PDA

View Full Version : Fixing broken teeth on gear



Axkiker
02-28-2015, 12:21 PM
So if you read my thread about the 5914 purchase you know I have a few gears to fix. Since broken gears can be from any machine I decided to start a new thread regarding the process to fix and my questions.

My first question is can anyone give me the pros and cons to either rebuilding with braze or tig? The gears I have are not cast and I have the ability to do either or. If you choose one method over the other could you recommend a rod and or flux to use.

Second does anyone have any ideas as to cutting the teeth without a dividing head. I have made gears before using the original as a pattern and indexing the new blank using the old gears teeth. The gears I made were made out of uhmd so im not sure it will work as well cutting into either braze or tig...

Ive considered trying to make a jig where I could hold a shaft the same dia as the ID of the gear. Then making a sorta circular pattern at the rear where I could print off an index pattern using a graphic program the same tooth pattern of the gear. I would then need to manually spin the shaft and line up marks using the template printed . Hard to explain but it sorta makes sense in my head. An alternative is buying a dividing head. I really want one but cant hardly justify the cost.

Lastly is to buy some gear pitch gauges and see if I can get gears from Boston gear.

Any thoughts ?????

Bob Fisher
02-28-2015, 01:16 PM
My Logan came with a missing tooth on the bull gear on the back gear. I got lucky on EBay and bought the entire back gear Assy for $70. I brazed up the missing tooth and carefully filed it to shape such that it would mesh well with its mate. Worked well and now I have a spare, that should guarantee that I'll never break another tooth. You can make an excellent approximation of a proper gear profile with diligent filing. Bob.

Carm
02-28-2015, 01:24 PM
A brazed repair can last for a long, long time. Use borax flux. I'd rather puddle teeth with a torch than TIG the thing, but that works too.
Like Mr. Fisher said, get after it with a file & diligence. Having the gears mounted makes for quick work provided you have access. Bring into mesh and work the shiny 'til you get what you want.

dp
02-28-2015, 01:28 PM
Keith Fenner recently did a youtube video of replacing broken gears. It is a multi-part series that starts here. You can jump ahead to part 7 where he begins repairing the original gear.


https://www.youtube.com/watch?v=qEGAs8OMuJc

Axkiker
02-28-2015, 01:39 PM
Keith Fenner recently did a youtube video of replacing broken gears. It is a multi-part series that starts here. You can jump ahead to part 7 where he begins repairing the original gear.


https://www.youtube.com/watch?v=qEGAs8OMuJc

Yep... ive watched that video several times. He brazed but I think the gear is Cast so not sure if that makes a difference.

Also FYI one of the gears I need to fix is missing several teeth in a row. So I dont think hand filing is an option. If someone has a way to file in multiple teeth in a row please let me know

Thanks

Carm
02-28-2015, 02:10 PM
"Also FYI one of the gears I need to fix is missing several teeth in a row."

Think stone age. Make a replication by cast, cut a template etc. Do your braze build up, cut the OD w/lathe if it isn't out of action. Rough w/hacksaw and start at the first repaired tooth with the template. Get close and stair step each tooth. If you can mesh with a mating gear, use blue or just cut the shiny.
If you have the money, buy the gears. But the above will be quicker than getting a D/H and setting up w/cutters etc.

Axkiker
02-28-2015, 02:39 PM
Found this youtube vid... its kinda what I had in my head for a DIY dividing head using a printout for index degree markings.

I think I would personally make a more ridged / robust unit... You get the idea

https://www.youtube.com/watch?v=-we6ORrJgUI

Axkiker
02-28-2015, 03:43 PM
Found this youtube vid... its kinda what I had in my head for a DIY dividing head using a printout for index degree markings.

I think I would personally make a more ridged / robust unit... You get the idea

https://www.youtube.com/watch?v=-we6ORrJgUI

Also, after thinking about this a bit more I wonder if instead of using the print out as an angle finder I could use one of those little digital angle finders.

J Harp
02-28-2015, 04:38 PM
Index right off the gear you are repairing. Mount the gear on an arbor, mandrel, or what have you. Turn a taper on the end of a bolt so it will bear on both sides of a tooth space without bottoming out. Drill and tap a hole for the bolt in a piece of bar and find a way to clamp it to a fixed part of the machine you are using for the job.

Set up whatever cutter you are going to use and feed it into a good tooth space until it touches both tooth flanks. Then fit the bolt into another tooth space and clamp it to the machine so it will stay put. Back the bolt and the cutter out of the tooth spaces and turn the gear so the first space to be cut lines up with the cutter, run the bolt in till it it fills the tooth space in front of it. The gear is now located for the first cut, it's probably a good idea to clamp the gear to something so it cant turn while cutting.

Mark Rand
02-28-2015, 06:26 PM
I've repaired the horizontal+vertical, partial-circle, 16DP feed gears (with integral ratchets as part of the casting) of a shaper, by welding up with nickel rod and re-cutting the teeth. Ended up machining off all the teeth, building the whole lot up, turning the result down and re-cutting all the teeth.

For the Beaver milling machine that I'm rebuilding, I just bought replacement gears, from HPC in the UK, and machined them to suit.

If it's a spur gear with a relatively simple geometry, it will almost always be simpler to buy or make a new gear rather than trying to repair a few teeth on the existing one. (exceptions are full side traction engine or similar sized gears).

The 98% Nickel rod I used has given good service and is probably more resilient than the original cast iron teeth were.

jdunmyer
02-28-2015, 06:35 PM
I've repaired a cast gear or 2, using braze. In addition to the braze, I D&T'd across the broken section, ran in SHSSs, then brazed over/around them. Seems stronger to me.

dp
02-28-2015, 06:50 PM
Yep... ive watched that video several times. He brazed but I think the gear is Cast so not sure if that makes a difference.

Also FYI one of the gears I need to fix is missing several teeth in a row. So I dont think hand filing is an option. If someone has a way to file in multiple teeth in a row please let me know

Thanks

You can try using mill board as a dam on the sides of the gear, or mix up some refractory.

Brazing should work as a filler for steel. Steel has a higher melting point than brass by about 1000F. You can buy a single involute gear cutter rather cheaply and pick up the teeth for the cutter using the gear itself and a mechanical stopping wedge.

Axkiker
02-28-2015, 09:36 PM
Index right off the gear you are repairing. Mount the gear on an arbor, mandrel, or what have you. Turn a taper on the end of a bolt so it will bear on both sides of a tooth space without bottoming out. Drill and tap a hole for the bolt in a piece of bar and find a way to clamp it to a fixed part of the machine you are using for the job.

Set up whatever cutter you are going to use and feed it into a good tooth space until it touches both tooth flanks. Then fit the bolt into another tooth space and clamp it to the machine so it will stay put. Back the bolt and the cutter out of the tooth spaces and turn the gear so the first space to be cut lines up with the cutter, run the bolt in till it it fills the tooth space in front of it. The gear is now located for the first cut, it's probably a good idea to clamp the gear to something so it cant turn while cutting.


Yup thats exactly how I made the gears for the lathe I have which works. Since I have 3 gears of different sizes to fix for this lathe I thought about possibly making something which would make all 3. Plus It could be used for other tasks..

Just thinking aloud

jep24601
03-01-2015, 09:38 AM
My brother bought a lathe that had a number of gear teeth "repaired" by simply drilling and threading a hole into which a screw is inserted (presumably with Loctite or epoxy) and which is then filed to shape - not pretty but it seemed to work.

Axkiker
03-01-2015, 03:16 PM
So I was thinking about something. A dividing head uses plates with holes drilled in them to index the gear for cutting. Why could one not use one of the little digital gauges that are out there?

Basically if you were making a gear from scratch you would make your first cut then zero the digital protractor. Then spin the blank to the appropriate angle. Make your cut and zero the gauge again. Repeat Repeat Repeat until the gear is complete.

My first thoughts are that the little digital protractors just arent accurate enough. After some research I have found some affordable units that claim +- .05 degree of accuracy. Anyone have an idea if this would be accurate enough? I am having a hard time corelating the .05 to real world tooth difference. I would kinda think that there could possibly be more slop in the indexing head etc. When you consider all the moving parts, plates, pins etc the gauge idea really sorta simplifies things.

Anyways just an idea.

Carm
03-01-2015, 03:52 PM
Axkiker, an attempt at explanation, post#15

Dividing heads and their cousins, worm drive rotary tables, work by averaging around a circumference, somewhat akin to half-nuts and leadscrew. Provided drive is either positive or accounted for (backlash) wear in the components has little effect on the result. Using the pins on the plate once again averages and eliminates mental health issues trying to read minutes and seconds.
Any error in the protractor will be compounded/ cumulative.
Not to say it couldn't be done, but there would be a time and frustration penalty in verification.

Fred P
03-02-2015, 01:36 AM
Some years back I did a repair on a knotter gear on a hay baler. Three or more teeth were broken out. I made a plaster of Paras mold of the teeth on the good side of the gear. I left one side of the mold open to fill from. I tinned the area of the missing teeth with Braze. Put the mold in the wifes oven for about an hour or so to dry it out. Placed the mold over the damaged area, warmed things up good and filled the mold with braze. Just kind of dripped the braze into the tooth voids. The new braze teeth needed very little dressing and the gear ran for many years.

Fred P...........

batt-man
03-02-2015, 03:30 AM
Some years back I did a repair on a knotter gear on a hay baler. Three or more teeth were broken out. I made a plaster of Paras mold of the teeth on the good side of the gear. I left one side of the mold open to fill from. I tinned the area of the missing teeth with Braze. Put the mold in the wifes oven for about an hour or so to dry it out. Placed the mold over the damaged area, warmed things up good and filled the mold with braze. Just kind of dripped the braze into the tooth voids. The new braze teeth needed very little dressing and the gear ran for many years.

Fred P...........

Now that's really smart !!!

sawlog
03-02-2015, 04:39 AM
If I was faced with the OP problem my first thought is to try and buy a stock gear of the same pitch and number of teeth and machine the gear to fit. If I could not do that then I would braze the gear up and machine it so I could cut the teeth in it, to do the indexing I would index off the opposite side because that would index the gear correctly, without any extra fixtures.

SGW
03-02-2015, 06:44 AM
Re: +/- 0.05 degree.

It depends on the gear diameter. Say the gear is 4" in diameter. The circumference would be 4*pi or roughly 12.5". 0.05 degree is 1/(360*20) of a circle, or 1/7200. So on a 4"diameter gear, 0.05 degree would be 12.5/7200 inches or 0.00174 inches. Just under two thou. Since it's +/- 0.05 degree, you could have a tooth-to-tooth error of up to about 0.0035", which would probably be too much to tolerate....but maybe not. You would probably have to clean up the gear meshing with a file.

I think your best bet is to get a new stock gear and modify it to suit.

Axkiker
03-02-2015, 09:35 AM
Re: +/- 0.05 degree.

It depends on the gear diameter. Say the gear is 4" in diameter. The circumference would be 4*pi or roughly 12.5". 0.05 degree is 1/(360*20) of a circle, or 1/7200. So on a 4"diameter gear, 0.05 degree would be 12.5/7200 inches or 0.00174 inches. Just under two thou. Since it's +/- 0.05 degree, you could have a tooth-to-tooth error of up to about 0.0035", which would probably be too much to tolerate....but maybe not. You would probably have to clean up the gear meshing with a file.

I think your best bet is to get a new stock gear and modify it to suit.


Thats exactly what I was curious about.... I just didnt know the math to calculate.

What im also curious about is the tolerance of a true dividing head. I know someone earlier spoke about the geartrain of the Index removing all lash by basically advancing it. This makes total sense to me and I agree. What I wonder is regarding the way in which the index locates each tooth. I would think that there could be enough play in the pin mechanisim to cause some fluxuation. With all that said I do not have a dividing head and have never used one.

Carm
03-02-2015, 11:59 AM
"What I wonder is regarding the way in which the index locates each tooth."

The dividing head has a worm driving a worm wheel. Different ratios from different makers, but in any case, the helix of the worm is sufficiently steep to prevent backdrive although not backlash.
The Swiss cheese plate and shotpin are selected in cases where 360 gets divided into steps that involve minutes and seconds. Any error (tolerance) in the hole spacing or shot pin fit gets averaged over the circumference. Trying to step by reading the degree, minutes & seconds on the vernier is certainly doable but for most people that means a long list of numbers best prepped in advance, and the great possibility of concern whether one has missed a step.

Dividing heads used to, and perhaps still do in some shops, drive leadscrews through a geartrain. Despite what looks at first glance like a great place for a stackup of errors, the world today got here with their help making tooling, ordinance,
lab equipment, medical devices....

Axkiker
03-02-2015, 02:03 PM
"What I wonder is regarding the way in which the index locates each tooth."

The dividing head has a worm driving a worm wheel. Different ratios from different makers, but in any case, the helix of the worm is sufficiently steep to prevent backdrive although not backlash.
The Swiss cheese plate and shotpin are selected in cases where 360 gets divided into steps that involve minutes and seconds. Any error (tolerance) in the hole spacing or shot pin fit gets averaged over the circumference. Trying to step by reading the degree, minutes & seconds on the vernier is certainly doable but for most people that means a long list of numbers best prepped in advance, and the great possibility of concern whether one has missed a step.

Dividing heads used to, and perhaps still do in some shops, drive leadscrews through a geartrain. Despite what looks at first glance like a great place for a stackup of errors, the world today got here with their help making tooling, ordinance,
lab equipment, medical devices....

Yeah I understand they are plenty accurate.. Im just trying to relate how much more / less they could possibly be than say using the digital protractor idea I have. I guess a lot also depends on the quality of the index and possible wear..

Just something im curious about really

Axkiker
03-02-2015, 07:36 PM
Another idea I had was to use a stepper motor and control. What im curious about is how stepper motors are used to control the movement of something when the needed position doesnt fall exactly on the resolution of the motor. I guess this is more of a stepper motor question

Example:
If I have a stepper that has 360 degrees of resolution. I need to have it turn an object to 1/2 of a degree. What does the stepper do? I understand you can use a belt and pulley to increase this resolution. What happens though if even that doesnt allow for the exact alignment of the stepper ???

Axkiker
03-05-2015, 03:41 PM
Awww now, I know someone out there has some stepper motor knowledge.... Should I start another thread ?

lakeside53
03-05-2015, 05:33 PM
If you have a stepper with a 1.8 degree steps, and run it in half steps, then you can directly resolve 0.9 degrees. If your worm was 40:1, that's now down to 0.0225 degrees. If that's not good enough for you, put on a 5:1 reduction belt and you now have 0.0045 degrees. You can skip the belt and put on a 50:1 harmonic drive gearbox (zero backlash) and now your errors are tiny.

What's important is that you don't just calculate the one best incremental count per division - that way errors accumulate - calculate actual steps from origin to best fit the needed position. Backlash needs to be accounted for (but you'll need a table derived imperially from you head if not constant) or drive one direction only against some friction, and lock the head each time you cut.

I'm about to convert my dividing head (the DH in my other other thread) to cnc. Not sure if I'll just buy a DivisionMaster or roll my own.. I'm half tempted to drop one of my servos on the DH and put an encoder on the spindle. lol.. but I'll probably just go stepper for now.

Axkiker
03-06-2015, 08:59 AM
If you have a stepper with a 1.8 degree steps, and run it in half steps, then you can directly resolve 0.9 degrees. If your worm was 40:1, that's now down to 0.0225 degrees. If that's not good enough for you, put on a 5:1 reduction belt and you now have 0.0045 degrees. You can skip the belt and put on a 50:1 harmonic drive gearbox (zero backlash) and now your errors are tiny.

What's important is that you don't just calculate the one best incremental count per division - that way errors accumulate - calculate actual steps from origin to best fit the needed position. Backlash needs to be accounted for (but you'll need a table derived imperially from you head if not constant) or drive one direction only against some friction, and lock the head each time you cut.

I'm about to convert my dividing head (the DH in my other other thread) to cnc. Not sure if I'll just buy a DivisionMaster or roll my own.. I'm half tempted to drop one of my servos on the DH and put an encoder on the spindle. lol.. but I'll probably just go stepper for now.

Good info... I think part of what I didnt understand was that you could run the motor in half steps..

Im curious how one would go about calculating the steps from origin to obtain the best fit. Any thoughts. What you said makes sense I just cant grasp how to calculate it.

I have never heard of DivisionMaster before, but it really looks to be exactly what I need. I dont really want to spend the time to reverse engineer everything. One of their kits would be simple and they dont seem that expensive.

lakeside53
03-06-2015, 12:05 PM
You can run the motor in microsteps if you like (driver based) but you may accumulate accuracy issues with doing this. Keep it simple!

There are many ways to find the nearest total integer count from origin (say zero degrees) to an actual tooth position. You can track total counts from zero to the desired position, subtract the current count, and simply drive by the difference of steps to the next location. You obviously need to factor worm ratios, stepper modes etc. Based on a simple integer response to "how many teeth" question on the display, I'll probably simply pre-calculate best step count matches to a particular tooth count and load it into an array of long integers. Less code if you want to back up, re-start, hit only one or two random teeth etc. etc With sufficient worm or other ratios you might be able to skip all this stuff and just go with floating point calcs, but I prefer not to for embedded control.

Part of me wants just buy the DivisionMaster (or whatever - there are other indexers available) and get on with gutting gears. The other apart - make it all myself. I actually built one a year or so back for another project and should just dust it off and make it work for my DH. I'll toss a coin soon.

Here's one I fiddled with on 2013. http://www.cnc-arena.com/en/forum/cnc-indexer-arduino-based--185758.html
It works fine and kudos to the original creator. I don't want to be insulting.. but the code does not appear to be from a professional programmer (maybe not native English speaker) so it takes some understanding so modifications are a bit tricky. I was going to rewrite it all but got side tracked. lol.

Axkiker
03-06-2015, 12:47 PM
You can run the motor in microsteps if you like (driver based) but you may accumulate accuracy issues with doing this. Keep it simple!

There are many ways to find the nearest total integer count from origin (say zero degrees) to an actual tooth position. You can track total counts from zero to the desired position, subtract the current count, and simply drive by the difference of steps to the next location. You obviously need to factor worm ratios, stepper modes etc. Based on a simple integer response to "how many teeth" question on the display, I'll probably simply pre-calculate best step count matches to a particular tooth count and load it into an array of long integers. Less code if you want to back up, re-start, hit only one or two random teeth etc. etc With sufficient worm or other ratios you might be able to skip all this stuff and just go with floating point calcs, but I prefer not to for embedded control.

Part of me wants just buy the DivisionMaster (or whatever - there are other indexers available) and get on with gutting gears. The other apart - make it all myself. I actually built one a year or so back for another project and should just dust it off and make it work for my DH. I'll toss a coin soon.

Here's one I fiddled with on 2013. http://www.cnc-arena.com/en/forum/cnc-indexer-arduino-based--185758.html
It works fine and kudos to the original creator. I don't want to be insulting.. but the code does not appear to be from a professional programmer (maybe not native English speaker) so it takes some understanding so modifications are a bit tricky. I was going to rewrite it all but got side tracked. lol.



I tried the link provided but it was broken.

Do you know what what language the division master uses? Ive done some c# so hopefully I could get by

Axkiker
03-10-2015, 01:16 PM
Okay im back to working on these gears. I have the first brazed up and turned down on the lathe. I found a set of gear took gauges on ebay and bought them. This first gear is a 72 tooth 16dp gear.

I thought that was all I needed until I was reminded that I need to know the pressure angle before ordering a cutter. Can anyone tell me what the pressure angle would be or how to figure it out. Most everything I find says it would be a 14.5 or a 20.

Any ideas

Thanks

lakeside53
03-10-2015, 06:09 PM
The link isn't broken -works fine from my computers.

Division Master is a proprietary system so you won't be changing the code ;) The Ardunio based system on the link above is C++, but mostly just straight C.

Axkiker
03-10-2015, 07:30 PM
Anyone able to tell what pressure angle this gear is by looking at it. I dont have others to compare it to so im left with asking others input.

http://i1167.photobucket.com/albums/q633/Axkiker/Mobile%20Uploads/2015-03/00C62B69-E94F-41B8-BB3F-F671AB6670C9.jpg (http://s1167.photobucket.com/user/Axkiker/media/Mobile%20Uploads/2015-03/00C62B69-E94F-41B8-BB3F-F671AB6670C9.jpg.html)

http://i1167.photobucket.com/albums/q633/Axkiker/Mobile%20Uploads/2015-03/02BFBA2C-2DDB-4B63-B7E2-7C06E93E45B8.jpg (http://s1167.photobucket.com/user/Axkiker/media/Mobile%20Uploads/2015-03/02BFBA2C-2DDB-4B63-B7E2-7C06E93E45B8.jpg.html)

Axkiker
03-10-2015, 07:33 PM
Here is the pitch gauge that matches perfectly.

http://i1167.photobucket.com/albums/q633/Axkiker/Mobile%20Uploads/2015-03/2CACA34F-C1B2-4D11-86A8-FDE9E07D5AE7.jpg (http://s1167.photobucket.com/user/Axkiker/media/Mobile%20Uploads/2015-03/2CACA34F-C1B2-4D11-86A8-FDE9E07D5AE7.jpg.html)

Axkiker
03-10-2015, 09:01 PM
After looking at a few pics im thinking this is a 14.5

Anyone agree or disagree ?

http://i10.photobucket.com/albums/a149/ufscrap/shaker%20shaft/14.jpg

Axkiker
03-11-2015, 07:16 PM
Nothing ???? Awwww come on guys I know someone has an opinion.. I would kinda like to buy a cutter for this job instead of hand grinding something.

Axkiker
03-28-2015, 06:56 PM
Anyone have specs on 14.5 degree involute gear cutters. I may entertain making my own but am looking for specs. Someone on here directed me to myfordboy and his youtube videos. He gave a good explanation on how to make gear cutters and did so using specs. So far the only specs I can find are for a 20 degree pressure angle

Im looking for something like whats provided from the link below

http://www.metalwebnews.com/howto/gear/gear1.html

Thanks

John Stevenson
03-28-2015, 07:27 PM
I reckon 14.5 degrees judging from this.

http://www.stevenson-engineers.co.uk/files/test%20gear.jpg



And the chart you want to go with that article.

http://www.stevenson-engineers.co.uk/files/14.5degree.gif

Juergenwt
03-28-2015, 08:04 PM
Good enough for working with wood or making some plastic gears (or something that looks like a gear). The product produced in this type "Dividing head" or using this type of cutter can not be used to make a gear for a machine tool. Properly good enough for some simple dividing where accuracy is not required.

Axkiker
03-28-2015, 08:45 PM
Good enough for working with wood or making some plastic gears (or something that looks like a gear). The product produced in this type "Dividing head" or using this type of cutter can not be used to make a gear for a machine tool. Properly good enough for some simple dividing where accuracy is not required.


Well, that maybe true for some machine tools... I however made gears for my other lathe in a much less accurate fashion and it works fine. Cuts threads etc just like it should. This time around I was looking to get a little more accurate without buying cutters or a dividing head.

Juergenwt
03-28-2015, 11:39 PM
Well, that maybe true for some machine tools... I however made gears for my other lathe in a much less accurate fashion and it works fine. Cuts threads etc just like it should. This time around I was looking to get a little more accurate without buying cutters or a dividing head.

I am not going to argue with success. If it worked for you - that is all that was needed.

elf
03-28-2015, 11:52 PM
Why don't you use your new CNC machine to cut the gears :)

metalmagpie
03-29-2015, 10:49 AM
My gear tooth gage set is labeled what the pressure angle is. Have another look.

I wrote up how I fixed a gear with broken tooth back in 1997. Have a look:

First, I considered having another gear cut. There are a lot of gear manufacturers who can cut any gear there is. However, the cheapest quote I got was well over $300, and this was out of my price range, so I discontinued this search.

Next, I considered buying a stock gear. I called most of the local stocking distributors, and got the Boston Gear catalog, and found the Stock Drive Web page, and so on. The problem was that my gear was helical, and had 94 teeth, and had only a 10 degree helical angle. It turned out that there was no way anyone stocks this gear. I did, by the way, consider replacing the 4 gears in this particular gear train with spur gears. This would have cost about $200 plus some machining. Better than $300, and has the side effect of having all new gears, but still basically way too much money.

The next thing I considered was milling out a slot, silver-soldering in a blank tooth, and cutting the tooth to shape with a shaper. There are three problems with this approach from my perspective. First, I don't have a milling machine. Second, I don't have a shaper, even if I did know how to grind a tool to the perfect profile. Third, several gear companies told me no matter what kind of repair I did on a cast iron gear, to stay away from anything that required heat, like a brazing operation, on just one part of the gear. This made sense to me, so again I abandoned this train of thought.

Finally, I decided to take the advice of a couple of guys from the Internet who were most helpful. I decided to drill and tap 3 holes, put in set screws, rough grind the set screws to approximate shape but slightly smaller, and to cast in place epoxy putty to form a new gear tooth over the 3 pins made from set screws. I figured the epoxy putty would bind to the remaining threads of the set screws, and that the tough metal the set screws are made of would sufficiently strengthen the new tooth. I have been successful at this method. The remainder of this text describes in detail how exactly to fix a gear tooth with a couple of bucks worth of JB Weld and about ten cents worth of set screws.

The first problem is how to hold the gear for machining. I found a solution that worked well for me. I mounted the gear on its shaft, which was 3/4" in diameter, and put the shaft into a 5C collet fixture that was quite solid and stable. I have seen these fixtures offered at very reasonable prices from a number of vendors.

I fixed the gear rigidly in the fixture, as described above, and then put the entire fixture in the bench vise. I took a very small fine file and filed the broken tooth off flat. The file I used had no teeth on either side, so it didn't mar the unbroken teeth.

After filing the tooth flat, I scribed a line down the middle of the tooth. This line was 10 degrees from the axis of the gear, which is the helix angle. Then, I center-popped 3 centers by eye. I then took the fixture over to the drill press, where it bolted down to the table. I located the holes with a wiggler, and drilled and tapped for (in my case) 4-40 set screws. This actually was the most difficult part. I find it difficult to get a very small drill bit to start where I want it, and this time was just the same way. Regardless, I got 3 usable holes, and they tapped just fine.

I then Loctited in the 3 set screws. I only tapped them about 3 threads deep (maybe 1 diameter) because I didn't want to weaken the gear body.

Then I cleaned some teeth on the opposite side from the break until they were as clean as I could get them. I cut a couple pieces of scrap black sheet metal, maybe 1" x 5", approximately 1/16" thick. Not too flexible. I cut out some rectangles of wax paper, and carefully folded them around the metal strips, taping them up the back side. Then I took the gear outside and lit my oxyacetylene torch, with just a little bit of oxygen, which generates a very yellow, sooty flame. I played the flame directly over the gear teeth and got a lovely velvety black soot layer. Then I took the gear back inside, and clamped the metal strips over the cleaned, sooted teeth, and mixed up some JB Weld, and gooped it in with a little piece of sheet brass cut just a little thinner than the gear. I covered about 5 teeth. The JB Weld says it sets up in 4-6 hours and cures overnight. After 6 hours, I put the gear in the vise, took a suitable block of metal, and gave the molding a robust tap from the side. It shattered!

Chastened, I cleaned up the teeth again, and tried again. What I figure I did wrong was to put on too much JB Weld, worked it in too much instead of letting it run in, and tried to remove it before it was strong enough. Putting on too much caused it to flow sideways over more teeth, getting a better grip. Working it in overmuch mechanically removes the soot, which acts as a releasing agent. Anyway, the second time around, I got a piece which spanned 4 teeth and came out perfect.

Then I got out the Dremel tool and a solid carbide burr, and went to work on the set screws. I didn't care very much about the final shape, I just wanted to leave as much meat as I could while allowing a reasonable amount of epoxy to cover. I used the mold to test the fit until I was satisfied that the pins were ground down enough.

Then, I took the same sheet metal strips I used when I made the mold, and again applied wax paper, this time to only one of them. I "sooted" the mold I'd made, again with the torch, and clamped the one side piece back onto the gear. I applied some JB Weld, then set the gear on its side (the 5C fixture has a 90 degree body, so you can work the piece either way) with the strip down. I had discovered that the JB Weld flows with gravity before setting up, and I wanted to use that. So, I put a "bottom" on with the wax-papered metal strip. Then I carefully pressed the sooted mold over the JB Weld, clamping it into place. It was gratifying to see the surplus putty run out the top, forming a nice little drop which would then tend to run down and fill any interstices before hardening. The next day, I took off all the clamps and the side piece, and carefully filed away the small excess amount of JB Weld. The advantage of letting it go out the side is that the side of the gear is a flat surface, which is very easy to machine. One of the guys who had given me advice on the Internet had suggested that you drill holes in the top of the mold, and put both sides on. I didn't want to fiddle around with cutting off any "sprues", or having them in any significant way impede me removing the mold.

Anyway, the next day, the mold came off cleanly, leaving a perfectly-formed tooth. It is actually not even very obvious, since the soot bonded with the epoxy to turn it black just like the gear.

The above procedure was done in 1997 and this gear fix is still working today.

metalmagpie

Axkiker
03-29-2015, 11:12 AM
Why don't you use your new CNC machine to cut the gears :)

LOL... that little thing couldnt cut a steel gear even if I had it running.

Axkiker
03-29-2015, 11:48 AM
My gear tooth gage set is labeled what the pressure angle is. Have another look.

I wrote up how I fixed a gear with broken tooth back in 1997. Have a look:


The above procedure was done in 1997 and this gear fix is still working today.

metalmagpie


I gotta say that is an ingenious way to fix a gear. If I had not already brazed mine up I would consider it.

Axkiker
03-29-2015, 11:58 AM
So if I do go through with making a cutter what steel can I use. At first I was thinking about using 01 as it would be easy to heat treat. I looked up some .250 x 3 x 18 which is the minimum length I have found and its over 50.00 without shipping. I can buy a full set of involute cutters off ebay for roughly 100.00

Im not very familiar with heat treating but from what I have read you cant really heat treat typical hot rolled steel that I have laying around. If I have to spend 50+ on the steel to make the cutters ill just buy the set off fleabay

Rich Carlstedt
03-29-2015, 12:12 PM
Why worry about hardening the gear ?
Just make it from mild steel. Less chance of breaking a tooth as well.
If you must have a hard face, just surface treat it (case hardening) with Kasenit !

Rich

Axkiker
03-29-2015, 12:14 PM
Why worry about hardening the gear ?
Just make it from mild steel. Less chance of breaking a tooth as well.
If you must have a hard face, just surface treat it (case hardening) with Kasenit !

Rich

Sorry, maybe I wasnt clear... I meant hardening the gear cutter

Axkiker
03-30-2015, 10:05 PM
So after a little more research im finding that I will need to either buy the 01 stock to make the gears or buy them already made. Anyone have a supplier for small quantities of 01 such as something around 3"x6"x.250 ??? Unless someone has another recommendation as to another steel I could use.

Otherwise it seems more economical to just buy the gear cutter kit off fleabay.