Cutting Tetrix gears

During the fall some of the FRC team mentors the middle school FTC team. This is a great opportunity for students from both teams to get some shop time. Cutting gears isn’t something normally time efficient but it is a great opportunity to teach multiple skills in the shop so when the FTC needed a gear reduction for their arm we decided to make our own.

The first thing we discovered was the gears are metric 0.8MOD with a 20 degree pressure angle. 0.8MOD is 31.75DP and functionally close enough to be considered 32DP. We found a resource Team 542 posted very helpful:

Cutting gears is time consuming. It requires a fair amount of specialized tooling. It likely will require prep work as often the students time is limited. I decided to post up some of these photos from the fall so you can see how it’s done.

We ended up using the relativity easy to source 32DP cutters. We needed two cutters, one for the pinion and the other for the bull gear. I ordered them from Ash Gear I opted for import cutters to save the budget.

First thing I did was to turn a mandrel to mount the 104 tooth gear. I then had the students make a rough blank with the center hole and the 4 screw holes. This was then transferred to the lathe and turned down to diameter. Normally this would have been done between centers rather than the 3 jaw but we were in a hurry. Cutting gears takes time. A mistake I won’t make again. I ended up re-making the gear because the 3 jaw was so far out.

The indexer was then setup on the manual mill and the gear was cut. Great opportunity for everyone to get a hand at cranking the handles. The indexer was pulsed manually. The indexer is a Haas 5C.

After the students finished the gear to was obvious the gear wasn’t concentric. Never trust a three jaw chuck. :o So I re-made the gear at my shop. It was a great opportunity to show the two gears side by side and explain were the error was introduced and share the photos.

At the shop I can pulse the indexer automatically so a few lines of g-code and I can let it cut… while I get real work done. Notice this is done between centers.

Cutting the 16 tooth pinion. With pinions I like to cut a length that can be sliced up into multiple gears.

Here are the gears installed. They held up fine.

After the season I was able to pickup a full set of 0.8MOD cutters off Ebay from a seller in the Ukraine. They are smaller which is nice especially for making smaller gears and pinions. The R-8 13mm arbor was ordered from a vendor in the UK. We are now ready to cut more gears next season.

D: Wow that’s a lot of teeth. Seems like a lot of time to put in for a single gear (especially with the non-automatic indexing!) but if it’s for the learning opportunity I can respect that. :slight_smile: The things I’ve done for the learning opportunity…
Your results look really good! Love the pictures of the lathe and setups, lots of stuff covered in the making of a single gear.
I’ve always wanted to set up a gear hobber or something similar so I can make my own gears without having to worry about the hand indexing and all that.

It was at least an hour, 4 or 5 students traded out. Everyone got a chance to cut. The indexer certainly wasn’t manual… all they had to do was press an advance on the controller to advance one tooth. I’d say semi-automated. :smiley:

The issue with hobbing is that without a hobbing machine that can keep the hob and blank in sequence you still have to gnash the gear prior to hobbing, generally with a slitting saw. Granted hobbing generates a more precise involute form. The way I figure it is… if I have to gnash the bank anyway… I might was well just use an involute form cutter and be done.

I’d love to see some examples of hobbing if any teams do it. I’m all about speeding up the process.

Ah, that’s what you meant by “pulsed manually”. I was thinking of you keeping a really close eye on the students every time they cranked the indexer. :stuck_out_tongue: I missed the part where you said “Haas 5C” ::rtm::

I was thinking about using a geartrain to keep them in sequence. If you have to gash the gears then there’s no point to doing the hobbing over just buying involute cutters imo. I actually had a spreadsheet that had all the gears that were needed to be bought from Vex to generate most non-prime-tooth-count gears from 1-100 teeth and beyond. The hobbing machine that I designed required a lot more machining than I was comfortable with, unfortunately.

When we made gears in a low qty, we made the blanks on the lathe and then talked to a local hob shop. They allowed our kids to come in and observe/learn about the process. It was a really cool experience overall.

They also did the hobbing for free for us! But the leadtime was about a month, they just found a day or two where the machine was not being used at all.

Yeah, making a hobber would be a lot of work. Have you seen the Jacob’s hobber? Castings use to be available and you see variants now and again online. For small FRC and FTC gears something about that size would be perfect.

I wonder if anyone has devised a cnc hobber on the cheap?

Yes, I have seen that, and I really want to buy it, but the only site the castings are available from is in the UK and the thing is expensive enough already.

A CNC hobber is called “CNC mill + automatic indexer”, I believe. :stuck_out_tongue: To my knowledge (a few weeks of on-off research) nobody has made a dedicated CNC hobber, although it should be easy to set up if you already have a CNC mill.

If you Google “cnc gear hobber” you’ll find tons of examples. Not the same as a cnc mill with 4th axis at all. A gear hobber has to vary the angle the hob is presented to the work at in order to make different kinds of gears.

When not done on a dedicated hobbing machine, hobbing mostly occurs on y-axis equipped cnc lathes, minus some homebrew setups people have built with mills. I know 368 at one point had setup their mill to hob sprockets. I think there were pics on here somewhere, but you need to mechanically slave the rotary axis to the spindle.

It’s the “on the cheap” part that’s not in google. :stuck_out_tongue: Industrial gear grinders and the like are beautiful things.

You can dramatically speed up your production by stacking gear blanks on your arbor. It will almost cut the time per piece to (time to cut one gear)/(number of gears)=(time per gear) as all you need to do is increase the length of your cuts to accommodate your extra gear blanks.

Martin Model has the castings available based on Tom Hammond’s patterns. They appear more robust than the original Jacob’s castings. Ships from Oregon. 14 Piece Gear Hobber Casting Kit + Drawings $495.

Here’s a video:

I’m going down the gear cutting rabbit hole… I’ve been watching some of the Sandvik InvoMilling videos. Cool stuff for sure! Here’s a simple spur gear.

Poking around Sandvik’s site I found the CoroMill 171 which can do “Gear module range: 0.8–3 mm (DP 31.750–8.467)”… which got me thinking about how cool would it be to do it all with one cutter.

Then I wondered about how to program it.

Sandvik had a link to Euklid GearCam. It looks cool and spendy. I’m sure the Coromill isn’t cheap either.

Then I happened across this video of Wilfried Smekens cutting gears with a slitting saw.

He posted his “Involute spur gear free gcode wizard using O codes” source in the comments of the first video. I assume it runs in LinuxCNC. I wonder if I can make it work with the Centroid? I’ll have to study it some more.

Now I’m intrigued. I think we could do this. We have a 4th axis on the Centroid CNC at the school. This would be a perfect use in the fall. :slight_smile:

I just finished giving you rep for the last post… time to find 10 more people to give rep to. :smiley:
I saw the Sandvik gear cutters a long time ago, but I have never seen somebody cutting involutes with a slitting saw before! You would probably need a very small diameter saw to cut aluminum without too much chatter at a decent speed though.


The slitting saw wouldn’t be my first choice of course. but the code he’s using to generate the path is fairly straight forward and should be easily adapted to any cutter.

I quickly tried converting it to the Centroid but it’s complaining. I re-numbered the variables but the sub-programs are handled a little differently. Might take a little more to sort it out.

Being able to cut various pitches and tooth count of gears without the expense of many specialty cutters is interesting. Being able to give the kids a programming exercise is a bonus.

It would be very easy to re-write in python and generate the long g-code but doing it directly on the control g-code is more elegant.

Spent some time on the phone with some really nice folks at Sandvik. I learned a lot.

The Coromill 171 and 172 are profile inserts so you have the same 8 approximate involute shapes you’d find in a typical cutter set. It would be comparable to what were doing now.

The InvoMilling uses the Coromill 161 and 162. The insert is flat bottomed the software generates the involute shape. Single cutter can do any count. A five axis machine is recommended but spur gears can be done with just four. The smallest size the InvoMilling Coromill 161 can do module 2.0. The InvoMilling software is expensive. With the InvoMilling gear cam you can also cut gears with standard end mills and such. It appears the secret sauce is the software more than the cutter.

I was able to talk to a programmer and we discussed the slitting saw method and a few others. He was confident if we could program for the saw then their Coromill 161 would work. The Coromill 161 would be more rigid for sure.

I guess I’ll keep working on porting the slitting saw code.

I re-wrote the slitting saw code in python so I could generate some g-code to play with. This is neat stuff. The rendering is a bit odd looking because of the zero is on the outside edge of the blank and the A axis rotates around the X. It appears to work. I wrote my version to automatically calculate the blank diameter based on the gear tooth count and module making it very simple. Enter the module, tooth count, blank width, and the cutter details, it then spits out g-code.This is a good starting point to play with. I’m definitely going to modify this code to handle standard profile gear cutters also.

The render is for a mod 3.0, 38 tooth gear, and a 30mm face width. The cutter is 60mm diameter and 2mm thick. Similar to the gear in the original video.

I really look forward to seeing this in action!
If you had standard profile gear cutters, wouldn’t that be considerably simpler code?

“The Home Shop Machinest” ran some articles on using slitting saws on small CNC Sherline size tools to make aluminum gears. I can dig out issue numbers if someone wants.

If one has an old metal shaper one can also make a carbide cutter for gear teeth for the clapper. Course shappers only cut on the forward stroke so…dead simple tooling…but you pay for it in 40% more cutting time.