X Y Z reduction

[viz]

I am interested in reduction, was reading this forum but found that people are using only belt driven reductions.

Belt is elastic and it stretch, or it does not affect the cutting quality anyway on speed like (5m/min) ?

so why not using gear reduction by rack and rack ? For example put on motor shaft 12 tooth rack and reduce it by 36 tooth rack ?

Any help and suggestions why reducing by belt is better ?

Thanks

[KenC]

Because we do not know any better solution. :D
Why don't you enlighten us with non-belt type reduction at least at the same cost & effort to built.

[viz]

Ok i attached 3d model and screenshot, at this moment it is not calculated distances from pinions, but the main idea i think is clear.

Nema 34 motor, 12 and 36 tooth pinions.

Maybe somebody tried this solution ? Advantages and disadvantages of this reduction.

[KenC]

I don't see any anti-backlash mechanism which is very important in any position control mechanism.

[viz]

I have an idea how to make anti-backlash, i will make 3d model and upload it today.

[Zouave]

Good motion control belts have so little elasticity as to be essentially a non-issue, except perhaps at very high speeds with rapid direction changes. At which point I kind of doubt that your belt's elasticity is your largest loss of accuracy.

[viz]

We can solve backlash by adding 2 pinions, one is rotated by very small angle, you can see it in screenshots.

But the main problem here is that 12 tooth pinion is too small. We can rotate one pinion by shaft key.

It is not hard to manufacture it and it is cheaper then pulley, but if belt reduction is good enough for example 10m/min, i will reduce my axes with belt and pulleys.

Anybody who have belt reduction what are the maximum speed you are working without loosing accuracy on 90 degrees corners ?

[KenC]

This is interesting :)
I don't understand how you mount the 12 teeth pinion gear.

[viz]

I fixed pinions with shaft key, one pinion key hole are made 90 degree and another one 91 degree, it could be turned over, maybe it is enough 90.5 degree, need to test.

Also i added rubber between pinions for smaller backlash, i think this system is very strong and zero backlash, but it is my opinion.

Maybe somebody who knows better mechanical engineering can tell us about this solution, it will be good enough or it's time wasting ?

(in 3d model i didn't calculated center distande between pinions)
sorry for my english

[KenC]

I recon the two small pinion gear must have opposing force on the large gear in order to eliminate any possible backlash. I don't see that happening if they are tied together by a keyway.

[Gerald D]

See https://www.google.com/search?q=anti-backlash+gear&hl=en&safe=off&rls=com.microsoft:en-za:IE-Address&prmd=imvns&source=lnms&tbm=isch&sa=X&ei=FA0IUMmXDNKHhQe4sJHIAw&ved=0CEcQ_AUoAQ&biw=1430&bih=928

[viz]

Nice, so this idea should work. I will search tomorrow in my country for standard anti-backlash pinion gears. Let you know if find something.

Also waiting answers from people who have made belt reduction, and accuracy they get.

[Richards]

It's a very interesting design, but a little complex (at least to me).

Anyone could use his own MechMate to make a belt-drive. (I built my first belt-drive for my Shopbot using direct drive motors. That design worked as well as the other four designs that I made later after the machine had been retrofitted with belt-drives.)

A belt-drive can be easily adjusted to keep the proper tension on the belt and the pulleys. How would you adjust the gears to compensate for wear (and backlash), or did I miss something in the design?

[Zouave]

It seems to me that you will get more wear on an Anti-backlash nut or gear than you would on belt drives, given the nature of the interface, material and chips getting into it, and more potential for binding and uneven wear. That is my -gut- feeling, not necessarily a factual statement.

[darren salyer]

With what little run time I have on mine, I can say its plenty acccurate for what I do.
Have I measured it to 4 decimal places? No.
Am I completing projects that look great to the naked eye, that people are willing to pay me for? Yes.
I believe wood expands and contracts seasonally more than any innacuracies my machine may possess.

[smreish]

I have used that type of split pinion as an anti-backlash for limit switches, cam timers and many other devices that require little torque, but resolution.

I know they work well on a little 100 revolution limit box, but for a constant motion machine like the MM, I wouldn't bet the farm on it for long term.

[Gerald D]

I am interested in reduction, was reading this forum but found that people are using only belt driven reductions.

Belt is elastic and it stretch, or it does not affect the cutting quality anyway .... ?

so why not using gear reduction by rack and rack ? For example put on motor shaft 12 tooth rack and reduce it by 36 tooth rack ?

Any help and suggestions why reducing by belt is better ?

Thanks

Belt is not so elastic and nobody here is seeing a problem with cut quality because of the tiny bit of belt stretch. Therefore, it is a waste of time and money trying to build special gears when the belts actually don't show any problems.

[KenC]

Anti-backlash gears don't really comes with high torque rating & really not a cheap solution.
but then again, IMHO it is a possible alternative,
found this...
http://www.ondrives.com/data/pdf/gears/precision/abpsg/abpsg.pdf

Base roughly on our usual stepper motor torque figure 400oz.in (2.8Nm),
the smallest anti-backlash gear will be a Module 2.0 38 teeth (80pound excluding shipping), to make a 3:1 gearbox, the driven gear should have 114 teeth... for 4:1 152teeth.

[viz]

Thank you all for explaining ! I decided to use belt reduction.

[JasonC]

Dont forget theres planetary reduction too.

[viz]

Yes but the price for planetary reduction is about for one axis 200EU

[Richards]

Let's revisit why we use stepper motors.

1.Stepper motors are open loop. They don't need feed-back, i.e., no encoders.

2. Stepper motors are simple. Send a pulse and they move a specific distance.

3. Stepper motors are low cost. You need an appropriate power supply, a stepper driver, a pulse source and a motor.

4. Stepper motors can be driven direct drive, i.e., 1:1, or geared. Geared means either belt-drives or gearboxes.

My Shopbot PRT-Alpha was direct drive. I retrofitted it with several different belt-drives and finally bought a Shopbot update that included geared motors. The direct drive model worked fine except for circles and curves. It had excessive chatter. The belt-drive models also worked fine (3:1) and greatly reduced the chatter on curves and circles. The 7.2:1 gear box was almost identical to the belt-drive in cut quality. I didn't need it, but I wanted to see if it would improve the cuts over the belt-drive. I don't think that it did. The quality compared to the belt-drive was identical as far as I could see.

What I learned is that any of the three methods work.

I would prefer either a belt-drive or a gearbox to reduce "chatter".

A 600 oz*in motor works well with direct drive.

A 300 oz*in motor works well with belt-drive or gearbox.

The power supply is not overly critical.

Wiring the motors half-coil or parallel works. Wiring the motor serial works for low speeds only.

Sometimes we want to find a perfect solution when the paradigm is imperfect. Stepper motors work very well when you keep within their parameters. Power supplies for CNC machines are not critical. Use the Geckodrive formula (32 * SQRT(inductance) ) to get close and then don't worry about it.

Use a direct-drive motor if your quality requirements are not too critical. Add a belt-drive or replace the motors and use a gearbox is your quality requirements are higher.

This is not rocket science. Lives are not at stake. A $30 sheet of MDF is not going to ruin you financially. (If it is, you need to charge more because operator mistakes happen even when the machine works perfectly.)

If you use the wrong motors and stepper drivers, you can fix it later on at a relatively low cost.

[viz]

Thank you Richards.

What timing pulley are you using for reduction, 3mm pitch or 5mm pitch ? Because here in my country we need to wait for one month to get 3mm pitch pulleys

[bradm]

I suggest that you choose your timing pulleys and belts based on local availability, fit, and price. As long as the pulleys and belts match each other, fit on your shafts, and are of reasonable size you'll be fine - and happier if you ever need to replace them.

[viz]

Yes but sometimes it's better to wait :) Maybe 3M timing pulley is more accuracy than 5M timing pulley in reduction ?

[bradm]

No, 3M timing belt allows for smaller pulleys and higher speeds. However, the difference doesn't matter for a mechmate usage. Take a look here (http://www.roymech.co.uk/Useful_Tables/Drive/Timing_belts.html) or google on "timing belt differences" for more information.

[Richards]

I used the XL size pulleys and belts (0.200" pitch). That size seems to be a perfect fit for the motors that we use (1/2" diameter shaft). You can use the 18-tooth pulley on the motor and the 72-tooth pulley on the shaft for 4:1 reduction.

Some people use even smaller pulleys on the motor, but then you would have to drill and tap the set-screw holes through the teeth of the pulley because the shoulder of the pulley would be too thin to reliably hold a set screw.

(The MXL size pulleys and belts (0.08" pitch) are frequently used on process control applications when small motor driving small loads are common; however I doubt that they could handle the torque required on a MechMate.)