Linear motion without bearings

[galaxyman7]

Hey, I am designing a machine that can basically replicate all of its parts. I am trying to make it as cheap as possible. I came up with this design for linear motion that I hope will eliminate the need for bearings or lead screws, ect. I just want some engineering opinions on whether it will work or not. One of the gears is driven by a motor, the others basically copy the movements of the first. I am hoping that by having two pinions that are tightened together with a timing belt will eliminate backlash. The cart will be connected to the 4 axles protruding from the gears. The gears will be enclosed between two sheets of plastic, to prevent any motion back and forth, with slots for the axles. (obviously the picture is not accurate, the teeth on the gears will be much smaller)

Keep in mind, the machine has to be able to mill this rack and pinion in order to self replicate!

[wildwestpat]

Hi

The basic mechanism would work but with questionable accuracy and speed limitations.

1. To keep positional accuracy under control a simple rack has a problem in that the progression is not perfectly uniform with rotation of the pinion. There are various ways of improving this such as helical tooth forms. This problem is distinct from backlash.

2. A timing belt is designed to have virtually zero stretch and as such will not provide the spring action you are looking for to sort out backlash by loading onto both faces of the rack.

3. In the machine replication process there is obviously a starting point. The start has to be to define:-

a. accuracy required both in terms of the tolerances required on the parts to be made.
b. overall size of those parts
c. material of which the parts are to be made
d. available materials for construction of the starting machine
e. materials and components for second generation machine
f. method of securing increased accuracy or at least maintaining it generation to generation.

The design process is iterative and will be a compromise which can be summed up as COST!

The challenge is to keep or improve the accuracy with successive generations. Your design as posted will in my opinion suffer from increasing inaccuracy with each successive generation. The main problem is with the tooth form and the backlash which will be compounded with the positional accuracy of the parent machines drive in the direction of the axis. Think in terms of the absolute position of each tooth on the rack and the ideal possition and you will see the effect of backlash adding to the uncertainty of the position of the parent's place on the axis. Add in any machining tolerances and you get an ever increasing problem from one generation to the next.

Yes this is an interesting problem and is where the skills of the human toolmaker remain in the equation.

Kind regards

Pat

[LaserImage]

I don't know if it's just the way you drew it, but you won't get any movement with the top and bottom touching.

[keebler303]

Originally Posted by LaserImage I don't know if it's just the way you drew it, but you won't get any movement with the top and bottom touching.

No, there will be movement.

Matt

[RICHARD ZASTROW]

Try www.nexengroup.com linear drive, RPS

Dick Z

[lumberjack_jeff]

Those roller pinions are really cool. Do you have a feel for pricing?

[galaxyman7]

Ok lots of stuff to answer:

I was thinking of making the parent machine from very accurately machined gears. I would not be making those myself. This way I have a good start to make accurate machines for the next generation.

Do you think an extension spring from one gear to the next would get rid of backlash?

My understanding is that if the gear teeth are very small there will be practically no difference in progression. I could change the gears to friction rollers, if that would help.

This machine will be modular, which means you can take multiple pieces to make a bigger piece, therefore eliminating the need for a bigger footprint machine.

You are right about decreasing accuracy, it will be hard to make a machine with the same tolerances as the first. I need some help in this area.

To answer lumberjack's question, the pinions and rack should be machined by the parent machine, and made out of plastic. I am thinking either 1/8 LDPE, or 1/16 PVC, preferrably the PVC, since it is dirt cheap. This whole assembly with all of the gears and belts and everything should cost around $7, with the timing belts being most expensive at around $1.40 each, the 8 bearings for the gears costing .35 cents each or 2.80 total. The rest is cost from bolts, nuts, and plastic. My aim is to make practically everything out of 1/16 PVC and 1/8 LDPE.

By the way, to cut costs on this machine it will be driven by normal DC hobby motors that have laser positioning. This should also make up for decreases in accuracy due to backlash and differences is progression (hopefully). Something like this: http://www.active-robots.com/product...rs-sharp.shtml but hopefully cheaper.

Or I could use a sonar positioning system, which might be harder since the speed of sound varies.

The signals from the computer will be translated by a microproccesor.

[galaxyman7]

Just thought of this. Instead of using laser positioning, I could use an RC time circuit that would measure the resistance between the cart and the side of the machine. I would use a metal strip along the axis, and have a brush on the cart to make electrical contact. The resistance would change depending on how far away the cart is. Don't know if this would be accurate, but it would definetly be much cheaper than laser positioning.

[RICHARD ZASTROW]

lumerjack jeff PRICEY

Dick Z

[lumberjack_jeff]

Originally Posted by RICHARD ZASTROW lumerjack jeff PRICEY Dick Z

Indeed. I just got a quote of roughly $6 large for racks and pinions for a 4x8 router table.

[galaxyman7]

you are joking right? just making sure

[lumberjack_jeff]

Originally Posted by galaxyman7 you are joking right? just making sure

Sadly, no.

$1783 for four pinions.
$3943 for 30' of rack.
$0109 for the alignment tool