Vetinari's Clock

[rockets4kids 204]

I found it:

 

 

That's a nice clock for $3!

[grahamf72 169]

Oops, sorry about the bad link. Thanks for fixing it Rockets4kids. My excuse is I was in a hurry lol.

Yes it is a quite nice clock for $3 - I'm almost tempted to buy some more, make the "adjustment", and give them away as Christmas presents.

[bluehash 1,581]

That is irritating!

[ILAMtitan 86]

@@grahamf72, That is awesome!

 

I think the hardest part of the project was finding a clock with a loud enough ticking noise.  Most of them are actually advertised as being quiet.  The cheaper you go on the clock, the louder it will probably be.

 

I had the same backwards problem crop up a few times which led to the adjustable pulse width; even at 0x4000 it still happens sometimes on startup, but that might just be an initial state issue.  I found that a short burst of manual force to the clock mechanism would restore proper drive direction.

[GeekDoc 226]

I found that a short burst of manual force to the clock mechanism would restore proper drive direction.

 

Do you mean hitting it?  You mean hitting it, admit it.  ;-)

[bluehash 1,581]

Do you mean hitting it?  You mean hitting it, admit it.  ;-)

^hehe

[ILAMtitan 86]

Do you mean hitting it?  You mean hitting it, admit it.  ;-)

Ya, ok, I mean hitting it.  Rarely is the case in electronics development that hitting things actually has a difference, so I try to jazz it up a bit when it happens.

[Fred 453]

There's nothing wrong with a bit of percussive maintenance when it's required.

[JWoodrell 285]

Kinetic recalibration?

[narangsagar31 0]

which transistor is used in the vetinari clock?

[rockets4kids 204]

The current, voltage, and switching speeds are so low that you can use practically any transistor in this project.  The only requirements are that you have two NPN and two PNP transistors.

 

You can even get away without the transistors, driving the coil directly from the MCU pins.  I wouldn't recommend this on the MSP430, but people have done it.  With the small, cheap mechanism I am using, the current required to drive the mechanism does not exceed the absolute maximum values.  Nonetheless, I am using an H-bridge just for safety.  For MCU's with greater drive current, there is absolutely no need for external drive circuitry, so if you have any PICs or AVRs laying around, it would be best to just use those.

 

One thing I have wondered is whether you could use multiple GPIO pins to distribute the load.  Using two GPIOs on each side (a total of four pins) should bring things within recommended values.  Does anyone know of the issues involved in doing this?

[t0mpr1c3 91]

I hate it and love it with equal measure. I'd like to make one and give it away

 

Great project guys!

[none 7]

Since the clockwork I am working with is a continuous type ("sweep movement"), it needs an 8 Hz waveform to move 1 s, and I'm currently trying to figure out how to actuate it properly with the irregular timing.

Also I noticed that the stator seems to be slightly magnetized (to determine the rotation), as the same timing for both polarities causes different motion.

I was looking at different "Vetinari" clock projects and found another, minimalistic approach using an MSP430G2211: http://renaud.schleck.free.fr/horloge_vetinari.php

He does not use a sequencer but randomly varies the length of the second-ticks by multiples of 1/64 s, and keeps track of the overall error.

 

I think this thread's project's code could also be ported to a smaller controller if one would replace the large 128 value array with an 8 int array and bit operations / pointer arithmetic.

I would also think there must be a more efficient way to randomly set 32 of 128 bits, but I'm not a computer scientist..

 

I'll report back once I've got the motor under control - thanks for the inspiration!

[Fred 453]

I'm not sure a "sweep movement" clock will work for this. Even if you randomise the timing and the second hand moves at varying speed, you won't get the desired slightly unnerving irregular tick.

 

Sent from my Nexus 5 using Tapatalk

 

 

[none 7]

If the motor can be reliably controlled at different speeds, probably even the regular tick motion could be created (although more silent), and thus also the irregular tick.

But currently, stopping and starting the motor will cause step-loss which of course is not an option. If this can't be overcome, i.e. the motor needs to keep running as fly-wheel, indeed only an irregular sweep is feasable, but I think this still may be interesting to notice.