Clock Project

[[email protected]]

Would like to build a digital clock with a long-duration mechanical
escapement ("artistic" part).

I found Bill Bowen's site

http://ourworld.compuserve.com/homepages/Bill_Bowden/

and could probably decipher something here, but thought I'd try for
some advice/ shortcut.

I envision empirically determining the periodicity of my "escapement
mechanism" and inputting that time period into my clock via dip
switches. This number (not much resolution needed due to mechanical
vagaries of long-period escapement) would then be latched into the
clock electronics to provide the clock increment.

One could call this a "variable increment clock", I guess.

Could someone point me to a circuit that comes close to this?

Marc on the North Coast

 

[Anthony Fremont]

Would like to build a digital clock with a long-duration mechanical
escapement ("artistic" part).

I found Bill Bowen's site

http://ourworld.compuserve.com/homepages/Bill_Bowden/

and could probably decipher something here, but thought I'd try for
some advice/ shortcut.

I envision empirically determining the periodicity of my "escapement
mechanism" and inputting that time period into my clock via dip
switches.

I had to read this a couple of times before I got it (I think). If I
understand you, you want to use a pendulum of arbitrary length (and
arbitrary period) and then have this advance the clock mechanism the correct
amount for each swing. For example, say you used a 1 meter long pendulum
with a period of 2 seconds. You would want to advance the the clock 2
seconds each cycle. Correct?

This number (not much resolution needed due to mechanical
vagaries of long-period escapement) would then be latched into the
clock electronics to provide the clock increment.

Not sure what you mean here, there is nothing vague about the motion of a
pendulum.

One could call this a "variable increment clock", I guess.

Could someone point me to a circuit that comes close to this?

Bryan Mumford has something like this going, but he uses a pendulum to
generate the pulses to an electrical drive system for the hands. He keeps
the pendulum going with magnetic impulses. The impulses are triggered by
the pendulum itself, so that the pendulum is the thing actually keeping
time.
http://www.bmumford.com/clocks/em2/index.html

 

[North Coast Igor]

I had to read this a couple of times before I got it (I think). If I
understand you, you want to use a pendulum of arbitrary length (and
arbitrary period) and then have this advance the clock mechanism the correct
amount for each swing. For example, say you used a 1 meter long pendulum
with a period of 2 seconds. You would want to advance the the clock 2
seconds each cycle. Correct?


Not sure what you mean here, there is nothing vague about the motion of a
pendulum.


Bryan Mumford has something like this going, but he uses a pendulum to
generate the pulses to an electrical drive system for the hands. He keeps
the pendulum going with magnetic impulses. The impulses are triggered by
the pendulum itself, so that the pendulum is the thing actually keeping
time.
http://www.bmumford.com/clocks/em2/index.html

 

[North Coast Igor]

Thanks for the reply, Anthony.

You understood okay, but I was thinking of a complex escapement
mechanism that is amusing (?) to watch and has a period of many
seconds. Accuracy is not a goal here; the goal is more: "you can keep
time with THAT ?"

The latched increment input into a clock driver is the part for which
I'm seeking help.

Thanks again.

Marc

 

[ehsjr]

Thanks for the reply, Anthony.

You understood okay, but I was thinking of a complex escapement
mechanism that is amusing (?) to watch and has a period of many
seconds. Accuracy is not a goal here; the goal is more: "you can keep
time with THAT ?"

The latched increment input into a clock driver is the part for which
I'm seeking help.

Thanks again.

Marc

I'm lost on your description, too.
Do you want to _add_ the incremental value
to the previously stored value each interval?

Ed

 

[North Coast Igor]

Right, Ed. Time value is incremented by the escapement period time
value. (This is true of every mechanical clock, when you think about
it.)

I didn't provide detail on the "escapement mechanism" b/c it's not
important regarding the portion I need help with. The Increment will
be many seconds and you understood perfectly. The "ecapement
mechanism" operates over a repeatable-enough period of time and THAT
time period will be the "increment".

My own thought since I described the problem is that instead of
counters, gates and DIP switches, a small microprocessor chip would be
the easy way to program the increment function (especially regarding
re-programmability). --probably cheaper, too.

Thanks for responding.

North Coast Igor - Marc

 

[ehsjr]

Right, Ed. Time value is incremented by the escapement period time
value. (This is true of every mechanical clock, when you think about
it.)

I didn't provide detail on the "escapement mechanism" b/c it's not
important regarding the portion I need help with. The Increment will
be many seconds and you understood perfectly. The "ecapement
mechanism" operates over a repeatable-enough period of time and THAT
time period will be the "increment".

My own thought since I described the problem is that instead of
counters, gates and DIP switches, a small microprocessor chip would be
the easy way to program the increment function (especially regarding
re-programmability). --probably cheaper, too.

Thanks for responding.

North Coast Igor - Marc

So conceptually:
impulse from escapment-> add N to totalizer
is all we have so far. First, flesh out the
impulse generator: microswitch, opto, what?
Debounce it as needed. You can google "debounce circuit"

Apparently you've already figured out how to
come up with N - but why on earth do it that way
when you've already decided to use a micro?
It can count the time between impulses probably
better than you, and with better resolution and
no need to gate the value stored in a bunch of
dip switches. Just count directly using the
micro. That also takes into account any physical
abberation in the escapement.

But after impulse from escapment-> add N to totalizer
what happens next? Does the value in the totalizer
cause something to happen? Seems to me it should,
and logically it is something like "is value in
totalizer >= X" if yes do something, if no do
nothing. Do you want the value to cause the hands to
move some particular distance?

You need to do a complete set of specs for what you
want to happen. It _has_ to be more than just
adding a number to a total or else there is no
purpose to doing the addition. I don't mean this
as criticsm, but I don't think anyone here wants
to drag the specs out of you one at a time. I think
that's what I'm doing, and it's not fair of me to
waste your time because I just plain don't get it.

Ed

 

[Rich Grise]

Right, Ed. Time value is incremented by the escapement period time value.
(This is true of every mechanical clock, when you think about it.)

I didn't provide detail on the "escapement mechanism" b/c it's not
important regarding the portion I need help with. The Increment will be
many seconds and you understood perfectly. The "ecapement mechanism"
operates over a repeatable-enough period of time and THAT time period will
be the "increment".

My own thought since I described the problem is that instead of counters,
gates and DIP switches, a small microprocessor chip would be the easy way
to program the increment function (especially regarding
re-programmability). --probably cheaper, too.

First, you're supposed to bottom-post, to follow the natural flow of the
thread.

Given that, google "compound pendulum" - you'll find some really
impressive-looking designs for pendulums (pendula?) of practically
unlimited period, probably even adjustable ones.

Good Luck!
Rich