as i was the one designing PCBs for them, i heard some of the feedback. basically:
version with OpAmps was working fine but calibrating correct pendulum length was time consuming.
version with MCU was much simpler to setup but it was sort of finicky with respect to choice of coil and supply voltage.
not sure about long term tests though i would expect MCU version to be more stable due to crystal oscillator.

i would also like to point out to an issue with the coil driver circuit. in original circuit as shown in post #17 has Q1 and Q2 connected directly - without any current limiting. if you go that route, i would suggest making some changes. at the very least add a resistor between the transistors:

 

You may be interested in the circuit described by BigClive here: WildWood Pendulum

 

In case it's of interest, I used an opto interrupter at the bottom of the pendulum with a 5mm wide card to trigger a 555 monostable. It triggered the coil in both directions but the field strength of the coil when the pendulum is swinging away is weak enough not to matter. A pot to adjust the duration of the pulse to control the amplitude.

Then I used a dual opto interrupter and an ATtiny85 processor to sense the direction so it only fires the coil in one direction. The system can start a stationary pendulum by activating the coil for 20% of the known pendulum period when the opto interrupters are not being triggered, (so it starts to resonate) and once the amplitude is enough to trigger the opto interrupters the amplitude can be controlled with a feedback loop based on the proportion of the time one opto interrupter is/is not triggered.

 

thank you all for your responses!

 

i would also like to point out to an issue with the coil driver circuit. in original circuit as shown in post #17 has Q1 and Q2 connected directly - without any current limiting. if you go that route, i would suggest making some changes. at the very least add a resistor between the transistors:

AND - add a resistor from the Q1 base to its emitter to assure a complete and rapid turn-off.

ak

 

For the record, I am a retired electrician with little electronics expertise. Yes, I understand ohms law and other basic formulas, but that's it. That said, I would welcome any pdf circuits that are simple to build and function without needing an electronics engineering degree. Thank you all again for your responses!

 

did you check AAC resources like:
https://www.allaboutcircuits.com/textbook/experiments/chpt-1/electronics-as-science/

 

Hello,

This pocket reference book may also be handy:
https://archive.org/details/analog_engineers_pocket_reference

Bertus

 

Omg you are too much . That is fantastic. But let me first get this guy's board that I paid $50 for and see if it will even work. I might have to redesign the clock base to be sure I have space for the PCB and battery. I'm using 1 18650 lithium ion battery work a rechargeable board. So the clock will run by USB-C 5v little charger as well as recharge the battery at the same time. Or you can pull out the cord and the clock will run off the 18650 battery for about 6 to 8 months before charging.

 

I would like to hear from @kicksnj about his experience with these boards and information about the rechargable battery and board included in his clocks.

 

as i was the one designing PCBs for them, i heard some of the feedback. basically:
version with OpAmps was working fine but calibrating correct pendulum length was time consuming.
version with MCU was much simpler to setup but it was sort of finicky with respect to choice of coil and supply voltage.
not sure about long term tests though i would expect MCU version to be more stable due to crystal oscillator.

i would also like to point out to an issue with the coil driver circuit. in original circuit as shown in post #17 has Q1 and Q2 connected directly - without any current limiting. if you go that route, i would suggest making some changes. at the very least add a resistor between the transistors:

View attachment 353670

Did you create similar JLCPCB compatible files for replicating the Regulated Pendulum Drive (version with MCU)? If so would you be so kind as to share them? Thank you!

 

i made few variants... such as this. THT parts are not needed (X2, Q1, U2). i can lookup and share the PCB files but since MCU is part of it, PCB design will be of little use - you will need the software too and tools to load it into MCU.

 

I have a copy of the old software from Dick Bipes (see attached) and will use AI to help update it so I can load it on the MCU (MSP430G2233 chip) using LaunchPad from Texas Instument. The software was originally created using Texas Instruments Code Composer Studio Version 5.1 and a Grace configuration tool. The Grace tool is obsolete, but the generated source code is still usable. AI can modernize the code and remove all Grace dependencies, convert the code to current CCS, and make it compile directly for a MSP430G2233.

If you can help me with the PCB design I would like to incorporate a MSP430G2233 chip using a DIP IC Chip Socket Adaptor 2.54mm Pitch 7.6 Row Pitch 2 Row 20 Round Pins Soldering. Thank you!

 

Why not make DIY and accurate with a simple crystal oscillator and CD4060? The 32,768 Hz crystal and the CD4060 cost a couple of bucks. With accuracy of crystal.

 

The Dick Bipes firmware and schematic use a 32.768 kHz watch crystal connected to the MSP430's XIN/XOUT pins.

 

 

The Dick Bipes firmware and schematic use a 32.768 kHz watch crystal connected to the MSP430's XIN/XOUT pins.

Looks good and simple. Meets the expectations.

 

That looks costly. With this one, you shouldnt go above $1 for the circuit. The Pendulum cost is there anyway

 

that will produce pulses... i am not clock maker but unless i am mistaken, that is the smaller part of the job. and many existing pendulum driver designs are also open loop.

my understanding is that the whole point of using MCU in a project like this is to use feedback and help quickly tune pendulums natural frequency to be close enough to reliably and automatically corrects sync with the pulse train - and stay synced even as temperature etc change. people making clocks all the time (makers community) rather than of one-off expect something that can help them quickly tune different designs or those where material properties are not very uniform (like wood, many plastics etc) which change properties with temperature, or humidity.

gentlemen for which i did PCB design was reporting that he has spent months fighting with it. everything would look spot on for a week of so and then it would suddenly drift...

WWII era was all about cheap mass production - all parts were the same (within tolerances). mechanically tuning pendulum for one and same design can be done once and as long as tolerances are close enough, open loop driver will kick it nicely (regardless if butterfly or sledge hammer).

but times have changed, and today most wand customized products, different shape, size, material etc. and calculating or tuning each and every unique product can be time consuming.

 

Yes, you are correct.