Custom Gear Boxes

As a part of this years off season I was looking into making custom gearboxes in the off season. But I have small problem, I don’t know much about gear boxes, I know the basics from putting AM gear boxes together and making custom vex ones.

But how do I plan one out?(other than copying a hole pattern from an AM one) Lets start simple, I want a single speed gear reduction lets say 8.5:1. Now where should i go to look at custom gears? How do I cad the gears and there “meshing” with out a gear generator?

All your knowledge you could share would be very appreciated.

You have to start with some basic knowledge about gears. There are probably papers here on CD that talk about gears.

Let’s start with your example: 8.5:1 in a single reduction. If you look at MartinSprocket.com gears, just for an example and 20 DP, 20 deg PA gears (I didn’t explain what DP and PA mean.) The smallest gear listed is 12 teeth and it has a 0.6 inch PD (pitch diameter - the diameter at which this gear would run with its mating gear.) A mating gear of 100 teeth will get you pretty close to 8.5:1. The pitch diameter of the 100 tooth gear is 5 inches. The center distance between gears would then be (0.6+5.0)/2 = 2.8 inches. This would be the distance between gears. However, there are some practical limits. The 12T gear has a maximum bore of only 5/16 of an inch - this would be a pretty small shaft. The OD (outside diameter) of the 100T gear is 5.1 inches which may be bigger than the wheel you want to use if you are direct driving off the 100T gear. Obviously, there are lots of things to be concerned with and I just touched on a couple.

As far as CAD goes, you don’t need to model the gear teeth - they look really good in the model but don’t add much real value. You need the outside diameter, the bore (and how torque is transmitted), the thickness, and the pitch diameter.

For starters, do some research on gears and find the basic formulas - they aren’t too difficult. Build yourself a spreadsheet that does the calculations you need and after that work on the CAD side of things. You can download a spreadsheet that does these calcs but I believe you’ll learn more by creating the spreadsheet yourself.

I’ll add a little to what Tom said–those formulas (and the definitions for PD and PA) can be found in a machine design textbook (check a college bookstore, or somewhere similar).

One other thing to be concerned with, at least a little bit, is the width of the gears in the gearset. Wider gears can be less susceptible to breaking (due to somewhat lower bending stresses in the teeth), but obviously weigh more. Any 2 mating gears should be the same width–not doing that increases the stress in the wider gear.

General process: Find the total reduction, then figure out any intermediate steps (may be forced due to physical constraints). Then figure out DP, PD, and PA, because you’re going to want them for some of the other calculations, like width and “is this going to break under normal conditions”. The next thing you do (after running the aforementioned calculations) is to find your gear or gear stock (or a substitute, if that size isn’t available). Hole patterns and CAD are near the end of the chain in terms of when you want to deal with an item–right around the light-weighting stuff.

I would highly recommend using the search button, its very very very handy.

http://www.chiefdelphi.com/forums/showthread.php?t=95413&highlight=custom+gearboxes

One of the threads that I quickly found, great info. Look around on CD before putting threads up.

-RC

Is that strictly correct? I would say that doing so changes the stress distribution (because the non-meshing parts of the gear teeth will resist bending). While some stresses may well be locally slightly higher than in the same-width case (I’m thinking right at the edge of the meshing tooth), on balance, the amount of stress at a given point in the wider gear ought to be lower, thanks to the contribution of the extra material.

When I design gears, one gear is almost always a bit wider than the other. There are two stresses to think about. The first is bending stress and the second is pitting stress. The bending stress is (or may be) different for each of the gears in the mesh. The pitting stress is the same for both gears since it is the result of how the two gears contact together. Due to the stack-up of tolerances in a gear box or transmission, there is axial error between the two gears in a mesh. In order to guarantee a given contact face width, one gear is generally made a bit wider. For bending stress, a rule of thumb is that up to 1/2 a whole depth of excess face width counts toward the gear face width and thus reduces bending stress in the wider gear.

I agree. I think a more accurate assessment would be that you can absolutely make one gear wider, but only the contact width will provide the strength you need, so making the gear wider than the contact area does not allow you to make the gear weaker (or maybe it does ever so slightly, but that’s probably not a good way to do the design…). I’m not seeing how increasing the size of one gear ever increases the pressure on it…

To cite reality, CIM pinions are almost always [strike]5/16"[/strike] 5/8" face width, while the standard 20dp gears (both from AM and from lots of other suppliers) have a [strike]3/16"[/strike] 3/8" face width. This difference means you can still line up the gear and pinion after pressing the pinion onto the CIM shaft with little to no accuracy in terms of how far it’s been pressed on. Edit: heh, not sure what I was thinking… sixteenths are small…

To the OP:

For the gears, there are just some basic things you need to know, all of which can be find on suppliers’ websites:

  1. the pressure angle
  2. diametric pitch (DP) (these 2 need to match)
  3. outside diameter
  4. pitch diameter
  5. face width
  6. number of teeth
  7. bore size and key

The pressure angle is the angle between the contact plane of the teeth and the line out from the center of the gear to the point of contact. The diametric pitch is the number of teeth per inch of the gear’s pitch diameter, which tells you (with circumference calculations) the distance between teeth. All you need to do is make sure these match and are suitable for your application. For FIRST purposes, 20 DP and 14.5º pressure angle is just fine.

The outside diameter and face width will just tell you how much clearance you need for your gear.

The pitch diameter is the diameter at which the teeth interact. If one gear has a pitch diameter of 1" and therefore a pitch radius of .5", while the other has a pitch diameter of 3" and therefore a pitch radius of 1.5", their centers must be 2" apart. That should be pretty straightforward once you know how it works. Just for fun, if these gears are 20 DP, then the first one has 20 teeth and the second has 60.

The number of teeth then tells you the gear ratio. In the above example, 20:60 simplifies to 1:3. So your gear ratio is 3, 3:1, 1:3 however you want to say. You just need to keep straight which way it goes. The bigger gear always turns slower but carries more torque. Conversely, the smaller gear turns faster but carries less torque. The gear ratio is exactly the ratio between these speeds and torques (ignoring inefficiencies).

Finally, the bore size and key type are important for making sure you have the shafts you need. You can always make the bore bigger, but it’s awfully hard to make it smaller. The key can be an ANSI keyway or hex key, both of which you can do yourself with a broach (don’t actually know much about that).

As for finding/making the gears. Aside from Andymark’s (somewhat limited) selection, you can find a lot of stuff at McMaster Carr, Grainger, and Martin Sprocket, if you look in their catalogs, which provide all of the above information. As for making them yourself, you’ll have to ask someone else. You can probably find information about the tooth profile in aforementioned machine design textbooks. I know 148 did some cool stuff with sheet metal, and there are probably other ways to do it too (wire EDM, others?). And who knows how pricey.

Well that was kind of long…