I want to build a rotary table that can handle continuous milling!
Do you think my drawing would work?
The upper gear belt pulley can rotate freely on the axle, and is connected to the axle via the spring, and is preloaded.
Printable View
I want to build a rotary table that can handle continuous milling!
Do you think my drawing would work?
The upper gear belt pulley can rotate freely on the axle, and is connected to the axle via the spring, and is preloaded.
As an alternative, "google" Howimat. This concept is ball driven like a ballscrew in the round.
Dick Z
I will check it out! I have also been thinking of using 2 servos and do the preload eletronically with a µC
Looks like it would work to me, although mixing gears, lubricant and timing belts might be a problem.
How about placing the two worm gears at an angle to each other, driving one directly and the other via a bevel gear? (if you see what I mean)
Or...
What about a buying or copying Commercial anti backlash worm drive ?
Hmm, thats an interesting idea! I'll make a drawing! :)
I made a drawing and realized that the worms would need to rotate in the same direction in this configuration :P
The dual servo version seems more and more appealing!
Just swap the main-shaft bevel gear to the front of the perpendicular one.
There are a lot of commercially available low backlash drives on eBay. Search for "harmonic drive" or "bayside drive". One of those sure will least trouble and expense.
Another possibility is to preload the heck out of your rotab drive with another gear and some bellvues. Not sure how much torque you plan to put on it though.
People also run 2 of the main circular gears spring loaded to spread the teeth just enough to get rid of the backlash. In this case, you'd make the spread adjustable with a threaded screw.
These last two solutions seem prone to binding and wear.
I would be tempted to also consider a big timing belt drive. properly set up, you'll get minimal backlash. People are even using these for linear motion with good success.
For really high torque situations, you might still find a brake is useful. Slew the 4th axis to the desired position, send an M-code to lock the brake, cut, and unlock. Small disk brakes from everything from mini-bikes to lawn tractors are readily available and work with an air over hydraulic and solenoid valve.
Cheers,
BW
Aaah yes...
:withstupi
Harmonic Drive is the way to go, if your willing to spend a little money.
http://www.powertransmission.com/iss...6/harmonic.htm
I want it to be able to use it for continuous milling, like gear hobbing and milling springs, not just indexing!
I think it's best to have the worms axially stiff because they can't be backdriven. So if you just can keep both worms in contact with the gear teeth with a little rotational preload it should be backlash free and much stiffer than the version with axial preload of the worms or split wormgear!
Harmonic drive seems nice! But i like to design and build stuff more than paying for complete solutions ;)
I was thinking of a gear belt drive at first.. but it's a little impractical when you want 1:100 ratio or so..
Even 50:1 is awkward. I'm thinking about retaining the rack & pinion Z axis in my drill mill (in order to keep a manual option),and I will require about a 50:1 back-lash free worm drive for that, hence my interest in your project.Quote:
I was thinking of a gear belt drive at first.. but it's a little impractical when you want 1:100 ratio or so..
Twin servo-motor would certainly work, but could be as difficult to get 'right' as your twin worm design, and would almost certainly be more expensive.
You could perhaps build something like this!?
http://www.globalspec.com/FeaturedPr...lDrive/74196/0
That's nowhere near enough for my hand re-sharpened drills :)Quote:
Nominal tangential load up to: 30,000 Lbs.,(135,000N)
Machine your own harmonic drive :)
Would probably be the best solution for your application, you can get 100:1 no problem, and I'm pretty sure most of the new 4th axis equiptment are using harmonic drives? It's how I ran accross them.
Decent load handling, probably indefinate gearing available, no backlash due to design, compact and readily available.
But if your determined to make it all yourself, I like the duel servo idea. But are you going to make the servo's?:stickpoke
Just joking, I get it,
MC
Just a couple of basic engineering things to think about.
Solid worm gears need some clearance to operate or they lock. Worm gears are not very efficient which is why they don't backdrive easily.
Let's say you adjust your drive worms for .0001 clearance. Have you thought about runout on the main gear? This is why Hass hobs their gears after they are mounted. Ever priced a class 9 worm gear set? I pay over $4500 for mine and they are only 6 inch in dia.
Spring loaded worm gears use very light springs to allow the halves to move when operating. Because of this it's very easy to overcome the springs and they have very limited top speeds or they wear like crazy from the sliding friction between the worm and wheel.
The wheel gear is fairly large and grows and shrinks with temp changes. Set one up in a 68 degree room with almost no clearance and it will be locked solid when the wheel temp goes up to 80 degrees. Worm gears generate a fair amount of heat when running because the sliding friction is so high.
Harmonic drives can be purchased in zero mechanical backlash configurations but the main drive flexspline is made of spring steel and will deflect under pressure (although not as much flex as a split worm gear). Grab the end of a robot arm and move it back and forth. Older robots are almost all harmonic drives.
Dual worms driven by servos are a nice idea but how do you get around the problem that the gear teeth are not going to be exactly spaced. You have to consider manufacturing tolerances into any design work. No gearset is perfect.
There are also high resolution direct drive rotary motors. Basically linear motors wrapped in a circle. These are used in the rotary axis of very high end machine tools and high priced robots. Very expensive and they still have at least +/- 5 counts of "backlash".
IMO a harmonic drive from e-bay makes a good option (they are very expensive purchased new). Be aware that they are not all "zero backlash" units and mounting accuracy of the components is critical. They also take up more room than a worm drive.
DDR motors are great but they require very high resolution feedback devices and sophisticated controllers.
Worm gears work but you just have to live with some backlash in a fixed design or give up top speed and realize life will be shorten in a spring loaded design.
The life of any machine designer is full of tradeoffs. :)
Bob
I have been thinking of building a BL AC motor/servo for fun ;) But thats another story!Quote:
But if your determined to make it all yourself, I like the duel servo idea. But are you going to make the servo's?
Good points about the problems with the wormgear solution! I'm thinking of perhaps using something like in the globalspec link! Drawing will follow!
This is what i came up with, it has 60:1 gearing. I think it would cost about the same because now you can skip the angular contact bearings and just use bushings!
That's clever; I like the use of a single spring to preload two gear sets (two reduction stages) :)
BTW I think I've spotted another problem with the spring loaded worms:
Effectively one worm moves the gear in one direction and the other worm moves it in the opposite direction.
In one direction, when the main-shaft worm is driving, the friction on the spring loaded worm will act with the spring (in effect the spring will be dragging the worm around against the worm-gear friction).
In the other direction the spring loaded worm is driving, the friction will be acting with the spring, so as the spring is turned it is pushing the worm against the worm-gear friction increasing the load on the worm and thus the friction. If the initial friction is high enough (i.e. if the load on the main gear is high) and there is sufficient backlash to allow it to build, there is a danger that the friction will increase until it locks. The main gear will then stop until the back-lash has been taken up by the main-shaft worm.
It may be better to use a axial spring to pre-load the worms (rather than a rely on twisting the worm to generate the axial force).
Another thought:
What about a traction drive? i.e. a tooth-less gear-train. It's very easy to get high reduction ratios in a compact planetary drive design.
As long as you have a positional feed-back from the output (high resolution encoder?) Your servo can compensate for the slightly unpredictable ratio.
Hmm that sounds interesting! I got a bunch of 3600 ppr encoders, if i use them in quadrature with a gearbelt i should be able to get pretty good resolution! The question is how much torque you could get from the traction drive!?Quote:
What about a traction drive? i.e. a tooth-less gear-train. It's very easy to get high reduction ratios in a compact planetary drive design.
Great information, in my situation I have a 9" troyke rotary table that I am converting to cnc for gear cutting. My plan is to advance in only one direction and use a separate motor to actuate the table lock. this seems simple and easy. Any other Ideas on how to do it?
That's down to the material technology you use. It would be interesting to see if you could persuade a someone to let you have some their magic traction fluid for a home made device :)Quote:
how much torque you could get from the traction drive!?
http://www.gears-manufacturers.com/traction-drives.html
"Dry Traction Drive - These drives eliminate the need for lubricant and allow nearly 100% efficiency in power transmission. Slippage between driving and driven members is presented by a spring-loaded system."
As it is used for low speed and not as often as a mill spindle i think dry with feedback from the output is acceptable :)
Sounds do-able :)Quote:
As it is used for low speed and not as often as a mill spindle i think dry with feedback from the output is acceptable
I await your drawing with interest :)
I'm playing around in 2D now to find suitable dimensions/ratios now...
20mm sun, 40mm planet and 100mm ring should give 5:1!?
Then i'm thinking of using 3 stages for 125:1, and with increased witdh of each stage to handle more torque at the output end. How does that sound?
I found this page! Check a bit down under "Advanced rotational speed reduction traction drive device. 3D."
http://www.rotaryeng.net/turbo-compound.html
Ah yes - small world...Quote:
I found this page! Check a bit down under "Advanced rotational speed reduction traction drive device. 3D."
I occasionally make engine animations for Doug Self's 'Museum of Retrotech'. Somewhere on his site is page about compound IC engines. Reading that prompted research that lead me to turbo compound engines and thus to the site you linked to. It was that site where I first saw a home-brew traction gearbox and what I had in mind when I suggested it to you :)
A complete circle :LOL:
FYI Cone Drive uses a belleville washer type spring to preload a two piece worm against a worm-wheel in their "AccuDrive" zero backlash applications.
There is also a near zero backlash planetary gear set developed for NASA called a gear-bearing. This is a combination of planetary gears that also functions as a roller bearing in an individual element. It's one of those "Why didn't I think of that?" type things.
It's available on NASA's technology sharing site.
Dick Z
A flight of fancy :)
This gives about 56:1 using two stages of cones.
BTW. there is considerable aliasing* visible in the gif animation, which is why the input shaft appears to be turning slowly.
*( the same thing that causes wagon wheels to go backwards when being chased by indians ;))
I like that gear-bearing thing! Started to draw one, but have a little problem with the simulation!
The planets rotation around the Ground ring gear is a little off. Is this correct?
For a planet to rotate 1 revolution around the ring gear, it have to rotate 66/21 turns around its own axis.
For a planet to rotate 66/21 turns around its own axis, the input have to rotate (66/21)*(24/21) = 3,1429*1,1429 = 3,5920 turns?
I found a PDF with these numbers:
Sun: 24
Bottom Planet: 21
Ground Ring: 66
Top Planet:21
Output Ring: 67
For 251,25:1 reduction
The secondary planets and the output ring is missing until i get the first stage done correctly!
With your numbers;
Sun: 24
Bottom Planet: 21
Ground Ring: 66
the sun has to turn 66/24 times for the planet to make one orbit.
The planet will revolve 24/21 times for each revolution of the sun.
Does that help?
I made a simple test to try out some ratios.
Sun: 18
Front Planet: 24
Ground Ring: 66
Back Planet: 30
Output Ring: 72
Now i got the simulation to run as it should! :)
I belive the Sun to Orbit ratio is 4,889:1
But i have no clue how to calculate the Sun to Output ratio! But would belive it's around 30:1
The planet set orbits forward 18/66 teeth for each revolution of the sunQuote:
But i have no clue how to calculate the Sun to Output ratio! But would believe it's around 30:1
The planet revolves (backwards) 66/24 times every orbit
The output ring moves forward 72 teeth every orbit (imagine if the output gear was directly connected to the planet) and backwards 30 teeth each revolution of the planet (imagine the planet stationary).
So, the output rings moves 72-30 * (66/24) teeth each orbit.
The output rings moves 72-30 * (66/24) * (18/66) for each revolution of the sun
Simplifying we get : 42*18/24 = 31.5:1
Nice! Thanks :) Back to the drawingboard then ;)
Did just draw a 105:1 gearbox!
The housing is 100mm diameter and 45 mm long.
Very nice :)
In my last model i have:
Ground: 72
Sun: 12
Planet: 30
Output Planet: 24
Output Ring: 66
I just checked, it seem to move about 6,5 degrees in the same direction as the sun!