Gear Alignment Question 2

[Auspex]

Hello,
I have designed my first gear/pinion set for a rotary table. I have a large 350 mm pitch diameter 0.4 Mod ring gear and and an 88 tooth antibacklash pinion will be driving it. My question concerns alignment. I am designing a mounting assembly and would like to know what type of tolerance would be appropriate for alignment of the centers, pitch, yaw, etc for proper funtioning. This will help me design for a properly stiff frame. Thanks for your help!
Shelby

 

[diamondjim]

Shelby,
You are using an 88 tooth anti backlash
pinion with an 875 tooth gear. You should
state the gear face and pinion face widths
as well as the materials. Obviously the
more exacting you can set these up, the
better life you can expect. I assume these
are metal parts and lightly loaded. Sounds
like an interesting application.

 

[Auspex]

Thank you for the prompt response. 303 Stainless parts with a max of 30 N-m Torque during short duration move profiles. Face width on the ring gear is 10 mm and 7 mm for the pinion. They are driving an air bearing stage where micron/nanometer level precision is crucial. The components to which each respective gear is mounted are very stiff, but I have limited experience with gears and are trying to understand how aligment errors will affect overall performace.
Thanks for the advice.

 

[israelkk]

First, it is a good practive to have the pinion width larger than the gear width. The pinion is ususlly the gear with the higher stresses.

Second, calculating the pinion using Lewis gear strength formula show that the tooth bending stresses are almost equal to the yield strength which is way too high and there is no safety factor even if we assume 7 mm pinion width. More than that, if the anti-backlash gear is a split gear (meaning two adjacent gears of 3.5mm width each) the situation is much more severe because, only half the pinion width actually carry the torque load.

Third, I do not know of any split anti-backlash gear that the springs between the two halves of the pinion can carry 30N-m therefore, actally you will have backlash. Those types of anti-backlash gears are not designed to take load and the are used to rotate sensors such as potentiometers resolves etc where there almost no rotating resistance.

 

[Auspex]

Israelkk,

Good catch and thanks for the feedback. The 30 N-m move is a conservative estimate that includes a fudge factor for the inertia values when I designed the gear. The masses are actually considerably lower, and the effective torque to achieve the desired move profile is around half of this.

A matching face width for the pinion is all that I have room for, but I will re-check the bending and Hertzian stress calculations to ensure that I still have an effective margin of safety.

As for the third point, I am aware of the limitations of an antibacklash pinion. I am not using a convenitonal spring,or a spring at all for that matter The design will have the ability to modulate the spring rate of the antibacklash pinion. In reality, I am not as concerned with backlash on the large force moves so long as I have the ability to settle in and fine tune the position to within a micron without backlash.

Do you have any advice on my original query concerning alignment?
Thanks again!

 

[israelkk]

I am used to AGMA standard and for precision gears it is customed to at least use AGMA Q10C accuracy which for minimal backlash should result in the center distance tolerance of +/- 0.025 mm. It is best that both pinion and gear axes are to be machined at same milling process in a single base plate where the axes should be parallel to each other within someting like 0.01mm. Any good CNC miling machine will fit for this job.

 

[Auspex]

So is the primary advantage of going from, for example, AGMA 10 to AGMA 11 a reduction in backlash? Any other benefits?

 

[israelkk]

The more higher AGMA number the more accurate the gear. However, the price goes higher, much much higher. As a result backlash can be smaller. However, the center distance tolerance should be smaller too to really reduce the backlash meaning, a use of a very accurate milling machine such that used for fixures building.

One more issue is that as the AGMA number goes up the hobbing process will no longer fit and the gear should be grounded, polished etc. However, small module grounding is a problem if not impossible.

 

[sreid]

Having some experience with air bearings and precision motion control, I wonder if a direct drive frameless brushless motor might be more appropiate. They have become moderately priced and eliminate all kinds of mechanical problems.

 

[Auspex]

I tend to agree with you. Air bearings paired with a brushless direct drive motor are a marriage made in heaven. It saves lots of headache associated with mechanical solutions like material compliance, stiction, and backlash. In fact, this stage will be stacked on top of a linear air bearing that is actuated via a direct drive linear motor for these reasons. However, there is one particular challenge associated with the rotary stage that lead to the decision to use a slotless brushless DC motor to drive this stage via a pinion and gear. Direct drive ring motors in this size range (~13" ID) tends to be more massive than can be tolerated for the rotary stage. The application requires the rotor components for this design to be oscillated axially at high frequencies. This becomes exponentially more difficult with large masses.

 

[sreid]

Auspex,

I hear you loud and clear. We have a fast tool servo driven by a voice coil that weighs 6 lb. and runs at 15 G.

You can reduce the motor rotor inertia significantly over conventional designs by

1) Making the interior out of aluminum.

2) High pole count motors can have quite thin magnets thus reducing peripheral weight.

3)The thin magnets do not need much back iron so the iron rim under the magnets can be reasonably thin.

 

[metman]

Auspex,
To control center distance accurately, you could go to a toolmaking shop that has a jig-boring machine. This will hold tolerance of +-0.0002" if a jig-grinding attachment is used on the spindle.

This still leaves you with what israelkk brought up "...However, small module grounding is a problem if not impossible." What about hand scraping and lapping? This would undoubtedly be expensive but if you really need it, cost might have to take a back seat.