[Art Ransom]

I am in process of building a 4 axis mill for wood columns with a capasity of 24" by 12' controled by Mach 3. I started out with no intention of motion control on the A axis but just speed control. Now I would like to have CW and CCW control to allow carving on the columns. There are several ways to acomplish this but backlash is the problem. My best idea so far is get largest gear possible ( 24" would be ideal) with a 1" bore and drive it with a 1" or 2" gear directly connected to the stepper motor. Suggestions, ideas and help GREATLY appreciated.

[ger21]

You could use a multiple stage belt reduction, and it would probably have much less backlash than gears.

[turmite]

Hi Art,

I have been following your posts here knowing what you have planned. I built my own a axis, or better said, I designed it and had it built. I do not do nearly as heavy work as you do but I want to give you my input as it might be helpful to you. I used a 300 oz in stepper and belt drive system with like a 5:1 reduction. What I found was the reduction was no where near enough and the motor was not nearly strong enough. I eventually added mor reduction to the point I am at now, which is 17:1. While this works.....kinda, I still have movement of my part when putting the bit into the wood. I carve gunstocks and set my carving system up a little differently that may be inducing some of this probelm. I offset my center to reduce the amount of high dollar wood I have to waste on the blank. Your centers will of course be on center and therefore won't be a factor, but the sheer size of the part tells me that you are going to need a hoss of motor to do what you need.

There is a fella in Dallas you need to get in touch with but I have tried his email and his url both of which seems to be dead. Masterwerkes was the company name and his name is Phill Pittman. Mucho talented man and I think he posted a long time ago here on the zone that he built an a a axis using a car differential! I'm getting a vision of Tim the Tool Man Taylor here with MORE POWER!!!:banana:

Mike

[HuFlungDung]

Use a single stage worm reduction gearbox. If you are merely indexing, a brake would be a good idea, otherwise you will need a worm reduction gearbox with very low backlash.

[Art Ransom]

Thanks for reminding me about Masterwerkes. I talked with Phil 6 months ago and found his web and wii tru and call him tomorrow. I found som worm reducers but they had up to 2.r degress back lash. Apparently you are also getting back lash with the belts. On a 24" piece 1 degree of back lash there is .2" of back lash. I would like to get over 75 to one.

[Capteod]

If you just rotate in one direction you do not need to worry about backlash in the A-axis.

I have built a number of Barrel CAMS -over a hundred using a simple A=axis stepper system attached to a cheap INDIAN Manufactured Rotary table. The repeatability of was very acceptable. I checked them all for a customer and the resulting Cpk was 1.87. My solution was to use the inexpensive rotary table, even though the backlash was horrible. Remember that backlash is only a problem if you are changing the A axis rotation. In my case I use an optical switch which has an accuracy of 0.01 degrees.

[ger21]

Apparently you are also getting back lash with the belts. On a 24" piece 1 degree of back lash there is .2" of back lash. I would like to get over 75 to one.

With the right belts, you should get very little to no backlash. Certainly not 1°. Look for GT2 profile belts at www.sdp-si.com

[pointcloud]

What's wrong with using a rotary table? I did, I mounted my servo directly to a hand operated table. Works great. I can mout a piece 8 feet long by 6 feet in diameter. You will not need a brake, or a gear reduction, and you can adjust backlash anytime you need to.. I am using camsoft, galil, with rhinocad/cam, with a digitizing laser, and cutting with a router in foam for the lawn and garden industry. It works well for me.. Good luck..

[HuFlungDung]

Yes, I was thinking of some kind of worm box that you would alter to improve the backlash.

I'm thinking the main advantage of a worm reduction is to eliminate jittering of the part because of the torque advantage that the cutter has, working at a large radius. A single thread worm reduction is very difficult to backdrive, and I do not imagine that you can achieve the same stability with any kind of parallel shaft reduction unless you have a monster sized motor holding it back. Even then, small vibrations take time to damp out electronically.

[Art Ransom]

Hate it when someone shows me that I took a simple problem and made it complex. Can you give me some details? What is the gear reduction, size of motor and max RPM of output shaft? I figure that my max RPM will be 100 RPM though I suspect that 10 to 25 will be my normal range.

[HuFlungDung]

If using a standard commercial worm gearbox, I'd size it appropriately for the job at hand. Probably for a 24" diameter workpiece, I'd want a worm gear maybe 8" diameter at the minimum, which would be 5 diametral pitch, 40 teeth, and the reduction would be 40:1 This will give you a max of 62 rpm with a 2500 rpm motor.

Having looked inside of some cheap rotary tables, I'm not all that impressed with what they give you. The worm is typically not in an oil bath, and is not supported on both ends. However, the worm backlash is compensated for on an adjustable eccentric, although this may involve some fudging to get it really tight as usually there is some means to stop it from going in too tightly from factory.

The cheap rotary table also has very limited end thrust capability on the worm shaft....ie., "they don't got real bearings in 'em " So because they use crude plain bearings (ie, a shoulder on the shaft rubbing against the base metal of the adjuster housing), they can wear themselves loose.

Many commercial powertrain gearboxes have a worm shaft that has the worm thread integral with the shaft, meaning you cannot take the worm off the shaft, because it is the shaft. And, reboring the housing to make an eccentric adjustable worm would be quite a trick since the original casting was not designed with that intent. You need some room to work, to make an eccentric worm adjuster.

I would then resign myself to the purchase of two hardened worms and a matching bronze worm gear from Browning Gear, or Boston gear, wherever. Then, I would build my own gearbox. I have a machine shop, so this is practical for me.

Instead of making an eccentric worm housing, I would instead mount the two worms end to end on the worm shaft. It would take some work to face the worms off so that when placed on the keyed shaft, the helix lines up properly. Then, backlash adjustment would involve placing a precision thin shim between the two worms, in effect, seperating them, and taking up the lost motion in this manner. Trial and error would determine if the worm gear is accurately made so that there are no places in rotation where it all binds up tight...too tight. Then, the whole mess: bearings, worms, spacer would be sandwiched tightly together on a shaft with no endplay whatsoever.

I'm not saying that this is the only method, or that you might be able to make do with less perfection, but it may give you some ideas on what to do, and how to approach it.

[Art Ransom]

How about using a driven worm and 2 idler worms with one loaded against the CW face and the other against the CCW face? I don't have metal shop capabilities yet.

[keebler303]

you can't have idler worms past around 5:1 ratio. They are not back driveable in the higher ratios.
Matt

[Art Ransom]

How about using a ring gear with outside teeth? Since I have the room I could use a ring gear mounted on backing plate mounted on A axis shaft. If driving gear was 2.4" it would give me a 10:1 ratio. Would using a 1200 oz stepper should give me enough torque.

[turmite]

Art how heavy are the columns going to be. This weight, along with the friction, weight of the moving compontents will dictate whether or not this would work. I have no idea if what I am about to say is correct or not, so hopefully someone will correct it if I am indeed wrong. at a 10:1 reduction using a 1200 oz motor, you would end up with 12,000 oz equivelant torque.

That's 750lbs?? I think!

Mike

[pointcloud]

Art have fast is your stepper?

huflungdung was right... You should use a worm (2 in-fact)

10 to 1 is not much mounted streight, but a 10 to 1 worm is much stronger because the right angle, motor(worm) to gear. The way you are mounting it will be weak...
Small projects it MIGHT work... But if you are going to all the trouble then you might as weel do it right the first time...

Plus your steppers have a count, 1-4, 1-8.. So 10 to 1 on a 360 deg would be 36 motor turns, with a count of 4x36= 144 counts, or 8x36 =288 counts.. Not much for accuracy.. remember 360 deg. 60 min in a deg., and 60 seconds in a minute angle, then hundreds... 1 degree to far on a part it to much error, well my parts...

good luck with what ever you do....

pc

[pointcloud]

12,000 at the center of the gear, not as strong when you add a lever...

Just imagine adding a piece of pipe to a wrench. A 3/8 bolt, what's the break strength. Can you break it with a little pipe and wrench... Yes...

pc

[Geof]

Turmite; Try http://www.masterwerkes.net/about.htm for Masterwerkes, it worked a few minutes ago.

Art; Use your ring gear idea but drive it with a worm. You need to get a worm cut with the correct pitch to mesh with the gear and then set the worm axis at an angle to the gear so the teeth and screw flanks are parallel. The screw only contacts over a small area but the loading will be very low, the speeds are also fairly low and if you use bronze for the screw you can get by with minimal lubrication. Make the screw mount adjustable so it can be moved further in to the gear when things wear.

[Rance]

I can mout a piece 8 feet long by 6 feet in diameter. I cut foam for the lawn and garden industry.

Art, I don't want to sidetrack your thread but I have to ask pointcloud this...

Pointcloud,

"8 feet long by 6 feet in diameter"? Sounds interesting. Can we see a pic or two of some of your work?

[pointcloud]

Sure... I am haveing a little trouble. with my graphics card..... No. A lot of trouble... The parts are so big that my card gets bogged down and errors out of my rhino program... I have some renderings of a few fountians.. I am going to machine the lips on the a axis, so you can see the bottom bowls are 48 in dia... I am cutting polystyrene (white 5# per cu foot foam) Did the pics make it?