Can gears with ZERO backlash work properly? 2

[makary]

I have an application that transmits large torque but requires ZERO backlash.
During each gear revolution torque direction changes several times. This generates large shock loads and noise due to backlash. The gear set is internal-external gears.

I intend to eliminate backlash to zero, but don’t know if this will allow gears to work. Any input? Experience in this matter?

How to eliminate backlash is another problem. My first try will be semi-bevel gears achieved by continuous profile shift along teeth. Oposite continous shift on internal and external. This will allow to keep axis parrallel. If this proves too difficult, just standard bevel gear assembly with parallel axis (opposite bevels), and small – about 3 degrees taper on teeth. A spring will push one of the gears axially as to contineously eliminate backlash. My concerns in this solution are (1) will gears work with zero backlash (2) wedging effect due to small 3 deg bevel angle that could potentially generate large friction or even stop gears from rotating.
Any word of experience here would also be appreciated.

 

[MikeHalloran]

Regular bevel gears will jam together at zero backlash.
With a really shallow bevel angle you might have to press them apart. It doesn't sound like a winner of an idea to me.

You might try moving the external gear radially inward a bit, and inserting two idlers between it and the internal gear, and spring load the idlers toward each other.

... sort of like this gear drive:

http://www.jegs.com/p/Pete+Jackson/748317/10002/-1

--- except, imagine that the tie bar between the idlers is a little short, and springy.



Mike Halloran
Pembroke Pines, FL, USA

 

[makary]

Thank you Mike,
I dont have room for additional idlers. However, if spring loaded idlers work, that means gears can work with zero backlash. I hope tooth tips don't require special modification. I also hope the efficiency is still fine.
I know of split gears with preloaded spring between halves to eliminate backlash, but those are usually for timing and don't transfer large torque.
As for bevel gears - I would not mind having to press them apart as long as they don't jam and still rotate without power loss. On the other hand - if they rotate, it should not have to take much force to press them apart.
I know of shalow bevel drives with parallel axis, but backlash is minimized there by washers, not by springs. Can't do it either beacause I can not assure precise axial location of either gear.

 

[MartyMan]

Cone Drives does make power gearboxes with split/preloaded gears...

http://www.conedrive.com/servozerobacklashgearsets.php

 

[tbuelna]

makary,

The only types of gear drives that can operate successfully without backlash are those with dual drive paths that are elastically pre-loaded in opposite directions. The types of gear arrangements (among others) that fit this description are scissor gears and harmonic drives. Normal involute gear meshes require some degree of backlash, due to the realities of manufacturing tolerances, assembly variations, limits of structural stiffnesses, and distortions due to thermal effects.

Strangely, you said your system requires "zero" backlash, but at the same time you would accept the elasticity of having springs in the system. If accurate angular indexing is your primary concern, backlash and low torsional stiffness would present similar problems.

You didn't describe the specifics of your mechanism, but if you require less than a full 360deg rotation at the reversing output, you might consider something like a motor driven roller or ball screw (which have a high effective gear ratio, low friction, and can be rigged for almost zero backlash) driving a connecting rod and crankshaft linkage.

Hope that helps.
Regards,
Terry

 

[makary]

Thank you Terry;
Unfortunately I can not use any other gear arrangement but the one I described. This gear arrangement is the main component of a special type of rotary compressor (so called cat and mouse design). It allows acceleration and deceleration of vanes - several times per revolution.
The gear arrangement is: one stationary ring gear, and one pinion spur gear that meshes with the ring gear. Torque reversing happens 4 times a revolution of the pinion gear. This generates 4 large "bangs" every revolution. Pinion gear rotates in one direction because of a prevailing torque in that direction, and the effect of flywheel that the vanes have.

I was considering splitting the pinion and preloading it with heavy spring, but this have two major disadvantages:

1) The already long pinion will get twice as long, because each half has to be able to transfer very large torque (500 Nm average, 1500 Nm peak) - each half in one direction.
2) Low efficiency. Due to the very large preload the half that does not transfer torque at the moment, still generates very large friction, almost as large as when it transfers the torque. That's what I think.

I think this solution is suitable for relatively low power transfer, such as valve trains, but not to transfer 100% of engines torque where efficiency is critical.

So my hope was to make this low angle opposite bevel gears work, where a spring would continuously cancel backlash between the gears by pushing one gear axially. This arrangement results with parallel gear axis. However the low (about 3 deg) angle may not allow gear to be pushed back should the fit become momentary tight due to accuracy of angular indexing as you pointed out.

Instead of spring, I could resort to manually adjusting backlash to almost zero, but this is expensive in production and requires adjustments every so many hours of operation.

I hope this explanation makes it clearer. I'd appreciate you comments, especially if you have any experience in effects of zero backlash on heavily loaded gears.

I uploaded a picture here. I hope the link works:

http://files.engineering.com/getfil...d5d-c5fbda81e47a&file=Gears-zero_backlash.png

 

[Spurs]

Makary
Have you done a calculation on the torques and forces to determine if it is even possible to install a spring with sufficient stiffness to maintain "zero" backlash?

In additon, depending on the speed of rotation of the gears, the springs may not have sufficient "time" to actually mechanically move to make the adjustment to "zero" backlash.

 

[plasgears]

Surprisingly, elastic deflections and other clearances of all parts in the gear train can allow zero backlash operation. This is particularly true of plastic gears and gearboxes.

 

[makary]

Thanks Plasgears. This is very encouraging.

To answer Spurs: yes it is possible too install a spring with sufficient stiffness. The problem always is to install a spring that can produce huge force but has sufficien SOFTENESS. The stiffer the spring the easier it is to design and package it. Output force tolerance is another issue but...
In this case there is no need for the spring to make any "adjustments". Each half of the gear is always in mesh (in contact) with other gear.

 

[Clyde38]

If indeed you achieve a truly "zero backlash" gear train, will it be "zero backlash" for very long (increased wear)?

[link

http://www.linkedin.com/in/clydehancock

]

[/url]

 

[makary]

If done the way described earlier (opposite bevel gears) - it should be zero backlash forever as spring will compensate for wear (pushing axially).

 

[Grunchy]

Check out Ikona Gear, they have a patented non-involute gear design that features zero backlash. Here's a technology page with a little movie that shows how it works:

http://www.ikona.ca/technologies.shtml

Smart!

 

[Grunchy]

You know, thinking about it overnight, is it possible to incorporate something like a one-way clutch? You are driving a load that overruns its drive 4 times per rotation. Perhaps that's a good thing -- just let it overrun by the clutch slipping. That way you avoid having to do something like run zero lash or use an expensive proprietary solution.

 

[makary]

No I can't use a clutch. It's a special mechanism with components that must be rigidly connected between each other. The only "lose connection" is this backlash.

I'll call Ikona in a moment and will let you know what they say. It could be what I am looking for if it works for larger gear ratios than they spec.

 

[makary]

I just called Ikona. The number is no longer in service. They may be out of business.

 

[makary]

I talked to Ikona rep. Their tooth profile may be the answer. I've seen things before that did not deliver on their promises, so I'm going to wait till I can test it. There is however some merits to their claims so I am hopeful it's gonna work.
What concerns me is the efficiency that's somehow lower than the revolute gears. Always a trade-off.

Thanks Grunchy!

 

[MikeHalloran]

The video on the Ikona site makes it clear that the tooth to tooth contact is basically 100 pct sliding, with virtually no rolling. Hence the efficiency.



Mike Halloran
Pembroke Pines, FL, USA