This is a starting point.  To date I haven't managed to print an accurate enough escape wheel but I have made improvements to my 3D printer recently so I'm ready to have another go.

I have further tested and adjusted the 3D printer and changed the centre hole in the escape wheel for better fit on the shaft and things do seem to be better.

The escape wheel has printed pretty well but when spinning it on it's shaft there is some variation in the points of the teeth but I think this is due to the difference in filament laying angle around the wheel rather than the accuracy of printing.  A Cartesian printing system is not the best for a polar form of object.

I've looked into the various escapements to see which would be best for 3D printing but most were superseded by later ones on account of accuracy. I've mainly been trying the Anchor Escapement but have tried the Deadbeat as well. It is said that the Anchor escapement is less bothered by tolerances than the Deadbeat so the Anchor seemed best.  I have tried some of the less popular ones too.  However, even the Anchor escapement is critical of any variation in the radius of the escape wheel teeth and this is what has been the problem so far.  Either the pendulum stops or it skips teeth, depending on the torque driving the escape wheel and the Anchor to wheel spacing. 

Having not been able to get the Anchor escapement working and with its inherent problems I think I may have another go at a Deadbeat type.  Also, it has occurred to me to "fudge" the issue by driving the pendulum separately from the escape mechanism in the style of a Slave Clock.  If I go this route I think the Deadbeat escapement would be better. 

Looking at the action of the deadbeat escapement I think it may actually be easier to get this right the the acchor.  It would seem to be less critical of the tooth radius.  Basically, the tooth wants to slide off the anchor when the pendulum is at it's centre position.

Start of a design for a deadbeat escapement.

Hi Gina, just to get me following the thread....

The bit that is often overlooked is that the tooth passing across the lower surface is what gives a kick to the pendulum.

Hi Neil.  Yes, that's how it works.  Great to see you here.

I found this on Thingiverse, apologies if I have posted it before. It took some tweaking to get it to run, I can see the clock will be a much bigger challenge!

Apologies for the tongue-in-cheek 'clickbait' video title too!

I've seen it before somewhere.  Well, it's a sort of clock... 😀

Been designing the escapement.

Now printing the parts to see how the fit together.  I'll make up a test rig.

Deadbeat escapement, showing: (a) escape wheel, (b) pallets showing concentric locking faces, (c) crutch.

Hopefully the deadbeat escapement will work but otherwise I might have to try something that's more tolerant of conditions such as used in tower clocks.   Gravity Escapement.

I think the deadbeat escapement stands a good chance of working if I can get my printer to behave.  Otherwise, it would seem that the Gravity Escapement has little to go wrong.  It would need a serious redesign of the whole clock though.  It would certainly be different and that makes it appeal to me.

I wonder... 

@Stub Mandrel  What do you think Neil or anyone else?

The Gravity Escapement would not depend on any critical 3D printing and it would be different.  The clock is already pretty different from a "normal" longcase clock - they don't have all their works on view.  As those who know me know, I like to be different.

The escape mechanism would be below the main gears where it can be seen rather than behind.  Everything would be rearranged.  The hour gears would go centre top where the current escapement is, instead of at the side.   I shall need to have a good think about the rest.

This escapement turns once in 6s rather than the 60s of the deadbeat or anchor escapement so will need more gearing.

This shows the design as far as I had got before including most of the striking mechanism (green and pale blue levers etc.).

The pink and purple gears are the intermediate minutes to hours and can be swapped with the going train gears (red and blue in this model).  The striking system could probably moved a bit higher to reduce clutter.  The auto-winding system, starting with the green motor gear at the bottom, will be replaced with something simpler.  Maybe worm drive.  I hadn't got as far as the chain drive for the striking mechanism.

Actually, I'm wondering now if the striking mechanism doesn't make it all too cluttered.  Anyway, I plan to get the main timepiece working first, then add any extras.

Some calculations :-

Centre wheel has 80t, chain drive wheel has 20t so chain wheel takes 15m per revolution.  The gravity escapement shaft takes 6s per rev. so if we start from the chain wheel, the overall gear ratio required is 15x60/6 = 150:1.  150 factorises into 2x3x5x5.  Suitable gear ratios would be 6x5x5 - easy to implement.  This is assuming I do without a seconds hand.

Had the clock apart and done some measuring.  I may be able to get a gear of 170mm or even 200mm diameter on the chain wheel.  This might enable me to get the required 150:1 gear ratio to the escape wheel in 2 rather than 3 stages.  Currently, the 60:1 ratio from minutes to escape wheel seconds uses 80:10 and 75:10 gear pairs.

Copied from the previous Blog, this shows how I had designed the gears around the minute shaft.  One error - Great Wheel should read Centre Wheel.  The Great Wheel refers to the chain wheel.  This shows part of the striking mechanism which I shall think about.

With a gear ratio of 150:1 the nearest ratios work out at 12.5:1 and 12:1.  With a size limit just over 200mm this gives 8t and 100t at mod2 for the first pair and 8t and 96t mod2 for the second.  This would seem to be just about the limit.  I'm happy with mod2 - it's what I used in my Moon Dial Clock but I would rather have 10t pinions rather than 8t.  One solution would be to increase the ratio of chain wheel and centre wheel.  This would only mean reprinting a couple of gears - the 20t gear is attached to the chain wheel with a dog clutch constantly engaged.  I could keep the clear acrylic sheets that form the clock case if I keep the spacing the same.

Spacing is 113mm.  Some new calculations :-  From the Centre Wheel rotating one an hour to the new escapement rotating once in 6s we have a ratio of 600.  It would be nice to use 6x10x10.

This seems pretty good for the Great to Centre Wheel ratio.  ie. chain wheel to minutes wheel.  6:1 ratio and mod2 with gears of 16t and 96t.   Why this shows "design failure" I don't know - maybe an Autocad expert might know.  But it seems to work anyway.

So far it seems that I shall be replacing all the spur gears!

I have a pretty fair idea of how it's going together in two dimensions but the third I'm not sure of yet for the escapement.  The pendulum is behind the escape wheel yet the arms are in the same plane.  I guess the arms (or are they legs) have to bend back to reach the pendulum.

On 16/01/2021 at 18:45, Gina said:

Hopefully the deadbeat escapement will work but otherwise I might have to try something that's more tolerant of conditions such as used in tower clocks.   Gravity Escapement.

Ooh! That is so clever!

Do you like it Neil?  I think the deadbeat escapement is a dead duck!  To difficult to get right so I'm trying a design for the gravity escapement..

If you think the effort of an extra gear pair and design are worth it then I like  the gravity escapement, although I would add a lot of extra material for robustness compared to the gif.

I've worked out that I should be able to get away without an extra gear pair but with higher ratios - though I have to test that out. 

6:1 from chain wheel to centre wheel with the chain wheel going faster (can do this because the auto-winding can keep up with the faster chain drive).  Then two stages of 10:1 give the overall ratio from centre wheel (60m) to escape wheel (6s) of 600:1.  Or do you think I'm being too ambitious?