Basic Clock Design

I have use unidirectional carbon fibre re-enforced epoxy tube for the pendulum rod of my regulator. It is specified at -0.1 to -0.3 ppm/C expansion co-eff. I have compensated it by careful positioning of the pendulum bob suspension and a short length of steel in adition to the carbon fibre rod. I haven't been running it long enough to get any temperature data yet. I'll have to wait for the hot summers we get in the South of France. I did notice some initial creep at first even though the stress in the pendulum rod is only about 0.5% of its UTS. That settled down in the first week or so.

Russell.

Nice regulator Russel.

Martin

Yes, that is a very nice clock.

An article in a recent HSN mentions that carbon fibre is a little hygroscopic, and I remember hearing this from someone else too a while back. I'm not sure how significant this is. I wonder if it is possible to seal the surface or if there might be better resins? Have you seen anything Russell, or any affects?

On the question of pendulum amplitude, yes it is determined by the loss, not the mass as long as the losses are mainly proportional to bob velocity. Aluminium and lead bobs of the same size and shape should give the same swing everything else being equal, but the lead one will take longer to settle down. The desirability and relevance or otherwise of high "Q" divides horologists!

Thanks Russell. That fits in with what the 2 design book says - better than invar. Does the negative sign indicate it shrinks as it warms up ?

John

-

Yes, the negative sign does indeed indicate that it shrinks with increasing temperature.

As far as hygroscopic effects go I believe that there is a minute amount of swelling with water absorption however, with longitudal unidirectional fibres I would expect any swelling to affect the width and not the length of the rod. From my sailing days I remember that we used to paint glass/polyester hulls with an epoxy barrier to prevent osmosis so epoxy is certainly better than polyester resin. Personally I'm not worried about moisture as we have very low humidity in this part of the world .

Russell.

I suspect that the hygroscopic concerns would be weight changes rather than size. Looking around I managed to find this - aircraft research datat

It fits in with other comments - slow and takes some time to reach equilibrium. That source also mentioned that in this temperature range the effects are more rapid as the epoxy goes rubbery. At normal temperature it takes too long to form a useful test.

John

-

Yes, if the weight increases the cg of the pendulum will move upwards and it will slightly speed up. But 80c and 90% RHUL is quite extreme!

Interesting John. A quick calculation shows that if my 30g pendulum rod gains 1.5% weight the clock would gain about 1 second a day. However as the other John said 80 C and 90% RH is pretty extreme. Our house averages 20 C in winter and 30 C in summer with RH between 35 and 45% so I doubt that I have anything to worry about.

However, when I get the clock installed in its case ( the case just needs staining, waxing and glazing now) I will make a point of recording going rate against interior temperature and humidity. It will take a full year though to get any meaningful results though.

Does the paper you found give any information on the fill factor of the composite? Obviously the greater the proportion of fibre the less the effect will be.

Russell.

No I couldn't find anything like that Russell. From memory some time ago the fill factor for high strength can't be that high but I don't remember the numbers.

I did some sums on what I might do and it works out at 0.1gm. Two things would reduce that, temperature but I get the impression that it just takes longer to reach the equilibrium point and more importantly it wont gain that much water because the relative humidity will be lower. I can't see epoxy behaving the same was as a desiccant does.

I'm wondering about a "mini regulator" based around a 1 sec period pendulum but suspect the space for the drop weight will be insufficient for a reasonable time between winding. On the other hand I can't see why that can't be geared up.

John

-

Does any one have any ideas of what parts are about that could be used / adapted to make up an electronic remontoire. Preferably suitable for battery operation with a reasonable life.

John

-

John, do you mean rewinder or remontoire? And by electronic do you mean electric?

Having dug in my memory a bit about the Clock B presentation last year, I think actually what their rewinder does is the following. It doesn't use a differential, rather there is a dc motor driving a worm that engages with a wheel on the hour wheel arbor. The motor and worm is mounted on a carriage that is picoted on that arbor, and also has a long lever that carries the brass weight visible on the phot that you posted. Also on the carriage is a mercury tilt switch (or two?) that senses when the lever has dropped down to its "rewind" position and when it has been raised to its "stop rewind" position. Those could be replaced nowadays with an Arduino and an accelerometer for sensing and control (what do I mean "these days" - the clock was only completed a couple of years ago!). I would have thought it would be easy to calculate the energy needed, making reasonable allowance for the motor efficiency, gear friction etc, to see what life was possible with a battery supply. A small solar panel could recharge the battery (possibly supplemented by an LTD Stirling engine?).

Another approach that I've been mulling over is to reverse the system, so a small motor drives the input to the remontoire proper, keeping it wound to the correct tension, so the motor's running is controlled by the remontoire. Then the motor would drive the train from the "fast" end, opposite to standard practice. Potential problem with these schemes is motor lifetime, assuming a brushed motor.

If you are going to end up electric powered why not make the ME Jubilee clock. This is a starter for ten, it has been updated in later articles and there is a short pendulum version I think.

Edited By Bazyle on 26/04/2016 17:49:36

Thanks Bazyle. Nosing around I found these

**LINK**

**LINK**

**LINK**

**LINK**

The first one is interesting.

John

=

Sorry John. Last time I was reading this I just read the last post.

From what I can gather clock b used a harrison style remontoire on the escapement and some other method of rewinding. I think I've seen mention of every 20min for that but no mention of how other than by a motor. It's a pity there isn't a decent video of it about. Maybe some one close could go and take a good one.

I want what ever is used to be as mechanical as possible so a weight remontoire or some sort of rewind system would be ideal. The question really is the best and quietest way to drive it.

While nosing about I came across a totally OTT clock that has featured in the BHI magazine several times. It uses weights but with unusual drums.

Not thought it through but as it lets some out is winds some of it back in so more turns can be obtained with a given drop.

John

-

There are videos of clock B about

http://www.guinnessworldrecords.com/news/2015/4/video-how-%E2%80%9Cperfect-clock%E2%80%9D-redefines-timekeeping-history-300-years-on-377653

I guess copyright could be a little complicated. At the moment it is still inside the Greenwich Observatory workshop and only viewable through a window. Not sure if pictures are allowed in the museum. And the clock is only on-loan.

I think that is the only one Jack and too much chuntering and not enough of the clock really.

The winding drum can turn an 11ft drop into a 17" one. Lots of loss of torque though.

My maths are probably screwball there but it will do for now, For 132" coming off one 115 goes on the other,

John

-

AjohnW - yes, as I said, a Harrison remontoire on the escapement and the winding mechanism I described on the hour wheel. I guess the OTT clock you mention is the Schroeder clock, also by Martin Burgess, and recently described in HJ by Doug Bateman? I've tried making one of those drives - never got it to work, probably should have used a bigger weight.

The weights do look a bit big John. It's shown here and on youtube

**LINK**

With the sizes he gives it should give the same performance as a weight of 1/8 the size less any losses.

The clock made me thing of one of Michael's shiny gearboxes. Wonder how good the time keeping is.

John

-

Slowly progressing with my drawings. At the same time as designing the clock, I am learning CAD for the first time. I am using Onshape which seems generally to be all right, but seems to do odd things at times, for no apparent reason. I had to draw the pendulum suspension block several times as part way through adding other parts to the assembly, the block split into its components parts and appeared to fall apart.

The width between the frames is currently arbitary and will be set later once I have worked out hw large the winding barrel needs to be. Using the idea of a looped rope with pulleys, I should be able to get seven of eight days running time from a four foot drop.

I have settled on the following gearing. It may not be ideal, but it all seems to fit and look generally fine. I know that it still uses ten leave pinion in the train, but I am struggling to devise a better ratio that fits together.

30 tooth escape wheel followed by:

12:90

12:96

12:120 for the barrel wheel

I have 10:30 and

8:32 for the hour and minute hands.

As mentioned before, the frames are merely shown to hold the wheels and aroburs in place and are not the final design or size.

I don't know how many teeth you have on the escapement wheel but I was reading that the pallets usually span 90 degrees so 7 1/2 teeth on a 30T wheel and 15 on say a 60 etc. 7 1/2 becomes 7 in that spread sheet I posted.

Just a comment in a book. Like most things in this area not much info about.

John

-