Worm Gear Gearbox

Has anybody on a FIRST Team every tried to create a gearbox using worm gears. I would like to try to do this, it seems very interesting to me. Even tough they are not always very practical for drive trans it seems as tough that it might have a competitive advantage. Since when you cut the power it will not roll back in the opposite direction. If anybody has any questions, suggestions, or any other usefull information please post it here. :smiley:

258 used worm gears to power their arm with two CIM’s this year. They have some pictures posted here on CD you could look at, and I’m sure Bill Gold would be willing to explain what they did to you if you asked.

I did this as an experiment using the plastic gearboxes, shafts and axels suplied by first and purchased gears - there is a post somewhere of someone else who did this. The Chippawa moters fed into this I think it was a 30 fold reduction - so the speed was about 4fps. So comparable to a Fischer Price, but much more powerful, Unfortunately my entire robot platform was thrown out in a fit of spring cleaning - the gears cost about 300 bucks for two sets.

I had some pictures, put heck if I can find them.

SO worm gears will work, and you can make the transmissions with just a good drill press if you are carefull and precise.

I got the gears from “Boston Gear”

worm GH1076RH gear G1071RH

Like Cory said earlier, the team I was on last year (258) used a CIM to power our shoulder joint, and the other CIM to power the elbow joint in our arm last year. Each CIM ran into its individual gearbox, and then powered their joint via #35 chain/sprockets. In each gearbox the CIM was reduced with spur gears at a 3:1 ratio, and then reduced another 60:1 with a 60-tooth worm gear and a single thread worm (worm ratios being # of teeth in the gear / # threads on the worm), and then there was a sprocket reduction of 4:1 to the joints.

The cast iron worm gears didn’t crumble, and the only problem we had was due to some “over-engineering while under-thinking.” Long story short, we use nuts to hold the worm in place on its shaft, but didn’t think about the nuts loosening up due to the rotation of the shaft (which they did). So, after a while the nuts would back up, and when we tried to rotate the arm with all of our robot’s weight behind it (like trying to “chin-up” to hang) the worm would slightly slide out of place, and end up stalling the gearbox/motor. If we had just machined an aluminum spacer instead of using those stupid nuts, that arm would have perfectly accomplished all it was designed for. Oh well, live and learn…

Our worm gears were ordered from Martin Sprocket…
10DP 60-tooth worm gear (W1060)
10DP Single Thread Hardened Worm (WG10)


I’ve also seen many other uses of worm gears in drivetrains and other robot mechanisms. They can be great for situations where you don’t want backdrivability, or for a huge reduction in a relatively small amount of space. As with any gear-to-gear setup, be sure you have the ability to machine your gearbox fairly accurately, or else you’ll get either binding or huge backlash (or no gear contact at all). If you have any questions about these gears I’d suggest checking out an informational .pdf that Boston Gear has (http://www.seadawgs.com/GT.pdf), or go into Borders/Barnes & Noble and skimming through (or just buying) The Machinery’s Handbook. There’s tons of great stuff in both of those.

I hope this helps.


I will vouch for the workability of the worm gear gearbox in a drivetrain. In 02 my high school team tested it. Just make sure that you pay attention to the forces that your gearbox will be subject to, and lube like a madman. The forces don’t work out the same as with spur gears as there is more axial force on the worm gear and worm, so putting a large amount of force through a worm gear gearbox can cause an improperly considered structure to flex and bind even if you were precise in your machining. (it can still work just design the box structure and shafts to take the forces)

Also make sure that you check the torque ratings on your gears. The fact that they don’t backdrive also means that much of the shock of hits goes soley to the teeth of the worm gear.

We also got our gears from boston gear.

Last year we used worm gears to power our shoulder mechanism for the first part of the UTC regional. We removed the arm primarily because it unbalanced the robot.

We adapted the FP motors to go into the Bosch transmissions. On the output of the transmission we had a six point socket attached. The socket interfaced with a bolt that the worm was welded to. The worm assembly was setup with 4 bearings, 2 flange type radial and 2 thrust. The worm wheels were soid mounted to our shoulder shaft and the motors and transmissions were mounted to the arm. It worked well for moving the arm and locking it in place with no back drive, but alas few ever saw it work properly for many reasons. :frowning:


Team 107 also used a worm drive transmission a few year ago,it work ok but there are things to be aware of. The main issue is your reduction should not exceed 50 to 1 on the worm ,for the reason that it will not back drive or coast. You can compensate for this by adding a ramp down in your code, but in humble opionion worm tranmisson is better suited for arms or lifting or any application where there are not sudden loads ,like stopping a moving robot.
I may still have one of the transmission we used if you would like me to sent it to you to look at I would be happy to help. I will try to post some pictures of a new worm tranny I designed over this summer If anyone would like to see it.

thank jim schaddelee

team 107 Holland Christian / Metal Flow

I would love to take a look at that worm tranny. If you do not mind sending it to me. You can find my email in my profile.

I don’t want to jump in and be a nay-sayer… but I think one of the things you need to really weigh in before considering a design using a worm gear is the immense efficiency loss you’re going to have simple due to the nature of the gear set.

We’re talking in the range of efficiencies as low at 30% :ahh: !

The advantage you mentioned, about not being able to back-drive the gear set, is not always true. There is some critical pressure angle and coefficient of friction that leads to the point were you actually can back-drive this gear set. Jim said that it’s above a reduction of 50-1. I can’t verify this either way.

[edit: But Paul definitely can. Rock on!]

However, simply because of the efficiency issue, I would steer clear of them in a drive-train gearbox unless there’s a really good engineering justification. For a large reduction in a small amount of space, an off-the-shelf planetary gear set would probably hold better efficiencies.

That’s my two cents. Good luck :slight_smile:


Back Drive-ability is a function of surface finish, lubricity, and worm lead angle. Using a bronze worm gear and a hardened steel worm (the small screw looking one), the max lead angle is 5 degrees to insure self locking. This is not a guarantee! What is a guarantee is that above 5 degrees, the worm gear set will not lock. The lead angle is the reason the window motor locks and the van door motor does not lock. It really has nothing to do with the gear ratio. A gear ratio of 50:1 with a lead angle of 15 degrees will not lock.

However, the more thread starts on the worm, the higher the lead angle and the lower the gear ratio is. Remember, gear ratio for a worm gear is # of teeth on the gear divided by number of thread starts on the worm.


a little off, but for lego, wormgear 24:1 reductions work very well, for lego robots i mean. not too slow, yet (well since lego isnt efficient whatsoever,) efficient (make sure you wupport both axles from both directions, and since lego axles are plastic, support them as close to the gears as possible.
well, yea. worm gear gear boxes (wormgearboxes?) as an idea are very applicable ithink. except as with designing any kind of gearbox, will have difficult parts but if it works well, great!

…and yea, im understanding why my english teachers rarely give me any good grades on my essays.

Actually efficiency might not be an issue if you properly design the gearbox. A little bit of engineering and you should be able to get the efficiency fairly high. Of course a little research shows that the efficiency can reach 60%.
Ps. Please someone build a drivetrain using these gears. These are the funniest sounding words ever. Hypoid and globoid.

60% efficiency is still downright terrible for a drivetrain when you consider spur gears are 98% efficient and chain is 95% efficient

Here are some facts about worm gear efficiency:

Efficiency = Tan(Lead)(1-fTan(Lead))/(f + Tan(Lead))

Lead = Worm Lead Angle
f = Coefficient of Friction (Mu)
For a bronze gear and a steel worm f = 0.05
Easy to purchase worm gears have lead angles ranging from 4.5 degrees to 22 degrees. Here is the bounding range on efficiencies

Lead angle of 4.5 degrees has an efficiency of about 61%

Lead angle of 22 degrees has an efficiency of about 87%

This assumes everything is lined up!! Tolerances are very critical for a worm gear set-up.

The lead angle of 22 degrees corresponds to worms with 4 thread starts and the lead angle of 4.5 corresponds to worms with 1 thread start. The worm with one thread start has 4 times the gear ratio for the same package size than the worm with 4 starts. A worm gear box has several advantages for FIRST robots. The main two are package size and weight. We have used worm gears on our robots for three years, but only once on our drive transmission (see white papers under CCT).

60% efficiency is still downright terrible for a drivetrain when you consider spur gears are 98% efficient and chain is 95% efficient
Read the article. It’s actually possible to reach that degree of efficiency using regular worm gears.

Here is a picture of a hypoid gear. It is used on an automotive painting robot that my company manufactures.

I’m going to have to strongly disagree about efficiency not being an issue.

Even with your research, which I won’t refute, 40% losses in a single stage of a gear set is still quite large compared to the 10%-15% losses over an entire drive train with spur gears and chain.

It should be noted that Adam’s article did cite a 97% efficiency for a multiple stage gear train using worm gears, which I found to be quite impressive. (Costs? Yikes! :))

Worm gears have their place, but I would strongly discourage their use on a drive train if you’re just looking to reduce your gear ratio. I’d definitely like to see justification addresing the losses, which we can all admit are several times larger than spur gears.

(See Paul’s link for an appropriate application.)


The worms we used last year were 35:1 on the back of another transmission. I don’t recall the ratio for the Bosch tranny we used however. We wanted an extremely high torque setup for our application though.


Here is a design for positioning worm transmission from team 107 , the nice thing about our design is that there is 50 to 1 reduction on the worm which should make it not back drive. yet you can quickly change speeds and motor combination. It also can have either a pot or encoder mounted simply.

Thanks Jim schaddelee team 107

We used two FP’s last year to build a winch for climbing the bar. Be very careful about efficiency loss - we had ~ 30% loss. But it did lock up nicely and keep us hanging :slight_smile:

Here’s an mpeg of it running