How Would I Machine this CAD designed Ratchet

[SillyOldDufferModerator @sillyoldduffer]

Thinking about an odometer-like mechanism recently I came up with a ratchet drive in which a pawl pushed a wheel one click by pressing on a flat face like ‘C’ whilst riding up a curved opposite slope ‘A’ until it dropped down into the next notch.  (The pawl moves left to to right on a hinge and is spring loaded)   Something like this:

 

Though surface ‘B’ is rounded, wouldn’t matter if a cutter spoilt that.

Even so, how could those curved ratchet notches be machined?  I can’t imagine either a straight or form cutter doing it however tilted.  Have I smacked into a design that’s easy to CAD but hard to make for real?  Or am I missing the obvious?

Ratchets like this exist, so there must be a way.

Help!

Ta,

Dave

 

 

 

[JasonBModerator @jasonb]

A form cutter a bit like a 1 sided involute gear cutter would probably do it if the centre section could be soldered into a ring so that the cutter could have run out room or have the centre recessed to just below the valley of the teeth

You would also have to watch that the cutter’s diameter was not too large otherwise it would hit the opposite side

We did have  asimilar thread about hole saws which need a similar sort of cut

If you could have a small radius in the internal corners then that opens up more possibilities

[JasonBModerator @jasonb]

Something like this would cut the tooth profile, just index it round.

[Paul LousickParticipant @paullousick59116]

Is the boss in the centre of the ratchet necessary?  Because it is very difficult to machine into a blind hole that has a sharp corner.

I made a ratchet wheel for the lubricator on my engine by cutting teeth on a long piece of rod with a chamfer tool, then cutting a slice off the end for the ratchet. (lots of spare material for future projects).

[Julie AnnParticipant @julieann]

As drawn the size of the gap seems to get smaller towards the centre. So a form cutter will not work. Even if the blank is tilted slightly a form cutter will probably not generate an accurate profile. It will be a compromise, like cutting a bevel gear on a manual mill.

If the item must be as drawn then get it 3D printed.

Otherwise back to the computer!

Julie

[SillyOldDufferModerator @sillyoldduffer]

Feels like back to the computer.

This is the shape of the removed metal, which I hoped might suggest a form cutter profile, nope!

Jason’s ‘D’ cutter does what Julie feared, I think.  Could be cut but the teeth won’t ratchet, the yellow lines aren’t parallel:

Thanks for the suggestions, seems tricky!  I’m off to sleep on it.

Dave

[Phil PParticipant @philp]

I made a radial toothed clutch for my little barring engine which was a similar problem. The only way I came up with was using a slotting head and a dividing head linked to the X axis of the mill.

 

 

[Michael GilliganParticipant @michaelgilligan61133]

Not sure [and I am certainly making no claim to fame] but I wonder if Jason was recalling the question I asked years ago about cutting a verge crown-wheel for a clock.

If so, it was never fully solved.

I will have a look for the old thread next time I wake-up in the middle of the night.

🙁

MichaelG.

.

Edit: __ ‘twas much easier to find than I anticipated:

https://www.model-engineer.co.uk/forums/topic/how-do-i-cut-a-verge-crown-escape-wheel/#post-212482

… still sipping a Glenmorangie

Edit: __ some of the content in that 2015 thread has gone awol so, for completeness I am quoting an observation I made in a related eMail:

Although JasonB and Mark C both provided very useful sketches … on close examination, both of them actually ‘demonstrate the problem’ rather than answering the question.

Mark was particularly helpful, and provided me with some more-detailed drawings; but the big ‘niggling’ question remained. [Zoom-in on a tooth in his SolidWorks rendering, and you will see that the tooth-tips are inappropriately shaped.]

[John HaineParticipant @johnhaine32865]

Change the shape of the pawl to fit a space you can cut?

[Nigel Graham 2Participant @nigelgraham2]

The crown-wheel form would be the easiest, as the cutter runs right through the gullet and into the fresh air in the middle.

Alternatively, can the wheel be of more conventional form with the teeth cut across the disc thickness? (Shaped like a circular saw.)

Your original form is not an example of something being easy to generate in CAD but not make physically. You can do that in manual drawing too; and it seems quite a common mistake.

[duncan webster 1Participant @duncanwebster1]

Blank on rotary table on a shaper. Form tool cuts face C, then turn it round a bit to machine face A. You might need to adjust the x axis as well depending on exact design. You need a groove between the inner edge of the teeth and the boss for cutter runout. The contact face twixt wheel and pawl makes line contact for most of the stroke, so you might as well have the teeth a lot narrower.

Enigma was not like this was it?

[Michael GilliganParticipant @michaelgilligan61133]

At the risk of stating the obvious, Dave … The fundamental source of your difficulty is that the shafts are at 90°

Ratchets are so much easier when the shafts are parallel.

Go Compare: https://www.automotioncomponents.co.uk/am/products/linkages-rotary-components/pivot-bearings-air-motors-cylinders/other-rotary-parts/ratchet-pawls/r4040/groups/g+m+c+s+a+view

… now engage your inner Production Engineer 🙂

MichaelG.

[Speedy Builder5Participant @speedybuilder5]

Spark erosion machining ??

Bob

[JasonBModerator @jasonb]

If both the flat and curved faces need to be at different radial angles then like Howards shaper method, go round once then offset by the required angle and offset again.

Some sizes would help OD of wheel and depth of tooth

A Thortons ratchet wheel cutter used on the end would do it

Or add a small fillet to all the internal corners and then it can be done on the CNC with a ball nosed tool.

[SillyOldDuffer]

On 23 February 2025 at 20:03 SillyOldDuffer Said:

[…] The pawl moves left to to right on a hinge and is spring loaded

Then unless the pawl is ‘trivially thin’ there would be a range of different velocities required at its contact patch with C, all at the same time … this being impossible I think it would naturally revert to making line-contact and scrubbing its way across C

[ you probably need to model that ]

MichaelG.

[Tony Pratt 1Participant @tonypratt1]

The pawl looks pretty but if it has to be exactly as drawn then spark eroding is the only method I know which will replicate the forms or maybe lost wax casting both of which are going to be expensive.

I would go back to the drawing board 🙂

Tony

[Charles LamontParticipant @charleslamont71117]

I assume the pawl is mounted on something which rotates about the same axis as the ratchet. I started off thinking the pawl was moving in a straight line, as the description says it moves from left to right, in which case the geometry would be all wrong. I think MichaelG has made the same interpretation.

For a rotary motion of the pawl, A is a helical face cam. If you don’t mind a radius at the bottom of the notch, it can be milled radially with a small slot drill. This one is for the indexing pin in my lathe toolpost (for scale, the thread is 9/16″, and sorry about the photo, it was done so long ago that it is real film):

[SillyOldDufferModerator @sillyoldduffer]

Gosh the team’s been busy whilst I slept!  All noted, thanks, though I won’t address each individually.

Very impressed with Phil’s work, though not quite what I was after:

Duncan asked if this was Enigma.  Thought it was the answer at one stage, but, apart from one variant Enigmas don’t have a gear train.   Instead each rotor has a drive pawl, all pushing in parallel, except pawls 2 and 3 can’t engage with their ratchet until a notch comes into alignment and lets the pawl drop in:

The original pawl and ratchet aren’t hard to machine, but I didn’t know that at the time.   A relief, because the odometer images I found all seem to have an awkward bevelled ratchet of the form I’m asking about.

Jason, Tony, and others  suggest compromise solutions, which I’m open to!   But I feel they dodge the issue by making thin teeth on an outer diameter, which hides the problem.  As drawn by me, with wide flat teeth,  the ratchet has triple curved surfaces – only one face is flat, and the rest are all twisted curves.

John Haines has me thinking. Fit the pawl to whatever can be cut!   But I don’t know what that is!   Actually the pawl isn’t the problem – a ratchet isn’t a precision power-transfer mechanism, so the pawl can be quite crude.  As drawn by me the pawl is a 45° hook, and the ratchet gear is a projection of it.   In the example, I didn’t bother to project the hook to fit exactly, but could do as ‘proper job’.   CAD makes it look easy.

Nigel thinks this isn’t an example of 3D CAD creating an impossible to make object, because the same “mistake” can be done in 2D .  I suggest he’s “not wrong”.  Solid Edge generates the geometry mathematically, hiding the  complexity behind a few buttons pressed naively by me.  In sharp contrast I have a book describing how 2D draughtsmen should taclke drawing curved objects, and it’s not for the faint hearted.  The need for mathematically correct curves aren’t necessarily a “mistake” either – aerofoils, cams, and some gears call for them, even if they do challenge the tool-room!   That said, if Nigel or anyone else can draw this in 2D, please bring it on!

Michael is correct about scrubbing too, but I don’t think it matters in my application.   Would mess up a clock though…

I looked to Meccano for inspiration, their Part 148 being easy to make with a form cutter (such as Jason’s Thornton):

Meccano’s contrate opens another possibility, I think this one is made by cutting a conventional spur wheel edge on, and then pressing the disc in a die to bend the teeth through 90°:

 

In answer to Jason’s query about dimensions.  To fit my enigma the ratchet wheel needs 26 teeth on a diameter anywhere between about 25mm and 80mm.  Small would be easier to match to the drive mechanism, but there’s much choice.   How deep the teeth need to go sideways into the wheel depends.  Far enough to accommodate the pawl, so a small proportion of a large diameter wheel, but proportionally much more into a small diameter wheel.

Many thanks!

Dave

 

 

 

 

 

[JasonBModerator @jasonb]

So for anyone wanting to play with this a 52mm outer diameter, 30mm inner boss 26T

[Graham MeekParticipant @grahammeek88282]

It looks no different to a Verge escapement wheel, except that it has a boss in the centre. Such a boss could be added later. The outer form could be produced in the same way as the clockmakers produce these types of escapements.

I made one for John Wildings 16th Century clock using my Unimat 3 with just a simple form cutter.

Regards

Gray,

[Nigel Graham 2Participant @nigelgraham2]

Dave –

I thought similarly to Graham’s idea.

Assuming the centre has to be of that shape (full height), can you make it two parts?

Machine the toothed outer as a cylinder so the cutter has somewhere to go (it may need a special cutter making), then solder that to a cylindrical inner.

…..

The design “error” I asked of, has nothing to do with how the item is drawn. It can be a “fag-packet” pencil sketch, drawn on a manual drauighting-machine or by a full industrial CAD package.

It is simply that you can draw anything any way you like and find it was easy to draw but a nightmare to make.

A work colleague once showed me a drawing he’d been issued. “Tap M3 X 3mm deep full-thread”, it said, of a strip of aluminium only 4mm thick. I also saw a drawing for a component resembling internally, the inside of a cone pulley. One step half-way down the cavity called for cutting a very small rectangular pocket into the surface, with very small corner radii. I am not sure if 2mm d. X 150mm l. end-mills are available. These were professionally-made drawings!

I’ve made similar mistakes: a lovely drawing but… how the heck do I make that bit?

[SillyOldDuffer]

On SillyOldDuffer Said:

 

 

Jason’s ‘D’ cutter does what Julie feared, I think.  Could be cut but the teeth won’t ratchet, the yellow lines aren’t parallel:

 

 

Dave, they are not parallel on your original either. What is it you actually require

[Charles Lamont]

On Charles Lamont Said:

I assume the pawl is mounted on something which rotates about the same axis as the ratchet. I started off thinking the pawl was moving in a straight line, as the description says it moves from left to right, in which case the geometry would be all wrong. I think MichaelG has made the same interpretation. […]

@Dave … This ^^^  is a very important observation

My interpretation was [and is] that your pawl somehow slides [on a detail not shown] linearly AND naturally rotates about its own arbor [also not shown]

Could you please clarify ?

MichaelG.

[Michael Gilligan]

On Michael Gilligan Said:

On Charles Lamont Said:

I assume the pawl is mounted on something which rotates about the same axis as the ratchet. I started off thinking the pawl was moving in a straight line, as the description says it moves from left to right, in which case the geometry would be all wrong. I think MichaelG has made the same interpretation. […]

@Dave … This ^^^  is a very important observation

My interpretation was [and is] that your pawl somehow slides [on a detail not shown] linearly AND naturally rotates about its own arbor [also not shown]

Could you please clarify ?

MichaelG.

That part of the problem doesn’t worry me.  The ratchet is pushed anti-clockwise when the pawl moves from left to right, and on return right to lift, the hook rides up and over.  The mechanism is like this, not modelled correctly in that the pawl should be longer, but it animates OK.

First view, the pawl is moving left, has ridden up the slope, and is sat on the top about to drop into the next notch:

It started from deep inside the notch, having pushed the ratchet one notch by pressing on face ‘C’.  Slider ‘A’ is  moving right and constrained by a block, but able to move up/down at hook ‘B’ due to the hinge:

Top view next showing the pawl at an angle to the notch, and why the two red lines are near parallel.  Not critically so but not wildly divergent, because that would delay the time the pawl rides on top.

 

The action is similar but not identical to a ratchet or a verge.  In a ratchet a pawl prevents rotation in one direction, whilst a verge intermittently releases a wheel via a rising and falling pawl.  In both cases torque is applied by the wheel.   In my application the wheel free-wheels and the pawl shoves it round one notch, before returning and dropping into the next cavity.  The pawl driver slides L-R, and the pawl is hinged to ride back over the reverse slope.   I don’t see much problem with that, my problem is how to cut the ratchet wheel.

Sorry about slow reply.  Nothing but interruptions today!

Dave

[Michael GilliganParticipant @michaelgilligan61133]

Thanks, Dave !

MichaelG.

 

Edit: __ Thinks … the required action is ‘not dissimilar to’ the count-wheels in Synchronome and Pul-syn-etic clocks [except for that pesky 90° turn you have introduced].

MichaelG.

[Author]

Posts