Strange. I just downloaded it again with no issue. There have been no similar reports on the CarveWright forums.
Hi DickB
I too have played with the simple two transistor circuits and could not get them working.
I am very interested is learning more about your microcontroller based impulse circuit.
From your description I take it that you have a sense circuit to detect the pendulum over the main coil and based on the period of the swing you code will adjust the amount of power to the coil to increase or decrease the impulse to the pendulum.
Will / could you supply the circuit diagram and HEX files?
If you would like any help in the design or development of the circuit or the PCB let me know.
I love the mechanics of a large gear clock except for the fine tuning escapement
Gareth
I would be interested in seeing your project(s) - why not post?
The circuit itself is pretty simple. I use a single coil for sense and drive. There is a very simple filter on the coil with diode clamps to limit the voltage. This is fed into a TI MSP430 low-power microcontroller's comparator input. The micro has a watch crystal frequency reference and uses it to measure the pendulum period. As you stated, it adjusts the output pulse duration to speed up or slow down the pendulum a fraction. An output port drives a transistor to put current back into the coil. And there is a 3.3v regulator for the micro. I also added a bi-color LED to help set up the pendulum. It flashes red if the pendulum is too slow, and green if too fast. You adjust the pendulum bob accordingly. Once the pendulum is adjusted within the window of close enough for the electronics to manage, the LED stays off.
I have gotten circuit boards laid out and built by ExpressPCB.
I've built a total of three clocks now, and they run very well. I have been able to adjust the pendulums in less than 10 minutes using the LED - no waiting for hours or days to see if they are keeping correct time. They keep accurate time after that. I was shooting for a year on 4 D cells, but didn't make it. I calculate 3-4 months. The first clock has been running for a bout 2 months now. We'll see.
Some videos:
Mystery Clock - YouTube
Mystery Clock & LaunchPad - YouTube
I just recently finalized the design and made the clock available as a project in the CarveWright PatternDepot. I'm selling a hardware kit of parts, including electronics, for those building the CarveWright project. I'm not putting the electronics and code in the public domain at this time. Contact me through my web site carveshop.com or PM if you would be interested in purchasing the electronics.
Thanks for the info. I know more about batteries than wood gears, so I will comment on that aspect. If you want your clock to run longer, use 4 AA energizer ultimate lithiums. They will go MUCH longer in that application.
I am not sure if the slightly higher voltage is a factor in your clock though, so you might want to double check that aspect.
Thanks for the tip on batteries. But the AA Energizers you mentioned are specified at 3000-4500 mAh (http://data.energizer.com/PDFs/l91.pdf) while the D alkaline batteries that I use are in the 12000-18000 mAh range, so they will last considerably longer. The lithium batteries have an advantage over the same size alkaline batteries in high-current draw devices, but at low discharge rates (under 25 mA) they have little advantage. My clocks are low discharge rate devices.
My first clock has been running for over four months on the first set of D cells.
They are also Energizer: http://data.energizer.com/PDFs/e95.pdf