View Full Version : The worm turns

02-26-2004, 01:54 AM
I just finished making a worm wheel on my SB9 for a upcoming project. Just under 4" diameter, 120 teeth, 10 tpi acme form. It took me a while to figure out the jig but is has worked out pretty well. I used a very rigid fly cutter with a very carefully ground acme thread form tool. The blank is on a common mandrel with a 60 tooth change gear on a bearing attached to the bottom plate. The plate is attached to the cross slide inclined at the correct helix angle with a (not visible) large set screw to accurately adjust the angle. A screw bears on the change gear teeth to index it. After a full circle of 60 teeth were cut the index screw was moved over .050 to cut the teeth in between the first set. It actually seems to have worked well.

I don't usually like to talk about projects in progress. This is the hard part and the rest will work out, I hope. I am making a 4" rotary table with some maybe unique features, possibly with provision for stepper drive.


02-26-2004, 02:08 AM
Nice work again Evan!

Since you don't want to talk about your projects in progress but already have, can you elaborate on this nice little rotary table? Considering submitting it to HSM/MW? I'd like to see a little more.

02-26-2004, 02:17 AM
So far there is no more to show. Stay tuned...

G.A. Ewen
02-26-2004, 02:32 AM
As usual your workmanship is top of the shelf. I look forward to seeing the finished project.

02-26-2004, 04:37 AM
Nice work. Simple but well thought out indexing fixture.

Doc Nickel
02-26-2004, 06:57 AM
Yeah, I'd say that's a rigid fly cutter. http://bbs.homeshopmachinist.net//biggrin.gif


02-26-2004, 07:55 AM
Where do you get all that beautiful aluminum?

The local scrap guys want more for it then new parts cost.

A chunk large enough to try to copy a motorcycle head is about $600. A motorcycle head is about $550.

My digitizer and software still suck. thou.


02-26-2004, 10:45 AM
A very clever adaptation of various techniques. Not only is your fly cutter rigid (looks like less than 4:1 L/D) it also has a significant rotating mass which should help keep the rotation backlash to a minimum. Set the plate angle by a sine bar method using a dial indicator for the short leg?

Neil Peters

02-26-2004, 11:59 AM
We happen to have a local outfit that does a lot of custom fabrication, including things like aluminum trailers. Good source of stock. They are even going to be bringing in 7075 and some other alloys for stock. I was talking to the shop foreman and he was worried about how to identify it other than the markings so I told him to just spray the pieces with layout dye. "!!!" he said.


No sine bar needed, in fact it is not necessary to actually calculate the helix angle [ Arctan((lead/Dia)/2)]. For a 1/2" worm the slope of the helix is .10 to one. The plate is 4" wide. Elevate the one side .4". Done.

02-26-2004, 12:02 PM
I can't see how you indexing setup works?
The bottom gear appears to have fewer teeth than the top. I would expect the same on both.
Or are you doubling up in some way?
Tom M.

Paul Gauthier
02-26-2004, 12:09 PM
Well done Evan

Paul G.

02-26-2004, 12:52 PM

At the left of the change gear is a block with a pointy screw threaded through it. The point of the screw goes into the teeth to index. The change gear is 60 tooth and I wanted 120 teeth. After cutting 60 teeth I loosened the two screws holding the block (oversize holes) with the index screw and moved it over the right distance so I could cut the next 60 teeth in between the first 60 teeth.

Not shown in the pic are a couple of other things. There are a couple of packing pieces under the plate to support it. A bar with two small C clamps was clamped across the change gear for each cut to make sure it stayed in place.

Paul Alciatore
02-26-2004, 01:45 PM
Fantastic! Thanks for the visual and explanation. It's amazing what you can do with a little thought and effort.

I certainly learned something today. Thanks for shareing.

Paul A.

02-26-2004, 01:45 PM
Evan, start writing some articles for HSM, the gene pool is getting a little shallow there, since Terry Sexton seems to have stopped writing.

It must be very iritating to the Deckel and Monarch types to see what a real macninist/toolmaker/artist can create with a 9" s.b. You are an inspiration to me.

02-26-2004, 09:36 PM
Nice work,So thats why you have been kinda quiet lately http://bbs.homeshopmachinist.net//biggrin.gif

02-26-2004, 10:32 PM
AWE inspiring!
Now, if that gonzo mandrel was between centers, it wouldn't flop around! LOL http://bbs.homeshopmachinist.net//wink.gif

02-26-2004, 10:58 PM
IBEWGYPSIE - What size aluminum do you need? I am close to a scrap dealer that has a device that analyzes scrap and determines composition, and therefore knows "what it is". What alloy do you need? Aluminum goes for around $2.00 lb.

02-26-2004, 11:01 PM
Evan you could right a "How to" book, about lathes,
Title would be 1,001 things that can be done with a lathe.

Nice work ...

02-27-2004, 03:43 AM
Very nice work Evan. That looks like a great

02-27-2004, 03:55 AM
Beautiful in simplicity and execution...Worm wheel dividing attachment for the lathe spindle?

02-27-2004, 04:12 AM
Yeah Weird,

I won't go into details about the several failed attempts to make this work. It could still use some improvement. I am not satisfied until the cutting action is like slicing butter. Funny thing is that when I try something, even if it takes a lot of time to find out that it doesn't work, I don't get upset. I just try something else. Beats the hell out of doing it for a living.

This worked but required a very slow infeed. More support between the blank and the change gear would be good. The mandrel I used is 1 1/4 dia between the change gear and blank with the bottom end turned down to accomodate the 5/8 hole in the change gear. Almost too flexible. The top end is a fit in the 1 1/8 hole in the center of the blank.

The height was calculated to place the center of the pitch circle on lathe centerline when place at the helix angle. If I do it again like this I will use a mandrel that is large enough to rest upon the outside edge of the change gear and will give much more rigidity to the setup.

I don't expect to do this again, that is why I am building a rotary table. This is a one off bootstrap job.

The worm will be mild steel and the wheel and worm will be run together with fine abrasive valve grinding compound until the fit is perfect. I will use progressively finer compounds until the surfaces are polished.

02-27-2004, 09:03 AM
Evan, nice work as usual. I do see some potential problems however.
Using an Acme thread form for the flycutter will not produce an involute shape. With the high tooth count, the gear approaches a rack, but engagement problems may still exist. These could result in wear or backlash.
The high tooth count produces a fine gear that will not transmit as much power as a coarser gear. Most RT's use 60 or 90 teeth.
The large flycutter produces a gear tooth on the worm gear that has a larger radius than the smaller diameter worm I imagine you intend to use, reducing the contact area of the worm and worm gear.
These factors may add up to produce a gear train which may wear or fail prematurely. It will probably serve to position the table if locking it for an operation such as drilling bolt circles, but if used for milling circular shapes, wear and backlash may be problems.
A better choice might be a 60 tooth worm gear free hobbed with a hob made to the same diameter as the intended worm. The posting on hobs has an excellent write up on doing this. Your fixture will serve as the base for doing this.
I have made a few worm gears by this method, and it has the advantage that the hob and worm can be made with the same set up, the hob being slotted, hardened and ground will produce an exact tooth form for the worm gear.

02-27-2004, 09:20 AM
Nice job. I hope all works out well. rock

02-27-2004, 12:34 PM

I am well aware of all the points you make. The fact the teeth are not involute will be somewhat alleviated by running them together with lapping compound. I intend to use the table for positioning and as such the load on the gear is minimal. It engages 5 teeth at once on the worm so the load is well distributed. I used 120 teeth for several reasons the most important being that I intend to direct drive the worm with a 200 step per rev stepper and wanted at least one arc minute resolution.

That will give a positioning accuracy at the 4" diameter of ~ +-.0005

[This message has been edited by Evan (edited 02-27-2004).]

02-27-2004, 10:10 PM
Evan,at work we use run in compound thats made from chicken bones,it keeps losing abrasive action until it disolves,it doesn't embed like silicon lapping compound.

I was thinking last night about your setup,if it where done using a rotary table driven by a driveshaft that was geared to the lathe spindle..........then if the table could bob up and down like it a milling attachment it would be a gear hobbing machine,nah,too sexy,couldn't possibly work for me http://bbs.homeshopmachinist.net//biggrin.gif

02-29-2004, 09:48 AM
That's great work, Evan! OK, I'll bite. How does one identify different alloys of aluminum with markout dye? I'm currently in the process of making a fly reel spool for my brother-in-law. (Reverse engineering is all I'm doing since I'm copying his original Daiwa.) He scrounged a 2" diameter piece of aluminum and I have no idea what allow it is. I did note, however, whenever I apply markout dye to it the dye sort of beads up... it acts like rain on a freshly waxed car. By your comment above, does this sort of reaction indicate a certain type alloy?

02-29-2004, 02:23 PM
The problem the shop foreman has is to make sure that each alloy can be easily identified on the rack even when cut to shorts so that the original markings are not visible. What I proposed is that he spray the new stock from end to end with a unique color for that alloy (say, blue for 7075) so it would be easy to tell from the unsprayed 6061. Cheap and easy and avoids mistakes (big price difference!). He liked the idea. Nobody is going to object to some layout dye on their stock.

BTW, on your question, it means there is some oil or wax on the metal. They use it to prevent "slip marks" on the new stock.

[This message has been edited by Evan (edited 02-29-2004).]

03-01-2004, 07:12 AM
Your setup for indexing the gear teeth looks great. But the problem I see is you did not explain very well that the fly cutter generates an involute worm gear shape by sweeping by the gear by means of the correct feed screw rate.
This is very well described in a a book my library has called " Metal Workers Benchtop Reference Manual " by Joseph Serafin. It has nice referance tables that show based on the pitch you use the change gear ratio to get the correct feed. They all are based on a pi or 22/7.
This system was used years ago because of the simple tool shape and a hob for a one off job would be vey expensive.
I intent to use this system to repair 2 teeth of a brased up worm gear.
Anybody intending to do this should get a copy of the book that has lots of other good stuff as well as this procedure. My library system has it so my cot is O.

03-01-2004, 09:05 AM
An Acme thread shaped flycutter will not produce an involute shape if the gear blank is stationary.
I am not familar with the book mentioned, but it sounds as if the blank is advanced and turned as the cut is made so it acts as a single tooth hob. This would produce an involute gear tooth.
To make a flycutter with an involute shape, get the Ivan Law book on gearcutting. This is a link to John Stevenson's description of the method. www.metalwebnews.com/howto/gear/gear1.html (http://www.metalwebnews.com/howto/gear/gear1.html)
There are several good links to using screw threads for hobbing in posting on hobs on this page.

03-01-2004, 12:01 PM
The involute shape issue is a non issue in this case. According to the involute tooth shape theory a worm is a rack and has straight sided teeth so the acme thread form is just fine on the worm. The wheel in this case has 120 teeth so the shape of an involute tooth would be a close approximation of a rack as well.

Paul Alciatore
03-01-2004, 01:05 PM
<font face="Verdana, Arial" size="2">Originally posted by Evan:
The involute shape issue is a non issue in this case. According to the involute tooth shape theory a worm is a rack and has straight sided teeth so the acme thread form is just fine on the worm. The wheel in this case has 120 teeth so the shape of an involute tooth would be a close approximation of a rack as well.</font>


I am afraid that the involute is the issue for the worm wheel. It's true that the worm itself does have the same straight sides that a acme thread has but usually with different dimensions. However, for a worm wheel that is to mesh with that worm the tooth shape does need to be a true involute.

I understand your idea of grinding/lapping them in and if you start with a coarse enough grade of abrasive and run it for a sufficient time, the two mismatched shapes will tend to work themselves together but it will likely remove as much material from the correctly shaped worm as from the incorrect worm wheel. Perhaps even more from the worm itself since the contact there is on 4 or 5 teeeth only while on the worm wheel it is on all 120.

The result of this will likely be a shape that will run together fairly smoothly but it will not give the uniform motion of a true involute. This is important in your case because you stated you wanted one arc second accuracy. This means that an error of only a tenth or so on the tooth form will likely destroy that goal by advancing or retarding the rotation by a significant amount.

On the other hand, the high tooth count is on your side in this because, as you say, it is fairly close to a rack in shape.

Just a seat of the pants guess but I suspect you will wind up with an error more like +/- 10 arc seconds or more at some points.

How do you plan to test this?

Paul A.

[This message has been edited by Paul Alciatore (edited 03-01-2004).]

03-01-2004, 01:48 PM
I hadn't really thought about testing the accuracy but that shouldn't be too hard. Fasten a block to the perimiter of the table, set up a dial test indicator and step the table one step at a time with the stepper motor. Ten arc seconds will be around 5 thou so if it is off anything like that it will be easy to read. Even if it is off it will be a periodic error which can be easily compensated for in software. Telescope drives use PEC routines (Periodic Error Correction) to remove exactly that sort of error from the worm drives for the tracking system.

03-01-2004, 03:35 PM

I agree with paul. It is likely that the position accuracy will vary over the angle traveled for the engagement of a tooth -- if the profile is not involute. In other words, an occilating variation of spur gear angle will occur with worm gear angle change. You should be able to test this with the methods you mentioned.

Plot the angle change of the spur gear vs. the angle change of the worm gear on a piece of graph paper for measurements over a complete tooth distance. Ideally this will result in a straight line. Significant deviation should show up distinctly as a curve.

The setup you have seems to be ideal for hobbing as others have mentioned. Perhaps you can make your own hobb when you make the worm gear? Now that would be an interesting, high precission bit of toolmaking!

Good luck with this. I'm interested in how your assembly works out.

03-01-2004, 05:25 PM
I have revised the fixture to improve the accuracy and rigidity. I will post some pics tonight perhaps. And yes, it should work for hobbing.

03-01-2004, 07:20 PM
since you need to turn your own worm anyway, just turn a long worm, cut a length off to become a hobb, gash it, hardenen it and hone it and redo the wheel with the worm hobb. You can still use the blank you cut, the gashed worm can cut the worm wheel better than trying to grind it in, and it should come out beautiful.

The roughed out worm wheel should provide you an excellent basis to complete with the home made hobb.

BTW if you want to get some good bronze blanks get in touch with these fellers
www.spectrummachineinc.com (http://www.spectrummachineinc.com) or 1.888.bronzes they can supply even certified goods in tube, bars,plates, gear blanks (finished and semi-finished) oil impregnaated, and powder metal form. These guys are good. If they ask where you heard about them tell a big hairy Canadian saw them in the Febuary "GEAR Solutions"

03-02-2004, 03:02 AM
There is more to this story but here is the "new improved" Mark II version. I will be sidetracked for a bit by the new micro lathe.