Basic Clock Design

[AjohnwParticipant @ajohnw51620]

A 1/2 sec beat pendulum works out at 248mm Michael – assuming the sums are correct this time.

B'ham library doesn't keep copies either, nor EiM. I do have several books with details of the Brocot in it but none so far really provide an answer to what happens when the pendulum continues to swing in what ever direction it's being impulsed in which it will tend to do. All make the point about the tangent and mention dead beat. I'm hoping Russell or some one can shed some light on that.

I can't edit the earlier post. The length of the arm also influences the impulse arc, longer –> shorter arcs.

Might be worth adding that the GIMP  can be used for animations. A bit tedious though but images can be rotated down to 1/10  of a degree so  there is enough resolution once the parts are drawn.

John

–

Edited By Ajohnw on 01/05/2016 16:09:19

[Russell EberhardtParticipant @russelleberhardt48058]

John,

The leading edges of the escape wheel teeth for a Brocot escapement are normally made radial so, if there is over-swing, the pallet just slides down the face of the tooth with no recoil and no damage. That is unless the over-swing is enough to cause the pallet to hit the base of the tooth cut-out! Pins can be fitted to the rear plate to limit the swing of the crutch and prevent damage. Of course over-swing creates more frictional loss in the escapement and is to be made as small as possible. In other words the driving force must be sufficient to give reliable running but not much more.

Russell.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Ajohnw on 01/05/2016 15:59:28:

A 1/2 sec beat pendulum works out at 248mm Michael – assuming the sums are correct this time.

.

Just testing

MichaelG.

[AjohnwParticipant @ajohnw51620]

I'm not sure I have explained well Russell. Maybe best to take the Graham first. The impulses are applied near the centre of the pendulum swing but only for part of it. The actual pendulum swing is larger than the swing while it's being impulsed. The escape wheel is still locked and the pallets slide over the wheel's teeth. Niel's animation shows that clearly.

I assume that the full swing on a Brocot has to be greater than the part where it's being impulsed. Maybe it doesn't have to be or can be so close that it doesn't really matter and it's just down to setting up or I haven't viewed it's action correctly.

However

The Brocot animation shows nothing but the impulse part of the pendulum swing. In practice the pendulum will swing further. In other words rather than stopping in exactly the right place for an impulse the pendulum will swing further causing it to go deeper into the tooth space and the point it meets at will no longer be tangential = some recoil.

The infuriating Laurie Penman does mention this after a fashion. Having a rant I bought his repair book which isn't bad but why chuck the same sort of thing into the escapement book. I bought that a long time ago too. Pays to look at books closely before buying. There are several infuriating aspects about it. He mentions the above effect but reckons it's negligible when correctly designed – not much help on that in places really – but also shows what happens if the pallet goes way too deep. He also points out that as the lock and impulse faces are the same they can't be exactly tangential in both cases.

He does mention something interesting. Some one he knows built a clock using a Brocot escapement with a 12" dia wheel, 8 day clock and it ran on a 12lb weight. Perhaps the drive power prevents the pallets from going too deep.

Any reading this shouldn't be put off it really because ease of making something which does need to be pretty precise is important.

John

–

[Russell EberhardtParticipant @russelleberhardt48058]

OK John, I see what you mean now. The pallet doesn't move in a straight line but describes an arc centred on the pallet frame pivot. For small movements it approximates to a straight line but if you have a large over-swing it will cause recoil. I don't think it is usually a problem but you could avoid it entirely by using two escape wheels, one for each pallet. You could then give the leading edge of the tooth engaging with the leading pallet a slight concave form and the trailing one a convex form.

Spanning a greater number of teeth will move the pivot point further away from the escape wheel and thus increase the radius of curvature of the pallet path and give some improvement.

Your earlier mention of impulse angles did confuse me a bit. Assuming zero friction (yes I know it's impossible) the impulse vector must always be perpendicular to the surface of the point of contact on the pallet face. So for the Graham escapement it remains constant but for the Brocot it changes smoothly through 90 degrees from lock to release.

Russell.

[AjohnwParticipant @ajohnw51620]

I think it's a case of following a lead from a thread on another forum. Build the escapement first, add a small pulley, mount it in something with an embryonic pendulum, wrap some string round the pulley and add a SMALL weight that may need changing and observe it's behaviour.

I've also been looking around at crutches – hope that's the right name, the part that connects the escapement to the pendulum. I saw one arrangement that fitted into a sort of keyhole shape at the pendulum  widening upwards. That could be arranged to provide clearance as the pendulum swings out. So far no info one where to put the part that engages with the pendulum. I'd be interested in opinions but as I am going to be very busy for the next few days may not post back for a while.

Designing Brocot isn't that difficult. It starts with a circle divided into twice the number of teeth. As the pallet is moved further out there is more space for the them. It's based on the diameter neglecting the semicircular part usually plus 10% but if in numeric terms this looked to be a lot more than the accuracy I could achieve I'd be tempted to make it smaller, say 5%.

The next thing is the impulse angle. For a given wheel as above that is set by the length of the arms / pivot to pallet centre distance and the span – the number of wheel teeth between the pallets. Clockmaker's were never taught maths at school or failed their exams miserably so draw a line between the to centres and then another from the centre of the pivot and the edge of the pallet diameter on the wheel side. Measure the angle and that is the impulse angle. Usually done it seems by drawing it all out at large size or via cad. The diameter of the pin can also change this.

The pallets enclose a whole number plus a 1/2 teeth. This sets the angle of the flats on the pallet. Draw a line from the centre of the wheel through the active side of a tooth on the right hand exit side. That is the angle of the flat on the pallet. If drawn the flat edge of the D (which would be the other way round) would be directly on this line just about to rest on the tip of the tooth. The round part of the D on the other side would be centred on the tip of a flank and touching it apart from any clearance that has been built in. This image shows it differently. The wheel needs rotating to align the flat on the right hand side with the flank of the tooth. It's just drawn the other way round.

 

The old American book I linked to mentioned a catch, well sort of. As the pendulum gets longer and the impulse angle gets shorter the length of the arms go up. He feels this isn't horological as there is an increase in weight. This does suggest not making the wheel any bigger than needed which depends on the pallet pin diameter. What he really means is that the arms tend to get longer for shorter impulse angles.

It's interesting to note that Harrison didn't like short pendulum swings but I suspect his arrangement including the type of remontoire he used on the escape wheel will need his cheeks as well. Adjustable curve pieces of metal or probably wood in his case with screws pressing on their tips for adjustment.

His remontoire is very similar to the electronic one I posted except it carries a cam with 4 depressions in it which are used to trip a rewind every 15 secs in his case. It drives the escapement. A simple weighted lever rests rests on the cam. On a weight driven clock it could be used to trip the drive train and the remontoire could be on the wheel driving the escapement as per the electronic one. Out of interest the person who designed the electronic remontoire reckons the 2nd hand moved precisely in time with the pip on the radio for many days at a time – I'm not sure when they stopped broadcasting the pips.

John

–

Edited By Ajohnw on 02/05/2016 15:19:09

Edited By Ajohnw on 02/05/2016 15:26:58

[AjohnwParticipant @ajohnw51620]

Not thought through but remembered that Harrison liked lots of teeth on the escape wheel. 240 from what I can gather which means the 2nd hand needs 4 pointers and 2nd's are marked out on a dial covering a 90 degree arc. As one hand passes the 59 the next one moves to 0.

Surely the number if teeth changes the impulse angle as well as it moves 1/2 a tooth at a time in time with the pendulum? So having more than the usual tooth count to pendulum relationship could be used to influence impulse angles.

He had some feelings on where the crutch should engage the pendulum as well. Not seen much else on that which is why I asked for opinions.

John

–

[Michael GilliganParticipant @michaelgilligan61133]

John,

Not in answer to any of your specific points, I'm afraid, but maybe of interest:

If you haven't already found it, here is an 'accessible translation' of Harrison's "Concerning Such Mechanism …"

The original is a 'Brain-dump' and very difficult to follow … This version is considerably easier.

MichaelG.

[AjohnwParticipant @ajohnw51620]

Thanks Michael. I've read it and the original. I've admired him for a long time – an empirical engineer as I would call him. People who don't just accept what they have been taught and assume everything else is wrong.

I just had a look at what De Carle has to say about Brocot's escapement. Makes good sense given some free pendulum swing after an impulse. Using the image I posted as a reference he would say it's incorrect and too deep and leave the entry pallet where it is but swing the exit pallet out a touch but not enough to allow the escape wheel to free run. That way the pallet meets the tooth at an angled part of the curve rather than square on. It only needs to be say 2 degrees of pendulum swing so isn't that much. He also mentions that the flat pallet faces should be radial to the wheel and draws them this way with the round faces positioned as outlined. This can be adjusted after it's been built.

On dead beat sliding friction is added to all of the free pendulum swing after the impulse.

De Carle reckons that many astro type regulators of "european" manufacture used the pin wheel escapement. It uses rads in a similar way to the dead beat. A link I posted gives the sums some where from memory. Also that the pivots in English ones were often jeweled and used end plates to avoid friction from the shoulders on pivots running against the plates. Also tighter bearings generally.

He was more or less around when these things were being used which might help. "practical clock repair" a clock maker suggested I should buy it used some time ago even if making a clock. It's mostly concerned with making the parts plus some odd bits and pieces on details. I'm going to try involute so some of that isn't appropriate. I do reckon that depthing will have to be a lot more precise and any short comings in dividing gear might stick out like a sore thumb.

LOL – Just add "Basic Clock Design", When it comes to design generally things aren't that basic. It doesn't appear to be too difficult to figure out the why's in this case – which helps – sort of.

John

–

Edited By Ajohnw on 02/05/2016 18:50:32

Edited By Ajohnw on 02/05/2016 18:53:52

[James AlfordParticipant @jamesalford67616]

I have noticed that some Brocot escape wheels have straight teeth, whilst others have one side scalloped out. Am I right in thinking that the cut-out is to accommodate the pallets if the pendulum over-swings and to avoid recoil?

Regards,

James.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by James Alford on 02/05/2016 22:28:11:

I have noticed that some Brocot escape wheels have straight teeth, whilst others have one side scalloped out.

.

I expect to be corrected pretty promptly if this is nonsense, but I think the scallop is to clear the pin … thus allowing the maximum possible diameter of pin to be used.

The example in this video illustrates the action very nicely, especially in the close-up view in the second half.

MichaelG.

[AjohnwParticipant @ajohnw51620]

Penman reckons that the curve should be cut first and then a slot added. The slot is sized to clear the D shaped pallet completely if the pendulum over swings for some reason. The curve starts at circa 30 degrees at the tip of the tooth and blends into an arc. It's sized to make sure that it clears the pallet when the other size is impulsing what ever happens. A straight side at the angle needed would do the same thing but the curve leaves a wider tooth base.

Sounds complicated to do but all pallet movements are arcs of the escapement pivot.

He also mentions that some clocks use a friction clutch to drive the pendulum. I have seen photo's of ones like that. It's just a sleeve that fits over the pivot. However I have a French 5" ( I assume French inches) clock movement that is fitted with the escapement parts Brocot sold and no signs of this on that. It's stamped Brocot and has the unusual pendulum length adjustment at the top. It's a bit poorly and not feeling well but the escapement is fine. When messing about with escapements it's often possible to make them free run and bend the teeth on the escapement wheel so I don't see it being any worse than others really. On the other hand if some one makes a clock with an exposed pendulum it's probably worth doing adding the "clutch" what ever sort is used. It could just be a cylindrical part, split with some means of tightening it up. Maybe even just pliers before it's fitted.

Nice video Michael. The wheel on that might only be circa 20 – 30 mm dia.

John

–

Edited By Ajohnw on 02/05/2016 23:42:55

[James AlfordParticipant @jamesalford67616]

John and Michael: thank youfor the clarification.

James.

[Russell EberhardtParticipant @russelleberhardt48058]

Posted by Michael Gilligan on 02/05/2016 22:51:21:.

I expect to be corrected pretty promptly if this is nonsense, but I think the scallop is to clear the pin … thus allowing the maximum possible diameter of pin to be used.

The example in this video illustrates the action very nicely, especially in the close-up view in the second half.

That's a nice clear view of the action. The pendulum is clearly overswinging by a considerable amount, the pallets entering well below the locking point.

The scallop isn't there to clear the pin, only the half pin enters the escape wheel and there is plenty of clearance for that. I think it's just decorative.

The active face isn't radial but ct back slightly. Presumably to move the pivot for the pallet frame further away from the escape wheel pivot for some reason.

Russell.

[Russell EberhardtParticipant @russelleberhardt48058]

Posted by Michael Gilligan on 02/05/2016 17:40:19:

If you haven't already found it, here is an 'accessible translation' of Harrison's "Concerning Such Mechanism …"

The original is a 'Brain-dump' and very difficult to follow … This version is considerably easier.

Thanks for that Michael it makes very interesting reading.

He makes an interesting point which has always confused me. He states that adding fresh oil to the escapement for a clock with a small pendulum swing will make it run faster, without any explanation. While if the pendulum swing is greater it will slow it down.

Now I understand that a pendulum period slows down if the amplitude increases but why should the opposite occur for a small swing? Intuitively it seems right but I can't see the logic

Russell.

[jaCK HobsonParticipant @jackhobson50760]

A possible explanation: Circular error is not linear with pendulum amplitude. For small swing there is very little error – oil may make the swing slightly larger but additional error is still insignificant – however, the oil may influence other mechanisims in the clock, some of which make it faster. For large swing, the oil may give the same % change in swing but the % increase in circular error is much greater and this becomes the main source of change in behavior, swamping the other effects of oil.

Edited By jaCK Hobson on 03/05/2016 08:18:06

Edited By jaCK Hobson on 03/05/2016 08:18:51

[Michael GilliganParticipant @michaelgilligan61133]

This thread is worth a look **LINK** … especially Post #8, which includes a delightful illustration from Goodrich.

MichaelG.

.

Also … See this video [animation] and the accompanying note.

It illustrates the 'clearance' aspect of the tooth shape rather better than my words.

Edited By Michael Gilligan on 03/05/2016 08:39:12

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Russell Eberhardt on 03/05/2016 08:02:07:

… He makes an interesting point which has always confused me. He states that adding fresh oil to the escapement for a clock with a small pendulum swing will make it run faster, without any explanation. While if the pendulum swing is greater it will slow it down.

Now I understand that a pendulum period slows down if the amplitude increases but why should the opposite occur for a small swing? Intuitively it seems right but I can't see the logic

.

Russell,

I think Jack is on the right lines: Harrison was a believer in large amplitude pendulum swings, and felt that the small angle was [what we now call] a 'workaround' which, though it avoided various problems, required unreasonably close control … He preferred to address the underlying issues.

In this example, it's about [what we might now call] the Signal to Noise Ratio of the pendulum & escapement system: A low amplitude pendulum has low 'signal' so the 'noise' of the re-oiling is significant; wheras a high amplitude pendulum has high 'signal' …

MichaelG.

.

The better that we understand mechanical timekeepers, the more obvious Harrison's genius becomes.

 

Edited By Michael Gilligan on 03/05/2016 09:05:27

[Russell EberhardtParticipant @russelleberhardt48058]

Agreed, adding a drop of oil thus reducing the friction will reduce the frictional losses and thus increase the Q factor of the system (the ratio of stored energy to energy lost each cycle). That will result in a narrower bandwith in electrical terma and thus increse the signal to noise. In this case the noise is random variations of timing. It doesn't explain why it will always go faster.

I don't deny that it is true. I'm just trying to find a logical explanation! Maybe my 70 year old brain is just running slow I must add a bit of oil (red wine and sunshine)!

Russell.

[AjohnwParticipant @ajohnw51620]

The low swing pendulum circa 2 degrees minimises circular error as Jack pointed out. Very heavy pendulums were used as well to minimise the effects of the changes in air pressure.

Harrison had another answer to the problem which was to make the pendulum behave sensibly even if the swing varies. It has to given the way his escapement remontoire works.

I have seen comments about flattening etc the end of Brocot teeth. A tiny change because in my view if taken too far it could stop it from working correctly. Think what the angle shown in Michaels link would do when it meets the pallet. A truly tiny flat on top of the tooth even sloping the other way would make more sense.

 My 0.6mod gear cutters just arrived. Bit more to go though.

John

–

Edited By Ajohnw on 03/05/2016 09:39:58

[Martin KyteParticipant @martinkyte99762]

Regarding the "adding oil question" All the escapements discussed so far are for pendulums with a small swing of only a few degrees. They can be viewed as essentially isochronous. That is the period is independent of the amplitude. The addition of oil to the escapement reduces the drag on the pendulum caused by the pallets moving across the teeth of the scape wheel and thus causing the pendulum to "speed up". He also criticised Graham for his uses of a large scape wheel which made the ratio of the pallet distance to the pendulum length too large and allowed variations in the escapement to have an overly large effect on the pendulum. (We are back to signal to noise ratios again)

Harrison used large swings and therefor was well into the realm where circular error was a factor. He realised that variations in the density of the air, as temperature and pressure changed, would change the period of his pendulum He therefor controlled rather than eliminated circular error by using curved cheeks each side of the suspension spring and was able then to use the residual circular error to eliminate the effect of drag on the pendulum moving through the air.

So the less dense air causes less drag speeding up the pendulum but the resultant bigger amplitude slows it down resulting in a near constant period.

Good innit :0)

regards Martin

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Russell Eberhardt on 03/05/2016 09:21:32:

Agreed, adding a drop of oil thus reducing the friction will reduce the frictional losses and thus increase the Q factor of the system (the ratio of stored energy to energy lost each cycle). That will result in a narrower bandwith in electrical terma and thus increse the signal to noise. In this case the noise is random variations of timing. It doesn't explain why it will always go faster.

I don't deny that it is true. I'm just trying to find a logical explanation!

.

I'm struggling a bit, Russell … but see if this makes sense:

Although I mentioned SNR, I suppose that in the case of re-oiling the escapement we are really talking about a small non-random message [lubrication >> less wasted energy >> more impulse] buried in a larger 'carrier'.

The low amplitude system; like a racehorse or a MotoGP bike, is sensitive to tiny changes in any factor.

The high amplitude system is relatively insensitive. [which is Harrison's key point]

MichaelG.

.

… Now add the 'red wine and sunshine' and tell us the answer when you have it.

[AjohnwParticipant @ajohnw51620]

I hate to see electronic ideas being used to explain something they can't.

A simpler way of looking it that pendulum looses some energy each time it swings. The escapement puts this back in which involves it generating a force. The force is generated by some systems of gears which will have friction as will any bearings. Add oil and there will usually be an increase in force because the friction has changed. Changes in viscosity and amount will also have slight effects. We are talking rather minor errors here not large ones.

As to the different action of Harrison it might be down to all sorts of things. He was heavily into reducing friction and used some unusual bearing arrangements. There is a multipart video on you tube showing a replica regulator being built. It shows one aspect of his low friction bearings. 2 rather large discs forming a V and the pivot rests in it. The disc rotatate much more slowly then the pivot so the bearings that carry these impart less friction. The other aspect was lignum vitae.

John

–

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Ajohnw on 03/05/2016 10:29:07:

I hate to see electronic ideas being used to explain something they can't.

.

Then just look away, John

People sometimes understand things better when they re-think them in analogous terms with which they are comfortable.

MichaelG.

[roy entwistleParticipant @royentwistle24699]

I think we have got away from a basic clock

[Author]

Posts