PDA

View Full Version : Gear rack linear accuracy?



wierdscience
08-05-2008, 02:02 PM
I haven't been able to find any info for the linear accuracy of gear rack.It's easy enough on threaded products like acme or ball thread,but I find dead ends when looking at rack.

Any suggestions?

macona
08-05-2008, 02:44 PM
Depends on if you get stamped, milled, or ground sections. Ground stuff is very accurate, milled is good enough for plasma cutters and router tables. Though most machines that use rack have closed loop control to keep lost motion under control.

Just Bob Again
08-05-2008, 03:17 PM
If the manufacturer doesn't quote some sort of standard, like an AGMA Q number, you don't really know. Standard "commercial rack" can be anything. An entry-level precision rack, like AGMA Q5 is (I think) about .003"/foot. Q10 is around 10 times better. You'd need the relevant AGMA standard, which I don't have handy. The standards aren't free and rather pricey. Not posted on the net. Maybe a gear handbook cites them. There's cumulative error, tooth-tooth error, center distance error, and a load of other stuff. Some people can make them accurate to tenths over 10 feet.

SGW
08-05-2008, 05:11 PM
You may be able to dig out some of the AGMA spec from www.wmberg.com

I think this is right:

AGMA 5 = pitch variation +/- 0.0025
AGMA 10 = +/- 0.0004
AGMA 14 = +/- 0.0001

oldtiffie
08-05-2008, 10:47 PM
Thanks SGW for the info. Is error on a "per inch" or "per foot" basis?

Good question wierdscience. I checked Machinery's Handbook (27) and there is a lot of info on everything but rack.

The fastening and support to/for the rack is paramount as any "out of line", "deflection" or poor "centre distances" will have similar effects as "backlash" and cyclic errors in and between conventional gears.

Of course not all racks are "straight" or "spur" but many of the better or heavier loaded ones are "helical/spiral" with the "off-set angle" being equal to the helix angle of the mating gear/s.

Cutting "straight/spur" racks using a DRO is relatively easy and quite accurate. Cutting "helical/spiral" racks is quite another matter.

wierdscience
08-05-2008, 10:56 PM
Thanks for the answers,rack for some reason is a black hole of information.I called the mfg of the stuff I am looking at.Their engineering dept didn't know,said they had never been asked:D

I have some pieces of the same rack from two different mfgs.Both are milled,with some blue sprayed on they fit nicely together with consistent marking on each tooth over the 2' length.It should be good enough for the plasma cutter.

Looking around anything with better than .009" per foot be it screw or belt gets expensive quick.I'm thinking the milled rack will suffice and at $56 for 6 feet the price is right.

lazlo
08-05-2008, 11:24 PM
Looking around anything with better than .009" per foot be it screw or belt gets expensive quick.I'm thinking the milled rack will suffice and at $56 for 6 feet the price is right.

Yep, same deal with Acme screw stock and ballscrews: the price goes up exponentially with accuracy.

If this is for some kind of CNC, you could use a spring-loaded anti-backlash pinion gear, and then the remaining error would be mostly pitch error. If you're patient you can map the errors and compensate for them in Mach.

wierdscience
08-05-2008, 11:44 PM
Yep, same deal with Acme screw stock and ballscrews: the price goes up exponentially with accuracy.

If this is for some kind of CNC, you could use a spring-loaded anti-backlash pinion gear, and then the remaining error would be mostly pitch error. If you're patient you can map the errors and compensate for them in Mach.

Error mapping will be in the future me thinks regardless of what I do.It's gonna be a 4x4' plasma cutter table.

I have plenty experience building gantries,but not much in the electronics side of things.I've only done single axis stuff so far,mostly indexers.I have a lot of reading to do:)

macona
08-05-2008, 11:49 PM
If its for a plasma cutter than generic milled is all you need. Plasma cutting, at least entry level, is not very accurate.

wierdscience
08-05-2008, 11:52 PM
If its for a plasma cutter than generic milled is all you need. Plasma cutting, at least entry level, is not very accurate.

That's about what I figured,I'll be making pad-eyes,base flanges and that sort of thing,not clock gears.

oldtiffie
08-06-2008, 12:08 AM
WierdS.

Don't neglect considering chains and sprockets. They can sit in a channel/groove and can have adjusting screws at either or both ends. Some new chains and sprockets can be quite accurate.

wierdscience
08-06-2008, 12:13 AM
I've done chain and sprocket before and haven't totally rulled it out yet,#25 or 35 would be plenty.It would lower intertia since it would allow me to relocate the motor on the y axis to the machines base and it would also mean less wire tray to buy without that motor traveling.

nheng
08-06-2008, 01:02 AM
Wierd, I went looking for the same sort of info a while back on linear rack and on gears in general and did not find much. Ok, so you've got a pitch accuracy of 1 part in 10^14, big deal because what happens as you turn the mating gear and the teeth traverse thru their geometry while never losing contact.

It seems to be that there are numerous small errors to be had in the gears unless perhaps those errors almost disappear with very high precision (unobtainium and unaffordium) parts. Den

lazlo
08-06-2008, 10:09 AM
It seems to be that there are numerous small errors to be had in the gears unless perhaps those errors almost disappear with very high precision (unobtainium and unaffordium) parts.

Normal commerical spur gears have AGMA ratings too, but I've never looked into it. If you look at the finish on a Boston or Martin gear, it sure isn't shaved or ground. So I'm guessing the gears you find in the Boston, Martin, SDPI etc have pretty loose tolerances.

kf2qd
08-06-2008, 01:33 PM
Gear rack is probably an order of magnitude better than that of the plasma arc. If the machine is light you will probably do just fine with soft gears and rack. Used to build 14" and larger gantry machines and we use soft rack with a hardened pinion. rack was 3/4 wide and showed no wear, but the pinion (1.5" dia.) would wear out in a year on a machine used 2 shifts a day. The hardened pinions never showed any signs of wear. We were running a 4000 lb. machine at a max of 600 in/min. ( at that speed in approx 1 foot). For a home machine - don't use the smallest pinion you can find - a slightly larger pinion will run smoother. Assuming you are using stepper motors - use a good gearbox - that will be where most of your backlash will take place, not in the pinion & rack. Spring load your pinion into the rack for 0 (zero) clearance - it doesn't need to be extreme - just enough spring preload so that the pinion doesn't bounce when you change direction on the axis.

Just as important is how you mount your torch. If the slide there is out of alignment it will add more location error into the process than will the accuracy of the rack. Are you going to run some kind of voltage follower for a torch height control? (100 - 125 V)

because plasma cutting is a non-contact process you can build fairly light, but watch for wobble in your cross beam, or run 2 motors on X, one on each end of the cross beam.

LES A W HARRIS
08-06-2008, 07:17 PM
I haven't been able to find any info for the linear accuracy of gear rack.It's easy enough on threaded products like acme or ball thread,but I find dead ends when looking at rack.

Any suggestions?

WS, what is the Diametral Pitch of your proposed rack? My newest specs are from 1980. It is a tablular chart, I will extract relavent data for your application, if you wish.

Cheers,

oldtiffie
08-06-2008, 08:46 PM
Thanks Les.

Would you mind posting it here too if possible or at least a link to where you have it or where it is?.

I have just used the requirements for a large gear of the required DP to approxiamate a rack.

Info on racks does seem to be a "missing link" for me at least.

lazlo's comment is exactly my observation as well:


Normal commerical spur gears have AGMA ratings too, but I've never looked into it. If you look at the finish on a Boston or Martin gear, it sure isn't shaved or ground. So I'm guessing the gears you find in the Boston, Martin, SDPI etc have pretty loose tolerances.

wierdscience
08-06-2008, 09:51 PM
WS, what is the Diametral Pitch of your proposed rack? My newest specs are from 1980. It is a tablular chart, I will extract relavent data for your application, if you wish.

Cheers,

Thanks Les,the rack I intend to use is 16 pitch 20*PA 1/2" face width.

wierdscience
08-06-2008, 09:54 PM
Normal commerical spur gears have AGMA ratings too, but I've never looked into it. If you look at the finish on a Boston or Martin gear, it sure isn't shaved or ground. So I'm guessing the gears you find in the Boston, Martin, SDPI etc have pretty loose tolerances.

I think those are pot broached,at least the smaller ones.That's one reason why I use rolled pinion wire and make my own.Better accuracy and tougher surface are the others.

wierdscience
08-06-2008, 10:17 PM
Gear rack is probably an order of magnitude better than that of the plasma arc. If the machine is light you will probably do just fine with soft gears and rack. Used to build 14" and larger gantry machines and we use soft rack with a hardened pinion. rack was 3/4 wide and showed no wear, but the pinion (1.5" dia.) would wear out in a year on a machine used 2 shifts a day. The hardened pinions never showed any signs of wear. We were running a 4000 lb. machine at a max of 600 in/min. ( at that speed in approx 1 foot). For a home machine - don't use the smallest pinion you can find - a slightly larger pinion will run smoother. Assuming you are using stepper motors - use a good gearbox - that will be where most of your backlash will take place, not in the pinion & rack. Spring load your pinion into the rack for 0 (zero) clearance - it doesn't need to be extreme - just enough spring preload so that the pinion doesn't bounce when you change direction on the axis.

Just as important is how you mount your torch. If the slide there is out of alignment it will add more location error into the process than will the accuracy of the rack. Are you going to run some kind of voltage follower for a torch height control? (100 - 125 V)

because plasma cutting is a non-contact process you can build fairly light, but watch for wobble in your cross beam, or run 2 motors on X, one on each end of the cross beam.

Cross beam will be 3x3x1/8" 6061 tubing,it's only 56"OAL so it's pretty stiff.Inside of it will be a 1/8" flat pushed inside the diagonal to add rigidity.

Planing on spring loaded pinions.I have the design in mind for the stepper,cogbelt reduction and pinion shaft all in one unit.Front face bottom of the 3x3 will be a THK linear slide,top face center will be a 1/2 x 1/2" CRS rail with a bearing either side making it slick as possible.

Pinion will be 16 P 20t 20*PA (1.38" OD) made from pinion wire.

Torch will be another slide with probably 6" over all travel,I want the possibility of 4th axis tube coping later.Any advice there other than the obvious daylight issue I should know about?

Torch height control,I was recomended the one from Cambell designs-

http://campbelldesigns.net/plasma-torch-height-control.php

Anything better,less complicated or less $$ etc I'm all ears.

Oh,I'm also getting a machine torch with the plasma when I buy it along with the hand unit.I figure the straight torch will nip a few problems in the bud.

LES A W HARRIS
08-07-2008, 12:34 AM
Thanks Les,the rack I intend to use is 16 pitch 20*PA 1/2" face width.

WS, Here you go.

http://i37.photobucket.com/albums/e97/CURVIC9/08%20GEARS/LAWH0101.jpg

http://i37.photobucket.com/albums/e97/CURVIC9/08%20GEARS/LAWH0102.jpg

http://i37.photobucket.com/albums/e97/CURVIC9/08%20GEARS/LAWH0103.jpg

http://i37.photobucket.com/albums/e97/CURVIC9/08%20GEARS/LAWH0104.jpg

So for 16 DP, tol's would be class 6 minimum, at class 9, tooth to tooth & total composite, profile & lead are required certs with corrosponding increase is costs as someone posted above. Summing up; class 6 = .002" pitch, .006"/foot index tol, & .011" mounting face to pitch plane.
Class 8 = .001" pitch, .003"/ foot index tol, & .007" pitch plane parallel to mounting face. To improve or check the later mount rack tooth spaces on 0.1080" dia pins to true up mounting face. Your supplier should have noted tolerances, to what they are selling?
Cheers,

wierdscience
08-07-2008, 11:17 PM
Thanks Les!

alsinaj
08-08-2008, 12:59 AM
Weird: You wrote: "It would lower inertia since it would allow me to relocate the motor on the y axis to the machines base."

Could you explain your system for mounting the Y-axis motor on the machine frame? I've seen gantry cranes that use a differential hoisting mechanism to produce up/down and left/right motion from two stationary motors. If the two motors rotate in opposite directions at the same speed, the hoisting hook moves up or down. If the motors rotate in the same direction at the same speed, the hook moves left or right.

If that is the type of drive you are referring to? If so, how do you program your CNC system to output differential drives to two motors instead of independent X and Y drives?

macona
08-08-2008, 04:00 AM
Other options are using a timing belt with the gantry attached to the belt.

Also using cable wrapped several times around a drum. This can be very accurate.

wierdscience
08-10-2008, 12:21 AM
Weird: You wrote: "It would lower inertia since it would allow me to relocate the motor on the y axis to the machines base."

Could you explain your system for mounting the Y-axis motor on the machine frame? I've seen gantry cranes that use a differential hoisting mechanism to produce up/down and left/right motion from two stationary motors. If the two motors rotate in opposite directions at the same speed, the hoisting hook moves up or down. If the motors rotate in the same direction at the same speed, the hook moves left or right.

If that is the type of drive you are referring to? If so, how do you program your CNC system to output differential drives to two motors instead of independent X and Y drives?

No differential drive.All I am doing is putting the y axis drive that would normally be on the gantry traveling back and forth down the rails on the frame base.This way the motor is stationary and pulls the gantry back and forth.Not moving the weight of the motor and it's cables/trays back and forth reduces inertia and decreases the chances of lost motion due to missed steps.It's also cheaper since it allows the use of shorter cables and eliminates one cable tray.

alsinaj
08-10-2008, 02:57 PM
WeirdS,

Thanks for your reply. Don't mean to highjack this thread, but I would really like to understand how to mount the Y-drive motor on the machine frame because of the advantages you describe.

Here's what I don't understand. The attached drawing shows an X-Y system with the Y motor fixed to the frame. The Y-drive belt/chain/cable runs to the Y-carriage via an idler mounted on the gantry. Let's say you send a signal to move the X-motor, but not the Y-motor. In that case, the gantry moves, carrying the idler with it, which causes the Y-carriage to move the same distance as the gantry moves. Result: sending a command to move only in the X direction produces both X and Y motion.

Where am I going wrong? If the set-up you have used is different from the one in my drawing, could you describe it? Thanks.

http://i293.photobucket.com/albums/mm56/alsinaj/FixedYmotor.jpg

wierdscience
08-10-2008, 10:43 PM
Your not going wrong,I had a brain fart and typed Y instead of X.Maybe this primative cave drawing will clear things up:o

http://img.photobucket.com/albums/0903/wierdscience/GANTRY.jpg

alsinaj
08-11-2008, 12:30 AM
WeirdS,

All clear now. Thanks.

oldtiffie
08-11-2008, 05:39 AM
Chains and sprockets can be adjusted for position and tension very easily. They can turn or change co-planar direction easily "as is". They can be connected by a square block - with or without swivels - to change to any other plane. eg a square block can connect two chains at 90 degrees - or with a swivel - any other angle. Having the chain feed onto and from the sprocket via a square section hollow guide to match the chain width and height works wonders. "Marine quality" is made for some of the harshest environments there are.

Chains and sprockets need little support, are not fussy like gears and rack as regards positioning, back-lash etc. and can and do work in some very aggressive environments as their sealing is very good. It can all be "off the shelf" stuff without the "special needs" and "special gearing" other wise required.



..................................
Also using cable wrapped several times around a drum. This can be very accurate.

Very true too - especially with high quality pre-stretched super extra-flexible wire rope (think "Marine/yachting/rigging") wound around warping drums or capstans. All they need are tubes if they need to be contained. They are readily made up at any ship-chandler or marine/marina shop - and certified as to their safe load capacity as well. Normal marine fittings can be used as regards connections and adjustments.

You might be quite surprised as to the loads that chains and cables can carry.