike Everman]

Large diameter fine screw threads - bicycle hub thingy where the pedals go ?

Look forward to the beer pointing
John

Oh, There's an idea. Has journals and bearings as well. Neeto.
So a 50mm thread would get you 153mm per turn. Not too shabby, as they say. Only eight turns for a 4 footer.

[greybeard]

Mmm - 9 threads at 35mm diam = 980mm not bad for a pedal powered cnc

[aminear]

I have done a cable drive, works just fine. Take a look here for pics.
http://www.machsupport.com/forum/ind...ic,3368.0.html

Arnie

[Mike Everman]

Neat, Arnie. I see, the cable is for synching both sides... cool.

[aminear]

Mike,
Yes the gantry (X axis) movement is sync'ed with the cable, and driven with the ballscrew.
The Y axis on the gantry is pure cable drive, then the Z is ballscrew.

[desbromilow]

G'Day,

mostly lurking (not much else to do when I'm not getting too much made)..

I suspect I'll end up making 2 machines (gantry mills) a small one to get started with, then the one I originally wanted to make.

I'm not working off plans, just looking at people's ideas and trying to come up with a "best of all worlds" solution.

Anyway, I am leaning heavily towards one of the designs where the X axis is one a set of elevated rails, and the table is solid, with nothing underneath it.
This brings up the concern about racking on the X-axis, and I was worrying about synchro methods for 2 leadscrews. now looking at arnie's design, I "think" I see a leadscrew on the extreme edge of the x - axis travel, and the use of the cable to ensure the "non driven side" remains in synch.
Does this work? or have I misunderstood what's been done?
I came on searching for cable drive systems, since I think this would be an easy way to synch the two rails - similar to a rack system with a rack (and pinion) at each rail, and the pinions driven from a common shaft.

I've seen cable drive systems inside photocopiers, and they seem to work fine, but I don't know the relative accuracy compared to our work.

My end goal is to be milling/routing all materials up to (and including) aluminium if possible, and with sufficient accuracy to be able to do PCB milling, and lithopanes.

Failing cable drive, I was considering racks (made by driving on the thread of a screw (using the cross section of the pitch as a rack), or a captive chain as a rack, but I fear the backlash would be higher than that of a cable system.

Thoughts and comments?

Des

[aminear]

Des,

You are correct in how mine works, the cable loops around to the non driven side, does a crisscross and ties into that side.

It is not without some flex, I think I need to tighten the cables a little more. I just haven't gotten around to doing it, I will have to modify my cable attachments to my adjustable side. After I do I will try to get some measurements on flex. Maybe deflection amount with amount of force, so we have real numbers.
The cable I am using is pretty small I don't think it is rated for more that a few hundred pounds.

[vwhotrod]

old thread but whoever is following up, I suggest a cable spool like the ones found in your garage door opener at home, is about 5 inches in diameter and it has the threaded groove for the cable to sit on, it handles trememdous pressure from the door weight and springs and is already concentric with a hub. Im going to try it myself

[RICHARD ZASTROW]

This may be an echo: www.sagebrushtech.com/prod-rotolok.php

Dick Z

[handlewanker]

Hi all, we had a Cinncinnatti tool and cutter grinder at our works, and the table drive was by a 3mm multi strand wire cable....very smooth action, no backlash.

The aspect of using a cable round a drum to drive the carriage is probably well established so no need to hum and har about the design, the problem arises when the cable length increases, is unsupported and sags and vibrates with speed, but that can be offset by a damping arrangement that slides along the cable as the carriage goes back and forth.

The drive drum needs to be attached to the cable, not just wrapped round it with friction drive, otherwise you get slip, and so constant recalibration required.

As was pointed out earlier the cable wrapped round the drum and attached to it will form a thread as it winds on and off causing possible over laying but this can be guided so no problem in that field.

I machined many Telfers or cable drums for the mining industry in the 70's back in UK, that had rope grooves machined on them to take ropes about 30mm diam and more, and the wire ropes came out the side of the drum and were fixed with a wedge or a clamp to the side of the drum end outside face.

Some of the Telfers were a big as 2 metres diam and 3 metres long.

Having a cable hanging vertically, as in a mine shaft configuration has no problem, but when the cable is to be used horizontally, as in a router X axis drive mode, then sag and vibration will occur, but the accuracy and lack of backlash is without doubt, provided sufficient tension and cable support is present.

Without cable support you get a giant spring as the cable sags, and bang goes your accuracy, and the longer and so thicker the cable the more tension is required.

The problem of length and sag is also present in a lathe feed shaft and leadscrew, especially in lathes where the bed is long.

In that situation the is a folding down support that moves out of the way as the carriage gets to the centre of the bed.

Summing up, cable is go....provided you support it.

Cable drive and rack and pinion share the same bed, but one is DIY positive (cable), and the other is Engineering intensive (rack and pinion).

Using cable drive means you want to ensure travel dimensions occur according to the stepper motor drive over a considerable distance, and if you apply the Mike Everman principle in the toothed belt design to supporting the cable it probably would be a simple solution.

This implies a twin cable drive set-up, each side of the carriage, supported along the length as per the Mike Everman design, but using a wire wrapped round a largish drum (possibly grooved) and attached to it.

The wire is actually in two pieces, fixed to the drum and each end of the frame, and winds on and off the drum seperately as the drum/carriage moves back and forth.

The thickness of the wire would have to be strong enough to carry an accelerating load at a safety factor of at least 3.

So if the carriage with work weighs 40 lb and the acceleration inertia increases this to 100 lb and a safety factor of 3, we have a wire requirement tensile stress that will break at 300lb max, probably piano wire of about .8mm (.030") would be flexible enough for this to wrap round a drum of about 6" diam without fatiguing from the bending stress.

Using a multi stranded cable (motorbike clutch cable) of the same diam would introduce some undesirable stretch, but piano wire being a single strand won't as much once tensioned.

I wonder how much tensile stress a piece of stainless steel Mig welding wire at .8mm diam would take to break.

I would be inclined, if using that wire, to lay two pieces side by side and attached to the frame with a swivel compensating bar, like the old swingle tree on the cable brakes of cars around the 40's and 50's era, you'd never break that.
Ian.

[RICHARD ZASTROW]

Also, the cable can lay in a trough. This would negate vibration and sag etc.

Distance measurement on the cabled axis can be a separate. Another cable/resolver measuring device can be attached to that axis, such as the retro devices for HBM spindles Z axis.

I used all the above on a welding lathe. Not very critical but practical. LOL

Dick Z