Servos for a plasma table?

[hotponyshoes]

Hi all,
I have been absent for a while due to spending all my time on my table construction!!

Anyway, it's coming along well. The gantry and table are mostly done.

My next problem is working out what power servo motors and what sort of controler I need.

It is a 10x 5 bed, both axis run on linear rails and I have done a bit of testing with some fishermans scales to get a rough idea of what power I need to pull the thing around.

To move either axis takes a pull of about 1lb. There are no 'rough' spots and its a smooth travel on all lengths.

I have put a bag of sand on the gantry to simulate the weight of the plasma torch, wires & hoses and a THC if I fit one.

Main problem is I have not decided on the drive system yet, I am tempted to use ball screw for the precision but a 15' length will push the cost a bit much.

I am looking into chain or toothed belt.
Could I use smaller motors with ballscrew (I presume they need less force than a direct drive belt)
My main question is what size motors do you guys use on simeler size tables?
Should I just go for the biggest I can fit incase I decide to change to a router at some stage?

Finally, Should I get a controller card for each axis or would one of theos ex-industrial 3-axis controllers work (ie. could I connect it to my pc?)

Thanks all!

[Robin Hewitt]

Motor size depends on what kind of acceleration/deceleration you want. Ignoring your rolling resistance, which sounds fairly low, your 1 lbf pull will accelerate a mass of 1 slug (32.16 lbs) at 1 ft/s/s.

The equasion is Force = mass x acceleration so the maths is real easy.

You don't really need a lot of torque when your x and y are rolling rather than rubbing.

If you add screws you have to up rate your motor to allow for the inertia of the screw and it will be a fairly heavy screw over that length because you can't add a centre support.

Chains are a bit coggy and as soon as you add springs to compensate for wear you will get sag.

Long timing belts could be a bit elastic which is why most folk seem to opt for rack and pinion. The problem with rack and pinion is pi immediately comes in to the resolution equasion unless you get one of those fancy tooth pitches which cancel it out.

[jcc3inc]

Hello,

The constant of PI is NO factor. All CNC controllers have provision to accommodate various distances per step. Don't let that small item effect your decision.

In my opinion rack and pinion is the way to go. You can buy good rack from www.stdsteel.com for very reasonable prices. We used 24 dp but 20 dp is more common and pinions are easy to buy. Use 20 degree pressure angle rack and pinions, spring load the pinion into the rack with a force of about .37 x max drive force. Thus a 100# max drive force would require a force of 37# holding the pinion into the rack. To make equal engaging forces for running forward and backward, pivot the drive mechanism so that the pivot line intersects the pitch line of the gear rack. In order to have no undercutting of the pinions, use at least 18 teeth on the pinions; something around 30 T is better as there are more teeth in contact with the rack.

These combinations will result in a good, zero backlash, smooth running machine.

Regards,
Jack C.

[Robin Hewitt]

Has anyone ever considered the spring steel band aka clock spring?

Lightweight and practically inextensible. It probably needs a friction drive but if bought in those handy packs we call "tape measures" it comes with a built in encoder

[energyforce]

You can go with belt drives too.

A = ( change in velocity / time )

So calculate your maximum velocity you want the machine to accelerate to. For instance lets say 12 in/sec. Your initial velocity would be say 0 in/sec. Therefore your change in velocity would be (12-0) = 12in/sec.

Now say you want to go from 0in/sec to 12in/sec in 0.5 seconds?

A = (12 - 0) / (0.5) = 12in/sec^2

Now to find how much mass you will have to move at this acceleration is tricky. I would put the mass on a scale and weight it. Say you weight it at 100lbs. Divide that number by 32.2 (dont ask).

F = ma = (100lbs/32.2) x 12in/sec^2 = 37lbs.

Yep thats 37 lbs to push that weight from 0 to 12in/sec in half a second. Now this does not account for friction. You say that you used the spring scale and moved it and it said 1lbs? Add 1 lbs to 37lbs to account for friction to make it 38lbs.

Now say you wanted to use a timing belt or rack and pinion you would need to convert this force into torgue.

Torgue = force times radius of the pulley/pinion.

Say you have a 6 inch diameter pulley/pinion. This is a 3 inch radius pulley/pinion.

Torgue = 37lbs x 3 inches = 111lbsin.

This is where ozin comes in. Use a unit converter online and convert 111lbsin into ozin.

Although for horsepower rating of the servo you would need to find out how fast the motor will be spinning to compute the rpms. Because power = speed times torgue. Ill save that for another day if you need this calculation.

Also you would probably want to calculate the interia ratio of the motor and make sure its at par. something between 3 or 5 i forget the units. That one is a pain in the ass. I bet more than 80 percent of the ppl on cnc zone done even calculate this and their machines seem to work just fine so i wouldnt worry about this i guess? What do you think?

Its my lunch break now! i hope i didnt make any mistakes, no time to check!

Erik

[Robin Hewitt]

F = ma = (100lbs/32.2) x 12in/sec^2 = 37lbs.

Force in lbf, mass in slugs, surely the acceleration is ft/s/s???


F = 3.1 slg x 2 ft/s/s = 6.2 lbf

If you drop a 32.16 lb mass in London it is accelerated downwards at 32.16 ft/s/s by a force of 32.16 lbf. So, 1 lbf should accelerate it at 1 ft/s/s

This really is a lot easier in SI units

[energyforce]

This really is a lot easier in SI units

Ready to re-start another revolution to standardize SI units Robin? lol

Erik

[Robin Hewitt]

Ready to re-start another revolution to standardize SI units Robin? lol

The European Union fearing a revolution just stopped their campaign to force SI down our reluctant British throats. It is now legal to sell a pound of spuds again, big relief

[evena4]

I am also running into issues with my gantry router. I built an 8.5' x 5' table to handle full sheets of wood and built the motion control with some 5/8 ACME screw to move it. I used steppers with a 3:1 reduction and find that the travel is a bit slow and the gantry seems to crab some (it is riding on 1" supported shafting).
All of the items to build it were found at salvage yards and Ebay etc.. I was thinking of simply adding a second screw to the other side and slaving it off my first shaft by using sprockets and #25 chain at one end of the table to fix the crabbing but then I started to wonder if I could run a chain drive at each side in place of the acme scews.
--
I have read your posts about chain sag and inertia/positioning issues and am wondering if it might just seem possible to cure the two by running the chain cogs horizontally and making a few supports out of UHMW to help the sag and then using a right-angle bevel gear reduction head or worm drive (possibly from my DeWalt worm drive saws that is retiring soon) to help control the braking of this mass.

I am willing to give it a shot if the rapids will be much faster (I am routing wooden shapes and accuracy is not exactly an issue right now though I have a plasma cutter I would like to run on this later.

This is strictly a hobby machine as I am a programmer by day.

Suggestions welcome.
-Manny

[Robin Hewitt]

I think lying the chains flat could make it tricky to keep them on the sprockets and you deny yourself the wonderful possibility of running axles from side to side.

You could support the chain with idler sprockets if you used sideways extender links to connect to the gantry, it would have to be able to pass over them