This blog continues from my threads in The Lounge, in particular  DIY Grandfather (Longcase) Clock Mark 2

I couldn't get the original (Mark 1) clock to work properly and also decided I was trying to cram far too much into it and it was very cluttered.  These points together made me decide to start again with just the main clock mechanism and get that working first then add other bits.  So I've dumped the perpetual calendar and that will go into a separate project.  I'm keeping the auto-winding mechanism as I'm too lazy to wind it, and the moon globe as that is no problem (much the same as the Moon Dial Clock).

I did another net search to to see if I could get any ideas for alternative clock mechanisms and found one with a reduced number of gears as shown below.  Fewer gears might help to reduce friction and would unclutter the clock a bit.  The doubt was whether the higher gear ratios would work but with a bit of testing I think they may.  One advantage of one less gear pair is that the escape wheel turns clockwise which would allow a seconds hand to be attached to the shaft.

I have been experimenting with printed gears to test the feasibility of fewer gears and I'm going to try it.  In the attached photo, the yellow gear is the centre/minutes wheel and drives the orange gear with a 10:1 ratio.  This in turn drives the escape wheel (light grey) with a 6:1 ratio giving an overall ratio of 60:1 as required.  The position of the escape wheel makes its shaft ideal for a seconds hand with  dial just below the 12 o'clock mark.  The smaller gear on the main centre wheel is for the chain drive.

Been designing the chain drive gear.   The first photo shows the gears laid out on the new acrylic panels and in the second, I have removed almost all the parts from the Mark 1 clock to make use of the panels and framework to test the gearing.  Once I have finalised the layout I shall transfer it to the new panels with the clock face further up the space.

The chain drive gear was too big so I've made a new one.

Calculation for the main drive chain sprocket which will be attached to the green gear in the photo above :-

1. Radius of the gear  = 84mm
2. Pitch radius if sprocket teeth wants to be less say 75mm
3. Link pitch = 20mm
4. Circumference of new sprocket at tooth pitch = 471mm.  Maybe 480mm which is 24 teeth.
5. New diameter of tooth pitch = 152.8mm so radius = 76.4mm
6. Angle per tooth= 15°

Produced the sprocket but the chain didn't quite fit well enough - the sprocket pitch was just a bit too long.  First photo below.  The other two show the gear arrangement with the gears roughly positioned in the original clock frame

Finished mounting the gears into the frame and tried it for friction.  There's too much for my liking, virtually all in the escape wheel bearings.  I don't think ball bearings are good enough, even 5x10x5 little ones, so I think it's PTFE as I had on the original.  I used 5mm polished stainless steel but I think I'll go for a smaller shaft.  I think I have some 3mm SS rod.

Yes, I've found a piece just under 100mm long.  I think that will be long enough to take a seconds hand on the end.  I have tried it in a couple of bits of PTFE and it's ultra smooth.  I had thought of having a larger escape wheel than the original but on thinking about it, this would increase the friction from the pallets so I think I shall stick with the 100mm OD.  The problem with the original was poor production - I need to get a 3D printer to accurately print Nylon, which has a lot lower friction than any other thermoplastic but is a lot more difficult to print.

I have taken a file to the teeth on the main chain sprocket and managed to get the chain to run on it.  Not perfect but it may let me test the clock  - I can re-print it later if required.

I have a shorter piece of 3mm SS rod that I might use for the anchor shaft to produce less friction than 5mm SS rod though I suspect the air friction on the bob and rod may be greater than a polished SS shaft in PTFE bearings.

Been thinking about the chain drive and alternative options.  Since the clock is being auto-wound continuously the reduction gearing from the sprocket to the minutes wheel does not need to be so great.  First I thought of putting the drive sprocket directly on the minutes wheel but I couldn't work out where the chain would go to provide a viable auto-winding system.  Next I looked into a separate gear for the sprocket but driving the minutes wheel on its main teeth rather than a separate smaller gear.  I'll continue looking at it...

My brain snapped into gear   And...  Inspiration...  I can use the sprocket from the original clock and attach it to a suitable sized gear.  The original won't do because it's mod 3 and the centre gear is mod 2.5 but just simple job   I can also use the smaller sprocket that is driven by the auto-winding drive.

The photo below shows the chain drive sprocket and also the escape wheel and anchor.  Below that are a couple of diagrams showing the the main drive chain arrangement and the auto-winding sprocket with its attached drive gear.  This shows that there is more at the bottom of the clock than at the top so it makes sense to move the centre 50mm further up the case (ie. the frame lower with respect to the works).  The main thing at the top is the moon globe but I plan to use the same design as for my moon dial clock, so it isn't included within the main frame.

I plan to use the space on the RHS of the case for the striking mechanism and its auto-winding system.  The drive for the moon globe will also be derived from the hour drive to the striking mechanism.  That will stop the drive lever from obscuring the main clock.

Chain drive gear designed - 56 teeth and 70mm pitch radius.  That means a ratio of 10:7 compared with the centre gear.  1.4x compared with 3x on the original giving roughly a 2:1drive weight advantage.  Of course the auto-winding ratios will need re-calculating.  I want a smaller gear on the auto-winding sprocket anyway as the original is too big for my liking.  I could use the same size gear as on the clock drive which would avoid awkward ratios when sorting out the auto-winding.

I've found out why the original escape wheel and the newer ones were out of shape.  The Pilot printer was significantly out of calibration.  X axis was fine at 100.04mm but the Y axis was short st 99.53mm so firstly I adjusted the Y calibration and put that right.  Then I checked the orthogonality of the XY axes - all of 2mm out in the diagonals.  That meant undoing the frame brackets and adjusting before tightening back up.  I needed to do this several times before I got it within a tenth of a mm.  This printer just won't hold calibration, hence the need for a new one.

Reprinted the chain gear drive as the first go was out of round.  The new one is better but not perfect.  I estimate it has about a half millimetre run-out which surprised me considering how carefully I calibrated the printer.  The centre gear printed on my Titan printer shows no discernible run-out.  The gear works well enough so I'll leave it at that.  I may try the escape wheel in Nylon.

Reprinted the auto-winding sprocket gear and the Nylon escape wheel, set up PTFE bearings for the escape wheel shaft and driven the shaft through the boss of the escape wheel, then set up the frame and acrylic plates to hold all the gears and escape wheel in position.  The friction is nice and low - I can spin the escape wheel with slight pressure with one finger on the chain drive gear.  However, the escape wheel is still far from right - the teeth deviate from a circle centred on the axle by a mm or more.  I shall check the calibration of the printer again.  If I can get decent prints I think this version of the clock might work.

Printed a calibration frame and measured it and the printer has gone out of calibration again already   0.5% loss of Y axis and 1.5mm difference in diagonals.  Enough to make the gear prints significantly oval!!  Evidently the problem must lie with the Y axis - movement of the print bed as the Y motion is obtained by moving the print bed.  I'm not going to get reliable printing until I get the new mini printer built.  I have ordered the parts needed for the new printer.  Meanwhile, I think I'll try my Titan printer with the smaller 0.4mm nozzle - that seems to keep its calibration.

Printed a new third gear on the Titan with 0.4mm nozzle and it's much better - can't see anything wrong with it   Might try the Nylon escape wheel or I might wait until I have a smaller nozzle available as the escape wheel needs to be very accurate.  It might be a good idea to use Nylon for the third wheel to reduce friction and maybe even go over to PTFE bearings.  Ball bearings are considered unsuitable for applications involving change of motion direction as with an anchor escapement though if I can get sufficient accuracy I might go over to a dead-beat escapement as that is reckoned to be more accurate.

Since posting last I have made good progress on the "GinaRep Mini" 3D printer but most projects are on hold until I shake off a heavy cold.  Also, this project has to take a back seat to astro projects and I need a calendar more than another clock - the Moon Dial Clock is working fine.  OTOH while I'm resting with this cold there is no reason why I shouldn't think about this project and after a considerable break I may have some new ideas :D. 

I have one serious problem with this pendulum clock - the pendulum and escapement.  I can't expect to make a perfect time-keeper with just a weight driven pendulum clock - even the most expert clock maker has not achieved this, so synchronisation to a true source of perfect time is what I have been trying to do.  So far I have not achieved this.  I know I need to make a better escape wheel but also I need to sort out a workable way of driving the pendulum from my time source.  Slave timepieces have been made in the past so I see no reason why I shouldn't be able to do the same.

Been looking again at slave clocks and those with pendulums seem to use solenoids mostly.  In my clock design I have been using levers and a spring running off the auto-winding mechanism which is timed from a Real Time Clock module.   Magnetic drive to the bottom of the pendulum might be better.

Another possibility I was trying before was to vary the effective pendulum length by lifting the steel spring support while still holding it in the gap.  This shortens the pendulum and reduces it's swing time.  I could measure the swing time and control the lift with a servo system until the swing time was exactly 2s.  I have worked with servo systems in the distant past and used to know all the maths involved including damping coefficients    I have a horrible feeling that I might be past that now!  What I don't want is an unstable system that deviates more and more from correct timing

Looking back at where I got to before, I see I didn't even get the clock to run continuously without any timekeeping control so on a practical level (when I get back to it) I shall need to produce a decent escape wheel and get everything set up for a running clock.

Thinking about pendulum timing control servo system, it might be worth doing this in two stages viz. coarse and fine :-

1. Measure the cycle time and adjust to make it 2s.
2. Compare pendulum phase with RTC phase and fine adjust to produce synchronism.

All this can be done with the Arduino I'm already using to drive the auto-winding system.  The mechanism for controlling the pendulum length could be a threaded rod driven by a stepper motor.

Having found that the pendulum won't synchronise unless the period is already very close to correct, it seems that some sort of period adjustment is required anyway.  In vintage clocks this is usually a thumbscrew on the pendulum bob to move it up or down.  Adjusting the other end is equally valid.  This could be manual or servo controlled (automatic) and I like a clock that looks after itself   Now since it seems that a servo controlled period could do the whole job of timekeeping it seems unnecessary to use magnetic synchronisation as well.

The above approach differs from my previous idea which was to manually adjust the period until close enough to synchronise by mechanical levers etc. from the auto-winder.

I might have abandoned this project as not being worthwhile in view of the amount of work required to get it to work but for two reasons - several of my friends are very interested in this project and also that I don't like to just give up