Reinventing a Synchronome


John Haine
 
Edited

As I'm making some progress on my reinvention (if I may be so bold as to call it that) I thought that I'd post some pictures of the story so far.  I acquired a 'Nome round about the beginning of January, in not bad condition but lacking the pendulum and pallet, with the intention of using the parts that I had to make a clock on the same principle but updated.  Apart from the fun of doing my own thing, I wanted to make something that would be quieter and maybe better appreciated by the domestic authorities.  Apart from making a pendulum with a 7 kg cast iron bob and carbon fibre rod, the main differences are:
  1. The same basic impulsing system is used, but with a pallet shaped based on FH-J's theoretical description in Electric Clocks, as proposed I think by William Shortt.  This is designed so the impulse roller is placed gently on the pallet just before the impulse is applied, on a surface that is a circular arc struck from the pendulum pivot.  As the pallet moves further, the roller slides down a ramp shaped to give a smoothly increasing then decreasing force, around the centre of the swing, then the roller ends up on another dead roll, from which it's picked up just after the impulse.
  2. The count wheel is replaced with an opto interrupter, which will (if I can squeeze it in) sense the end of the pallet at the extreme righthand swing.  Swings will be counted by an Arduino, which will generate an impulse every N swings where N may not be 30; plus a pulse to drive the slave clock every 30 seconds.
  3. The gravity arm will be controlled by a spiral cam on the shaft of a small stepper motor driven from the Arduino, so as to place the impulse roller on the pallet just before the impulse, and lift it off just after.  Because the roller touches down and is lifted off when on a dead roll the impulse timing should be pretty insensitive to the exact moments of touchdown and liftoff.
So far I have made the pendulum, pallet + collet, the stepper mounting and the lifting cam, and a new cock for the gravity atm pivot that's much thinner than the original to make more room for the pallet collet, so all the main mechanical components.  Here's a photo of the assembly - there are more in an album under photos.  You can just see the spiral cam poking out of the side of the motor mount, which itself sits on two pillars set in holes on the bedplate that used to hold terminals.  I have accumulated a collection of 'Nome components that should not be needed, which will be looking for a good home.
.
With the basic mechanics built, now I need to start thinking about the Arduino.  I'll post more as the story progresses.

And sorry about the major surgery to an innocent Synchronome!


John Haine
 

I forgot to mention that the cam follower is a 6 mm ball race with its inner bolted to an M2 screw through one of the holes that used to be used to attach the ebonite insulator on the back of the gravity arm.  Stepper motor is an "08" size.


John Haine
 

That should say that the opto will pick up the pallet at the extreme left hand swing!


John Hutchinson
 

That's an interesting project :)
Why CF for the pendulum rod? Better coefficient of expansion? 


 

This is just beautiful. The development of the pendulum clock was driven by clockmakers relentlessly trying to stop the pendulum doing too much work so that the pendulum would just do the timekeeping. It reached its pinnacle with the Synchronome and its successor the Shortt-Synchronome. None of the developers in the past had access to optoelectronics but now we do.We could interact with the pendulum without touching it. How elegant to just use an opto sensor not only to count the number of beats but at the right time to drop a gravity arm and pick it up.  Do keep us posted as to how you get on and I'll be very interested to see what sort of accuracy you get from this modification that approaches the Shortt-Synchronome. You can even get a seconds signal from this.  Wonderful.


John Haine
 
Edited

Thanks for the positive comments.  Carbon fibre has a near-zero coefficient of thermal expansion - there is some discussion over on the NAWCC discussion boards, and though the coeff is variable between materials it is generally better than invar.  It is liable to absorb moisture slightly, which means its weight can vary, but what this does to timekeeping I have yet to work out.  I do have a spreadsheet that takes into account all the material densities and shapes to calculate the period so it would be simple to do (in fact I ought to sooner rather than later).  It's also MUCH cheaper than invar.

Edit: Now done the calculation.  According to an NAWCC post the maximum weight gain of CF due to absorbed humidity is 0.2%.  Plugging this as a density gain in my spreadsheet gives a change in pendulum period of 1.9 microseconds, which I think I'll forget about!  The thing is, carbon fibre is so light anyway, especially compared to the bob, that its contribution to the MoI is minute - so changes to density also have a small effect.

Actually the opto will tell the Arduino when the pendulum reaches its extreme of swing, then the Arduino works out when to place the pallet, but will do this on the dead roll before the impulse so the exact impulse timing depends only on the pallet profile.  That's the theory, anyway!


Rob
 

Hi John,

 I am interested in this project. I was wondering how you are going to connect the bob onto the carbon fibre rod. I have been interested in cycling for many years and there it has always been a task to bond carbon an steel or aluminium together. I would be interesting to see how this project ends up

Regards

Rob Lockyer


neil
 

Hi John
I have also been working along similar lines, but all home made. Hopefully yours will be running before mine is.  I have 3 optos ( as seen in the pix) to sense the BTD of the pendulum but also to detect when the amplitude falls off to a preset point, when i will impulse the pendulum again magnetically. ( an electronic Hipp toggle if you like)  The coil is on the LHS with the LED across it is the impulse electromagnet, which "sucks in" the steel rod.  I'm using an Atmega micro, not the Arduino, but same thing. Just different language.  The bob is a massive chunk of cast iron on a brass sleeve for temp compensation, on a 3/8 invar rod. The suspension I built along the lines of the recommendations in Matthys book Accurate Clock Pendulums, but in hind sight it could be better. I have had it running, but other things have got in the way, and at the moment its a project still. 
Neil
New Zealand.
On 08/03/19 8:16 AM, Rob wrote:

Hi John,

 I am interested in this project. I was wondering how you are going to connect the bob onto the carbon fibre rod. I have been interested in cycling for many years and there it has always been a task to bond carbon an steel or aluminium together. I would be interesting to see how this project ends up

Regards

Rob Lockyer


John Haine
 

Rob, the CF is 10 mm OD tube, 8 mm bore.  There is a 25 mm slug of 8 mm aluminium rod araldited into the bore at each end.  At the top there's a 3 mm cross hole with a bolt through it that holds the rod into the lower chops of the suspension.  The bottom slug has an M4 rating screw tapped and loctited into it.  As the inside of the bore is quite rough, and I believe the CF is laid up with epoxy, I'm hoping that the lower joint will hold but may drill another cross-hole for a pin just in case.  


Rob
 

Thanks John,

Good old Araldite, sticks anything to anything else :-)

Regards

Rob

On Thu, 7 Mar 2019, 21:15 John Haine, <john.haine@...> wrote:
Rob, the CF is 10 mm OD tube, 8 mm bore.  There is a 25 mm slug of 8 mm aluminium rod araldited into the bore at each end.  At the top there's a 3 mm cross hole with a bolt through it that holds the rod into the lower chops of the suspension.  The bottom slug has an M4 rating screw tapped and loctited into it.  As the inside of the bore is quite rough, and I believe the CF is laid up with epoxy, I'm hoping that the lower joint will hold but may drill another cross-hole for a pin just in case.  


John Haine
 

That's what I'm hoping!  Time will tell.


John Haine
 

To avoid making the thread very long with photos etc I've created a folder over in "Files" called iSynchronome, and to get the ball rolling uploaded the initial schematic for the electronics which I now need to get on and build.


Darren Conway
 

Hi

It seems to me that you are using so little of the original clock you could have done it without a donor clock.

--

Dazz

Virus-free. www.avast.com


neil
 

Hi John,
                Looking at your clock pulser - a couple of suggestions:
 the diode D1 across the clock dial is back to front.
I would suggest 1k pullups in place of the 22k at R6 and R7, because the 22k will only give you about 0.2mA of switch current which is too low.
A protection diode across the LM7809 is a good idea and will protect it from reverse polarity if the input supply goes below the volatge on C3.

cheers
Neil
On 10/03/19 1:27 AM, John Haine wrote:

To avoid making the thread very long with photos etc I've created a folder over in "Files" called iSynchronome, and to get the ball rolling uploaded the initial schematic for the electronics which I now need to get on and build.


John Haine
 

Oops on the diode - thanks.

Actually the on-board Arduino pullup is 20k, so the effective pullup will be ~12k, giving a wetting current of ~0.5 mA.  We'll see how that gets on.


Dr Stuart Harrison
 

Dear John,
Thank you for your notes on your project. I think you will have difficulty lifting the gravity bar  reliably with what is effectively a 'Scotch yoke'.
I have tried to upload a file of my ongoing development of two Synchronome clocks but are having difficulty.
Kind regards,
Stuart Harrison


John Haine
 

Thanks for the observation Stuart.  Please could you explain why you think this, and also why the comparison to a Scotch yoke?  


wwwrogerj.codotuk@...
 

Hi John..Firstly I'm very envious of your mechanical, electronic and programming ability..I'm sure your 'Nome clone will turn out to be a work of art !
I have a couple of questions...First HJ develeoped the 'Nome over a number of years with enormous resources at his disposal. The result is that a 1936 'Nome is already an exceptionaly accurate  timepiece when set up correctly. I did see on this forum a suggestion that the "shock" of the resetting gravity arm worried someone but of course, at the time it was instrumental in HJ's plan to perfect the problem of sharrp, clean contact that did not rob the pendulum of energy..Such interferences to the pendulums motion that do exist are exactly repectitive, time wise.

In my own DIY  optically sensed and Hipp Toggle style, magnetically impulsedd clock I've found that the random external influences - temp and barometric pressure - are the final challenge and I've yet to proove that its timekeeping is any better than a standard 'Nome..Mine is regularly compared with Mike's 'Nome,( another forum member), using his Pi monitoring system. This is in spite of currently running a test with the temperature inside the case regulated to 22C..It is so frustrating to see the drift changing all the time in sync with Barometric pressure..Storm Gareth passing thru for instance and the much worse deep depression a few days ago..

So the first question. Any prediction (or even aspiration ) about how much better your ingenious design will be at timekeeping than a standard 'Nome ?
The second question is about the carbon fibre rod with near 0 temp coef. I have long held the opinion that the small coef of T of the Invar rod (+ the mild steel 4Ba threaded section)  is compensated in a 'Nome by the upward expansion of the bottom supported bob...What do you think and are you planning on a bob supported by a shoulder at the middle of its height ?
Roger


RICHARD ADAMEK
 

I didn't grasp the comparison either but from the steam world, it is not widely realised that the scotch yoke produces simple harmonic motion as opposed to the crank and con rod, where there is an uneven angularity effect which dogged designers of valvegears. Probably not relevant but might come in useful for somebody someday.

Richard in the wild and windy Boglands of Norfolk



On Wednesday, 13 March 2019, 13:28:34 GMT, John Haine <john.haine@...> wrote:


Thanks for the observation Stuart.  Please could you explain why you think this, and also why the comparison to a Scotch yoke?  


John Haine
 

Thanks for the comments Roger - my programming skills will be tested to the limit pretty soon!  I've seen your results on your website and as you say barometric pressure has a big effect.  Have you tried compensating it?  Douglas Bateman has made a clock rather similar to yours and fitted an aneroid capsule with good results - I think it's described in HJ relatively recently, and there was certainly an article by someone else in the last year or so. 

Alternatively, and something that I may investigate when/if I get this thing going, it turns out that there is a specific amplitude where to a first order the rate of any pendulum does not vary with atmospheric pressure.  If the pressure rises, the bob gets lighter and the clock slows down.  But also, the increased air drag reduces the amplitude which increases the rate - at the right amplitude these can be made to cancel out.  This is what John Harrison probably discovered, and it is one of the major reasons why "Clock B" at Greenwich is so good.  There's a guy in the US, Doug Drumheller, who has worked out the theory and suggested a way to make measurements to predict the required amplitude.  This is one reason why I am going to make the number of swings between impulses variable, so the average amplitude can be adjusted to see if I can get this working.  I'll check with Doug and see if it would be possible to post copies of his articles here.

I don't have high expectations that the timekeeping will be wonderful, but I'm hopeful!  Mainly I'd like it to be quieter.  Yes, my bob is supported on a "ledge" halfway up its bore, but there is a short length of threaded adjuster between the bottom of the rod and the nut.  I've got a couple of ideas about how to compensate the pendulum's actual variation, but want to get it going first!