[andy_ck87028]

Hi All,

I don't have an engineering background so at times I have some pretty basic but to me, fundamental questions to answer.

I can see that there cannot be introduced backlash with gearing through a timing pulley belt and very clearly backlash can be an issue when meshing two spur gears. What I cannot conceptualize is whether a spur and worm gear combination can have intrduced backlash or whether helical gear arrangements have introduced backlash.

I guess my primary question is: what gear arrangements have no backlash problems and as a subsidiary question (possibly already answered in the first part), do worm and spur combinations have backlash problems?

Finally, am I right in my assumption that timing belt pulley gearing arrangements have no backlash problems?

Although theorists might say that backlash cannot always be eliminated, I am looking for answers in practical terms.

Happy CNCing

Andy

(Go Lleyton Hewitt!!!!!!!!!!!!!!!!!!!!!! :boxing: )

[Al_The_Man]

The gearboxes I tend to use are low backlash planetary, 10 or 5arc-min backlash there are several manufacturers but quite a few Bayside come up on ebay.
Al

[HuFlungDung]

Worms only run properly with a worm wheel, which is different than a spur gear. A spur gear lacks the helical twist to the tooth which makes for a proper mesh with the worm. But you are correct, it is impossible to eliminate all backlash through precision machining and mounting of the gears alone. Worms create sliding friction against the worm wheel which causes some degree of wear immediately, if you attempt to run them "tight" together. Heavy oil lubrication is required, and this, in itself, will create some degree of sponginess to the mesh.

As Al pointed out, there are precision gearboxes with low backlash available, but there is always some.

The best way to eliminate backlash is to place the feedback system somewhere beyond the gear train, such as with a linear scale on the sliding elements, if you are translating rotary motion into precision linear motion. Or, mount a rotary encoder directly on the ballscrew, after any gear train.

Timing belts may have some backlash, depending on the style chosen. Its harder to see it, and is probably negligible on a system that moves freely. But heavy loads can make the belt slip a tiny amount from one side of the belt tooth, to the other.

[andy_ck87028]

Thanks guys for that feedback.

Just so that I have a direction that I can adopt, is it please possible to open open further on one of the points made in your answers:


The best way to eliminate backlash is to place the feedback system somewhere beyond the gear train.....

Is there an example of a feedback system that I can draw upon? I am planning to use a timing belt pulley but I think I also have to supplement it with a simple worm drive arrangement.

For the sake of reducing to basics, lets assume the gear train is two cogs only. What feedback system alternatives can be put in place beyond such a simple gear train? Alternatively, is there a URL available that spells it out in simple terms?

Thanks so much for your time in answering these questions. I appreciate how much effort you guys put in for beginners such as myself.

Andy

[HuFlungDung]

Andy,
You have not told us what kind of a system you are thinking of building. If you are using a stepper drive, then these most typically are "open loop", meaning there is no feedback provided to your computer that the system has made a move correctly.

The computer sends pulses to the motor, expecting that they are perfectly executed. Backlash in such a system could be factored out by estimating how many steps are required before motion begins to execute correctly in the opposite direction. This is not a true elimination of backlash, but merely is compensation for it. Mechanical loading of some sort (gravity, springs, etc) can sometimes also work to always keep the load against one side of the gear train, and eliminate the effects of the backlash.

The only risk is that at some point, Murphy's law will execute, and the mechanics of the system will change and cause the backlash to suddenly appear (a slide begins to stick, or a tool lifts the part unexpectedly across the backlash distance).

A servo system using encoder (linear scale or rotary disk) feedback to the control operating in "closed loop" mode means that the control checks that the movement commands were obeyed, and adjusts the machine position by commanding further movement until the scale readings jive with the commanded position. So, if the encoder is reading the motor shaft angle directly, then the backlash in the output gear train is not compensated, whereas, if you take the encoder off the motor, and mount it directly on a ballscrew, then the angle of the ballscrew can be accurately detected. The motor will then be commanded to rotate the gears until the ballscrew is correctly positioned, regardless of any backlash in the gears.

[Al_The_Man]

I am planning to use a timing belt pulley but I think I also have to supplement it with a simple worm drive arrangement.
Andy

Andy, I am curious to know why you feel you have to combine both timing belt and worm drive?
Also, unless you are using high rpm servo's, you may be limiting your max feedrate, unless you have already factored this in. As Hu said, it would help to know what you aim on doing.
Al

[andy_ck87028]

Thanks Hu / Al

What you say makes sense and I'll see if I can answer.

I previously decided to go stepper and have bought the driver card and software and so I am committed to an open system. Installing a servo with feedback control on the linear bearing makes perfect sense but .... not this time for me now.

So, being an open system means managing backlash which I want to do without having to resort to using stepper compensation but by keeping backlash out in the first instance.

I am driven by price and ease of construction and so I want to reduce the number of bearings and gears. I also want precision accuracy. My preference given those criteria was to use threaded rods (very cheap) and fixed at both ends (no bearings and easy to construct). I also like timing belts as there is minimal backlash. Given that design criteria, I am therefore committed to having the drive assembly actually on the gantry spinning a nut that when spun causes the gantry to move.

I can easily spin that nut using a simple timing belt pulley but that raises 2 problems for me.

1) I am cautious by nature and feel that gear reduction is required as I feel that gearing down will give me increased precision and control. I realise I will sacrifice feedrate but currently that is not important for me.
2) I have to replicate the assembly at the other end of the gantry to remove skew and therefore have to either use another threaded rod (cheap) or a very long timing belt (expensive and messy).

Given those problems, I would like to use a threaded rod with gear reduction. On the basis that I am employing simple gear reduction and as the two threaded rods at either end of the gantry are at right angles to the interconnecting threaded rod on the gantry, I was wondering if a worm drive might be the answer.

That was where my original question came from with respect to wondering if worm drives have an inherent backlash problem.

So, from the contributions I have got from this thread (thanks and thanks again) I feel that I have to now modify my design as worm drives seem too problematic. I think I now have two options shown below as A and B. Neither option should have any significant backlash issue.

The rack and pinion will be through a timing belt pulley and a timing belt laid flat and anchored at both ends. (Diagram B').

Andy

"To do or not to do; Try does not exist" (Yoda)

[ynneb]

Andy, do yourself a favour and download this. http://cnczone.com/modules/Downloads/userfile.php?id=4
It will make your life so much easier. You dont have to reinvent the wheel, just follow the pre beaten path to achieve results.

You have told me what you want to cut, this machine would be perfect for that.

[ynneb]

EDIT: For some reason the post was uploaded twice. Moderator please remove this post since it wont allow me to

[andy_ck87028]

Thanks Benny for the suggestion but I'll think I'll skip this time. I'm having far too much fun trying to work it out for myself.

I always have and always will get exceedingly bored with copying as it never gets me closer to the truth. To me the real challenge is about achieve understanding and innovation not about following instruction.

Even though I am probably re-inventing the wheel, believe me it is still free thought and discovery. Discovery is my catalyst for innovation and I love putting the effort into understanding the underlying problem to be solved.

My CNC router will definitely have my stamp on it but it will also have both influences and improvements from other designs such as yours Benny and the machine you have just pointed me to.

Andy