cutting spur gears on a mill

Well PR post a pic of a hobbing attachment - not a specialised hobbing machine - that doesnt count, its not for the hobby machinist.

My point here as I expect you noticed is that we all take for granted that a lathe is generally designed to be able to make screw threads. Its no big deal

But a mill isnt set up to make gears in a similar way (nor a lathe for that matter)

In both cases you can go for a dividing head and laboriously rake out each tooth

I dont know of a simple system that allows you to cut gears with ease and speed (I am discounting the modern use of stepper motors for the job for the purposes of this discussion. No electronics allowable

You need to lock the spindle rotation with the gear blank rotation by suitable gear chain and the whole should run and chew out a gear - you may need some movement of the cross slide if its a thick blank of course

Thats hobbing on gas

So is there a ready made attachment for mill hobbing?

Hence my pursuit of a cheap charlie way of knock up spur gears - we already know the worm gear method

and this is an expansion of this method

Indeed, looks like cutting a helical form, the rotary axis is driven from the table X feed not the cutting spindle.

Brian, you have not said how long those gears took, could you enlighten us as to how long each took and did you take any photos during the machining?

Interesting that you seem to have about 47 teeth from a quick count up, with a 13tpi cutter and a 27.3mm dia blank I would have thought 42 teeth, so the method is very unreliable.

From the one I cut I would not say it's super fast as I found the feed along the work needed to be increadably slow to get an even finish rather than a "shallow worm wheel" effect in a helical form along the face related to my feed rate.

The milling setup with the driven dividing head is one of Andrew's photos for one of his hobby projects, not sure he has an arm chair either as his front room has too much engineering tools in it to find space for one.

Edited By JasonB on 16/09/2021 07:12:54

As a spectator, a long and interesting debate, from which i have learned a few things.

1 IF this method is to be used, a spiral tap is probably the better way to go.

2 There are some quite short fuses about.

3 To maintain an accurate NUMBER of teeth (Not necessarily Form ) pre gashing is required.

Since this is required, will this be a lot quicker than using a gear cutter, considering the need then to reset from some means of Dividing to allowing the blank to rotate?

This ignores the fact that most of us will have a slitting saw, and the likely absence of the required gear cutter, and the cost of buying one.

A while ago, I did consider using a tap for gear cutting, but since a definite ratio, and an accurate gear was required, I took the coward's way out and bout a spur gear and a worm of the same Module..

Horror since the two will not be an exact match. But for unloaded very low speed operation, this is considered to be a small problem.

For other spur gears, I shall continue to use gear cutters. So far, cutting a gear has never taken as long as reading this thread. .

Howard

Posted by brian jones 11 15/09/2021 22:27:24

Well PR post a pic of a hobbing attachment - not a specialised hobbing
machine - that doesnt count, its not for the hobby machinist.
My point here as I expect you noticed is that we all take for
granted that a lathe is generally designed to be able to make screw
threads. Its no big deal
But a mill isnt set up to make gears in a similar way
(nor a lathe for that matter)
In both cases you can go for a dividing head and laboriously
rake out each tooth
I dont know of a simple system that allows you to cut gears
with ease and speed (I am discounting the modern use
of stepper motors for the job for the purposes of this
discussion. No electronics allowable
You need to lock the spindle rotation with the gear
blank rotation by suitable gear chain and the whole should
run and chew out a gear - you may need some movement
of the cross slide if its a thick blank of course
Thats hobbing on gas
So is there a ready made attachment for mill hobbing?
Hence my pursuit of a cheap charlie way of knock
up spur gears - we already know the worm gear method
and this is an expansion of this method.

Have a look at Jeff Thyers gear hobbing unit made
for use on Maureen seen in MODEL ENGINEERS'
WORKSHOP magazine, issues 258 & 259
and at the video here.
https://www.youtube.com/watch?v=vuXyHe3VUpY

I cut many gears on my Doreen using the
methods seen in the book Gears and
gear cutting ,the gear cutting is all based
on using this machine with a little ingenuity
and the Jeff Thyers setup you may be able
to adapt the machine to hob gears on your
Doreen.

John

If the OP is going to misappropriate my photos he might at least attempt to get the descriptions correct. The first photo is not of a gear hobber. It's a universal dividing head set up to cut a high helix angle helical gear, as shown here in more detail:

And the finished gear:

Such dividing heads are normally used on a universal mill, ie, one where the table swivels. The table on my universal mill only swivels ±45° so in order to cut a gear with a 70° helix angle it was setup on the Bridgeport with a right-angle head. This equipment would have been common in commercial workshops up until at least the 1980s when CNC started to make inroads. There are many forum members using ex-industrial equipment, so I doubt I'm the only member with this sort of capability.

In the second photo "caught" is hardly the right word, given the photo is in a public album. The idea that increasing the gear OD makes it a spur gear is nonsense. Long ago (50+ years when I was still at school) I cut a much larger gear using the same method, with a straight flute 2BA tap, just to see if it worked. As in the photo the larger gear was also a worm wheel.

It would be interesting to know how the OP would tackle this gear:

It would need one hell of a tap!

An underlying theme in posts by the OP shows an obsession with sofas and armchairs. Even if I had them nothing he's posted so far would have stirred me to get off them.

Andrew

But Brian, you should know of a simple system! I've mentioned it at least twice in this thread, provided a picture, and linked to Neil's Orrery Thread where he used the method to make many gears.

On the 20th August, I typed:

Using a tap to cut gears isn't new. I've experimented with it, but got much better results with a Sunderland Rack cutter and Rotary Table:

Neil Wyatt made a lot of gears this way for his Jovilabe Orrery a few years ago.

Unlike a tap, Rack Cutters get the number of teeth right and, with a little care, they cut an accurate involute. Rack cutting and Hobs do take longer!

Boils down to how much the operator needs an accurate gear as opposed a quick approximation. If gashing gears with a tap is 'good enough' for the job in hand, go for it! No-one objects.

The much despised armchair is a valuable engineering tool. They're good for research and thinking.

Is Mr Sunderland's method too complicated? It requires a:

1. Lathe and V cutter ground to turn the Sunderland Rack Cutter
2. Mill, milling cutter and horizontal rotary table (or other indexer) to relieve the Sunderland Cutter. (Could be done with a file.)
3. Mill with horizontal rotary table or other indexer, plus chuck to hold the gear blank.

The results are predictable. This home-made gear has the right number of teeth, it sits on ½" Meccano centres, and it meshes with the commercial plastic gear without shredding it:

Hobbers are preferred for mass-production because they churn out well-made standard gears quickly and efficiently. Sunderland is mostly used by industry to cut big bespoke gears, often by planing, but same principle.

Sunderland is slower, but there's no need to make a special hob, and it's possible to generate accurate involutes. The Sunderland method can be created in a small workshop without too much trouble, no pre-gashing is required, and results are reliable and predictable. I'd expect Sunderland to be faster than Brian's free method in practice because Sunderland produces fewer rejects. Sunderland wins whenever the centres, or exact number and shape of teeth matter. Sunderland versus Hob is a different debate!

Dave

Dave, I have made a single tooth, four cutting edges rack cutter (I know it doesn't make sense put like that) using a ground and polished carbide insert to turn a round piece of HSS. Coupled with a CNC mill with an A axis it is easy to cut gears, just a bit slower than a multi-tooth rack cutter, but since it is CNC you don't need to stand over it. As with a lot of this work more time is spent in the preparation and set-up, the actual machining time is not always the issue. The advantage of the single tooth is that one cutter with the correct pressure angle will do lots of different module or DP gears.

With CNC a lot of productive work is done sitting in a comfortable chair.

Martin C

If the gear profile can be drawn in CAD, then a special cutter isn't needed for CNC. Just use a standard ballnose endmill:

Andrew

Brian I only pointed out your error so that the uninformed don't become the mis-informed, please don't take it personally.

There isn't a simple solution because it's not a simple process. You don't find ready-made attachments for mill hobbing because the fundamental requirement for hobbing is that the work rotation be timed to the cutter rotation, and your average milling machine doesn't allow for that (there may be some ancient machines that take the table drive from the spindle drive by a geartrain and if you find one of them then you could possibly adjust the gearing to suit).

Your mention of the worm gear method bears some exploration. Unlike straight spur gears which can be cut with any diameter cutter or hob a worm gear needs to be cut with the same diameter cutter as the mating worm, or else the helix angle is wrong and tooth engagement departs rapidly from ideal. This is very prohibitive for the home engineer as buying custom cutters is very expensive so the most viable options are:

1. Make a hobbing cutter from hardenable steel that mimics the worm you will make with added clearance. Harden it and grind the cutter teeth sharp. Not a trivial exercise but doable though a lot of work for a single item.

2. Make a single-tooth cutter to the form and diameter of the worm, tilt it to the helix angle and cut each worm wheel tooth separately. A very laborious task.

3. Make or modify your design so that a stock thread can be used as the worm and then a stock thread cutting tap can be utilised as a cutter to hob the teeth onto a pre-gashed wheel.

No3 is often used out of necessity because it can give a good return of effort vs reward in terms of producing a workmanlike result for very little time and effort, The 'wrap-around' nature of the wormwheel teeth allows for a decent flank area for the worm 'cutter' to self-drive the wheel (unlike the spur gear's point contact) and if the material isn't too hard and the work well lubricated can produce very nice results. If the worm is pre-gashed (always recommended) you'll hit your desired tooth count every time.

The reason why pre-gashing is recommended for free-hobbing is because the pitch of the cutter (or tap, which is a hobbing cutter in this respect) is fixed but the pitch of the teeth on the wormwheel is larger at the OD than at the base circle and only matches the cutter pitch at the pitch diameter. When you start cutting the tips of the cutter will cut the edge of the blank and drive it somewhat however because the hob's pitch which is correct for the pitch diameter but starts cutting on the much larger OD it will naturally try to divide the OD into a greater number of divisions (teeth) than you require. Pre-gashing the work allows you to start your hobbing at a depth that matches the pitch of the cutter more closely to the pitch of the wormwheel at the pitch diameter which has two benefits:

1. You get the tooth count you want

2. The cutter will cut more equally both sides.

You see, by starting your cutting at the outside diameter the distance between each cutting start is 'short' for the tooth spacing at the OD, in order to continue towards full depth the cutter has to continuously shift the tooth profile around the blank away from the driving flank. When your cutting depth gets to the pitch diameter the pitch on the cutter is now correct for the cutting diameter and it will naturally cut the correct number of teeth. As you pass pitch diameter and approach full depth the pitch of the cutter is now 'long' for the diameter at the cutter tip which is why you get under-cut on small tooth count gears (not that you're going to get under-cut at 30 degrees pressure angle).

So, if you can gash the blank to at least the depth of the pitch diameter you stand a much better chance of getting a good result from your free-hobbing exercise in terms of correct tooth cound AND quality of cut because the load on the cutter and on the driving flank are greatly reduced.

So you see, worm-wheel fre-hobbing is borne of necessity but free-hobbing spur gears is not because there is no real requirement to. As you have demonstrated, it is actually possible but the result (subject to personal opinion) most often fall far short of what constitutes an even moderately well-made gear.

This is my personal take on the matter. Feel free to disseminate, disagree or criticise my thinking I will not be offended.

Pete.

Well guys its all good input and shows we are addressing a serious issue - making spur gears

For this discussion I discount CNC and stepper motor drives as this is a different diocese IMHO

I also discount the lathe chivving method cos its not on msg and you can make up some complex jiggery to hack away - you cant buy one off the shelf btw

Ialso discount geared hobbers - the tools alone are £150 - thats all way too expensive

So this is for Doreen

We need to focus IMHO. Remind you of the SOW

A quick cheap way to make a serviceable spur gear from a non ferrous blank. It does not have to be perfect form, only for light load and speed. The issue of No. of teeth may not be so important if its only say within a few percent for some applications - especially model making as these are never put to live use (as admitted by one of our luminaries)

More work is in hand AWS as I improve my method of chivving bolts

None of the armchair pundits has spotted the limitation of this method - meh

Vouchsafe I may, but I have a cunning plan

OBTW to answer JBs question, yes it did take me ca 10+ mins, but I was cautious at my first attempt. I think I can up the spindle speed x2, and it will surely benefit when I make a properly made 6 slit bolt, not off and angle grinder in a vice jobbie. Work for 2mrw

Edited By brian jones 11 on 17/09/2021 01:55:41

JB no. of teeth issue

Q

Interesting that you seem to have about 47 teeth from a quick count up, with a 13tpi cutter and a 27.3mm dia blank I would have thought 42 teeth, so the method is very unreliable

UQ

Yes the theocratical number would be 42 for starting at 27,3mm dia allowing for pcd using a 13tpi thread

but what happened in practice for my brass blank is that some rolling took place and my OD increased to 28.07 gives predicted 45 teeth but I actually got 46

But my blanks are only 17p so a bit of trial and error may result in predictable repeatable results

Now this may improve when I use my proper cutter

Hold that thought

13TPI =0.077CP=40.8DP (41DP)

27.3/25.4=1.075" diameter blank

41x1.075=44.1 +2 = 46.1 so you should have indeed got 46 teeth on a 27.3mm blank the 28.07 diameter should have yielded 47.

Certainly not cutting as one would expect .

PCD of a 41DP with 46T is 1.122" or 28.5mm

OD of a 41DP with 46T is 1.171" or 29.74mm

If we take Brian's new OD of 28.07 and his 46T then he has cut a 43.4DP gear with a "hob" that should have cut 41DP gear's CP is 0.072" or just under 14tpi

The amount of extrusion you have got with the increased OD is a sign that the cheap and cheerful bolt method is not working too well as you are not cutting cleanly through the metal but forcing it into submission and just making the unpredictable nature of the method even more up in the air

Edited By JasonB on 17/09/2021 07:21:31

Indeed Pete I am rushing my calcs and must pay more attention

Yes Jason if you saw my rough attempt at a spiral chivved bolt, its a wonder I got anywhere near right ball park

but will try harder

Have any of you spotted the inherent fatal flaw in my conjecture yet?

Not that difficult although it did take some time; I didn't need to grind anything. Reliefs were added with a file before hardening and tempering:

The worms and worm wheels, for scale the worm wheels are around 4" in diameter:

Andrew

Something intriguing I noticed, but i expect you all knew this, is that a rack and pinion system uses a 20deg PA form but the rack is a straight 60deg form.

Well AJ excluding cnc I cant see how you made that on Maureen BTSOOM

The helix was cut on a CNC mill for historical reasons, but everything else (hob and parts) were machined on a manual lathe and a big old horizontal mill that cost me £175.

Andrew

“Something intriguing I noticed, but i expect you all knew this, is that a rack and pinion system uses a 20deg PA form but the rack is straight 60deg form.”

No this is wrong, if using a pinion with 20deg PA then the rack also must have a PA of 20deg.

Where do you get the 60deg from?

Jan