Gearotic V3 to design Gears for Repair

[jdrichard]

Folks. I have been interested in being able to build and machine a watch gear for some time now. I would love to cut one on my lathe, but the vintage addins are really hard to find. I am good with designing the gear and getting it CNCed by someone else or even printing the gear and filing it by hand. That said, I ran into a program called Gearotic V3 where you design a gear and a gear train and send the file to a CAD program.

 

Have any of you tried to do this or used this program for that purpose? It does have clock escapements it the designer but not watch lever escapements. Thoughts, suggestions, etc. the program was 120US, but no issue as I love learning and an adventure.

 

 

Sent from my iPhone using Tapatalk Pro

[nickelsilver]

I make gears all the time, I use the norms in the swiss NIHS book which gives you all the math to draw the profiles for any module of cycloidal gear. Does the Gearotic do cycloidal or just involute? While several watch companies have moved from cycloidal to a highly modified (and proprietary) involute derived tooth form over the last decades, cycloidal is the standard. It won't mix with involute. Well, it can, but-

 

Gearing is a pretty huge subject. You can get an involute to mesh with a cycloid and they might even seem friendly together, but the efficiency will be diminished and they will self destruct over time. If you reverse engineer geartrains in watches you'll find that they rarely match perfectly to theoretically correct numbers; this is because they were modified based on experience or testing.

 

All that said, if you make a gear train for a watch "by the book" it will perform well and last. If you need to replace a gear or pinion there is usually the original to copy, and if not, a reverse engineered one will perform well. The larger the more leeway you have. A pocketwatch escape pinion can probably be off in diameter by as much as +- 0.02mm and still function; in a small wristwatch you will be cutting it close at +- 0.005mm.

 

What's really a problem is getting cutters. There are several companies in Switzerland who will make you one to your drawing for a couple to three hundred bucks. If for a wheel of say 70 teeth, that cutter will probably be OK for any wheel of that module up to a rack, down to around 35 teeth. For pinions it's much trickier, a cutter for a 7 leaf pinion of a given module will cut that and nothing else.

 

One does find cutters secondhand, but 99% of the time they are marked with some arbitrary number and you'll need a profile projector to trace the profile and compare to drawings.

 

That's all said not to discourage, just to illustrate a little bit what's in store. I'll be the first to say that making even a theoretically terrible gear on your own is a really great feeling when you get the last tooth- it's like you're a wizard. Put it on your keychain and get ready to do some math then.

 

 

[jdrichard]

I make gears all the time, I use the norms in the swiss NIHS book which gives you all the math to draw the profiles for any module of cycloidal gear. Does the Gearotic do cycloidal or just involute? While several watch companies have moved from cycloidal to a highly modified (and proprietary) involute derived tooth form over the last decades, cycloidal is the standard. It won't mix with involute. Well, it can, but-
 
Gearing is a pretty huge subject. You can get an involute to mesh with a cycloid and they might even seem friendly together, but the efficiency will be diminished and they will self destruct over time. If you reverse engineer geartrains in watches you'll find that they rarely match perfectly to theoretically correct numbers; this is because they were modified based on experience or testing.
 
All that said, if you make a gear train for a watch "by the book" it will perform well and last. If you need to replace a gear or pinion there is usually the original to copy, and if not, a reverse engineered one will perform well. The larger the more leeway you have. A pocketwatch escape pinion can probably be off in diameter by as much as +- 0.02mm and still function; in a small wristwatch you will be cutting it close at +- 0.005mm.
 
What's really a problem is getting cutters. There are several companies in Switzerland who will make you one to your drawing for a couple to three hundred bucks. If for a wheel of say 70 teeth, that cutter will probably be OK for any wheel of that module up to a rack, down to around 35 teeth. For pinions it's much trickier, a cutter for a 7 leaf pinion of a given module will cut that and nothing else.
 
One does find cutters secondhand, but 99% of the time they are marked with some arbitrary number and you'll need a profile projector to trace the profile and compare to drawings.
 
That's all said not to discourage, just to illustrate a little bit what's in store. I'll be the first to say that making even a theoretically terrible gear on your own is a really great feeling when you get the last tooth- it's like you're a wizard. Put it on your keychain and get ready to do some math then.
 
 

Hamilton 12S designed all gears


Sent from my iPhone using Tapatalk Pro

[jdrichard]

Sent from my iPhone using Tapatalk Pro

[JGrainger]

The practical side of the gear cutting setup is pretty straightforward except for the cutters, that's also one of the nice things about lantern pinions in clocks.. you can relatively easily make a flycutter for the wheels and just drill and use lantern pinions. The other machine parts of a working setup can be put together relatively easily..

In my case I've got a late 30's Hardinge BB59 bench lathe which is sort of like a Schaublin but with 9" swing and 5C collets. The bed profile is pretty standard for a bunch of bench lathes of that style and age - like a bigger WW lathe. The Schaublin 70 isn't a million times bigger than a WW (which is still more rigid than a Geneva pattern) and aside from period accessories, there are new Chinese Geneva lathe parts which could be matched up with other bits, some other headstocks such as little Emco / Sherline, etc.. I'm using a WW headstock until I've had time to make a dedicated milling spindle with ISO20 taper.. the WW headstock is plenty beefy enough for watch wheels. Add to that the Chinese/Indian vertical slides (I've got a myford copy) then only the motor and method of directing belting remain.. and an index plate, plate carrier, spindle lock, and indexing pin. I've used an old 1/3hp refrigerator motor and made the overhead drive from scratch - that was a bit laborious but basically a series of simple operations with a 4 Jaw chuck (I'd have preferred my mill but it's in storage until CV-19 makes it possible to collect).. could have made the dimensions a bit smaller to make the holes small enough for a drillpress and reamer if I only wanted to mill watch gears. I would recommend an optical aid for setting the tool.. and I'm not an expert in this but knew people doing it for years before my first try more recently.. still took a few tries due (mostly) to stupid mistakes.

Flycutters can be made in a few ways, aside from grinding them from a broken drill by hand (at the cheap and simple end of the spectrum) or using a cutter (like an engraving single sided cutter) they can be made in soft material and hardened, either by marking and drilling as per Daniels book, or you could probably use a circular turning tool of a calculated diameter then mill the excess material away to leave the profile and clearance - that wouldn't build in side clearance without a further honing step.

[jdm]

On 6/1/2020 at 8:34 PM, nickelsilver said:

I make gears all the time, I use the norms in the swiss NIHS book which gives you all the math to draw the profiles for any module of cycloidal gear. Does the Gearotic do cycloidal or just involute? While several watch companies have moved from cycloidal to a highly modified (and proprietary) involute derived tooth form over the last decades, cycloidal is the standard. It won't mix with involute. Well, it can, but-

Gearing is a pretty huge subject. You can get an involute to mesh with a cycloid and they might even seem friendly together, but the efficiency will be diminished and they will self destruct over time. If you reverse engineer geartrains in watches you'll find that they rarely match perfectly to theoretically correct numbers; this is because they were modified based on experience or testing.

Yes. See this doc for an example.

IS 05_Evolution calibre SW200 vers SW200-1.pdf

5 hours ago, JGrainger said:

I'm using a WW headstock until I've had time to make a dedicated milling spindle with ISO20 taper.. the WW headstock is plenty beefy enough for watch wheels. Add to that the Chinese/Indian vertical slides (I've got a myford copy) then only the motor and method of directing belting remain.. and an index plate, plate carrier, spindle lock, and indexing pin

Would something like the below help? Www.unitor.it or PM for more info.

[JGrainger]

5 hours ago, jdm said:

Would something like the below help? Www.unitor.it or PM for more info.

My own setup is complete but that looks like a nice option.. even takes me back to a previous job in which I was involved in making similar types of flexible drive, they're used in a number of fields. A short sheep-shearing drive would probably use the same size core (those are around 8-9mm for anyone looking to experiment, 6mm would probably work depending on the maximum work you intend), it can be driven directly by the motor, then geared down at the milling spindle so you stay within the torque limits of the flexible core while having the torque at the tool..

..but the setup you show or some of the other bits would allow JDR to cut gears (the nice WW gear cutting accessories are somewhat lovely but there are other equipment paths).