Hi everyone.

My name is Leandro and I’m a software developer from Brazil. I want to build a personal electronics project but not sure how to start. I have some knowledge about electronics but not at the level required for this project, so I’m asking for your help here. I would gladly appreciate any help and insights, even if it is just for part of the problem.

The project: the end goal is to attach an hours and minutes display (hh24:mm) to a mechanical watch that I already have. However, one requirement is that everything should be powered by mechanical energy, extracted from the mechanical watch barrel (or an additional barrel). So, a rechargeable battery and/or capacitors can be present, but only to store the temporary energy to refresh the display and power a small circuit.

Another complication is that the mechanical watch must “instruct” the digital part regarding how much time has passed (and not the other way around). So, I thought about having a mechanical component that can rotate, connected to the appropriate mechanical watch gear. Every complete turn of this component would “instruct” the circuit that one minute has passed and then update the display accordingly, using the previously stored energy (somehow). Some state should be stored in the circuit to know how to add and display this new minute that passed, so I believe a few bytes of store will do.

For the display, since the energy requirements must be low, I thought about using a small e-paper display (the smallest that I will be able to buy online), since it will only use energy during screen refresh (one refresh each minute during normal use). I know the lifetime of e-papers are not ideal for constant refreshes, but the display “burn out” or degradation after some millions of cycles is something acceptable.

At this point I don’t have a requirement for size, because bigger watches can have bigger barrels to store energy, so I believe the solution could scale up or down appropriately.

In summary:

• An e-paper display (SPI interface?)
• A “component” that will be rotated every minute to instruct the circuit to update. This component will be connected by gear(s) to the appropriate watch minute gear
• Another component / subcircuit that will convert mechanical energy to electric energy and store it for the next update. This will be connected to the mechanical watch barrel or a separate additional barrel
• A button to reset the circuit to 00:00
• The circuit itself

Again, any help/insight on any part of the project would be much appreciated.

Thanks in advance.

Leandro.

 

I read the description of your proposed project. I think that it will be difficult (but maybe not impossible) to get enough electrical energy from the mechanism to power the electronics. The electronics will be quite simple if a small microcontroller is used to drive the display. It could be put into aleep mode between updates. It would be probably necessary to breadboard this part of the project to find out what its real power requirements are. Once that is known, it will be time to consider what is available to generate the power to drive it.

 

Your problem is going to be energy, energy, energy. A electrical energy budget of 5 microamps at 1.5v should be sufficient.

The conversion of mechanical energy is going to be tiny and inefficient from standard watch movements.
https://phys.org/news/2015-11-self-powered-e-watch-powered-wrist-movements.html

 

That is a very interesting a doable project. I started a 10,000 year solar powered clock but never had the time to complete it. One of these days I hope to get back to it.



You can follow the project here.
https://forum.allaboutcircuits.com/ubs/long-clock-prototype.647/

One of the goals was lowest power consumption. For this I used this LCD display.
https://futurlec.com/LCDDisp.shtml

and ICL7211A display driver IC.

The total average current draw for the whole project is about 3μA.

 

Your problem is going to be energy, energy, energy.

The conversion of mechanical energy is going to be tiny and inefficient from standard watch movements.
https://phys.org/news/2015-11-self-powered-e-watch-powered-wrist-movements.html

Hi!

Thanks for replying.

Totally agree. Energy is the main challenge for this project. However, since I'm planning to use a very small e-paper display (relative to the watch/barrel) and its energy consumption is very low, I still believe its feasible.

I'm not trying to build a smart watch with many functions and connectivity. Just a hh24:mi display that needs power only during refresh of the e-paper (once each minute).

As far as I researched, the energy required for a full switch on an e-ink display is about 7 to 8mJ/cm2. Some (good) mainsprings in a wristwatch can store 1300 millijoules of energy. Lets assume for now a 1cm2 display. So 1300 divided by 8 would result in 162 total possible refreshes or 2.7 hours of uptime/power reserve. Still far from ideal, but for a test project, it could work (solar could be added in the future, better display, parcial updates, better mainspring/barrel and so on).

Still, I don't have enough knowledge to estimate additional energy requirements for the circuit nor even how to transform and store the energy.

Regards.

 

I read the description of your proposed project. I think that it will be difficult (but maybe not impossible) to get enough electrical energy from the mechanism to power the electronics. The electronics will be quite simple if a small microcontroller is used to drive the display. It could be put into aleep mode between updates. It would be probably necessary to breadboard this part of the project to find out what its real power requirements are. Once that is known, it will be time to consider what is available to generate the power to drive it.

Hi.

Thanks for taking time to reply.

Agreed! I plan to build it using a breadboard in any way it would work and then iteratively refine it. Using a small microcontroller is my preferred path as well. Ideally, since I have good software knowledge, even use an electronic circuits software emulator to start. The problem is: it will be my first time using either!

To your point: also think that the first part would be to "drive" the display in any way possible with any power source, make the refreshes and measure how much energy is being consumed. I'll start with that in mind. Any tips on how to actually using something to measure the energy over time? Does a simple multimeter is capable of that?

Regards.

 

Hi!

Thanks for replying.

Totally agree. Energy is the main challenge for this project. However, since I'm planning to use a very small e-paper display (relative to the watch/barrel) and its energy consumption is very low, I still believe its feasible.

I'm not trying to build a smart watch with many functions and connectivity. Just a hh24:mi display that needs power only during refresh of the e-paper (once each minute).

As far as I researched, the energy required for a full switch on an e-ink display is about 7 to 8mJ/cm2. Some (good) mainsprings in a wristwatch can store 1300 millijoules of energy. Lets assume for now a 1cm2 display. So 1300 divided by 8 would result in 162 total possible refreshes or 2.7 hours of uptime/power reserve. Still far from ideal, but for a test project, it could work (solar could be added in the future, better display, parcial updates, better mainspring/barrel and so on).

Still, I don't have enough knowledge to estimate additional energy requirements for the circuit nor even how to transform and store the energy.

Regards.

What you need to design is a: https://museum.seiko.co.jp/en/knowledge/Quartz06/

You don't need the time keeping functions, you need the equivalent electrical generator.

 

What you need to design is a: https://museum.seiko.co.jp/en/knowledge/Quartz06/

You don't need the time keeping functions, you need the equivalent electrical generator.

Yep, the Grand Seiko Spring Drive is a huge inspiration for this project. Its like I want to create a Spring Drive v2, but now with a display!

Regards.

 

That is a very interesting a doable project. I started a 10,000 year solar powered clock but never had the time to complete it. One of these days I hope to get back to it.

View attachment 292892

You can follow the project here.
https://forum.allaboutcircuits.com/ubs/long-clock-prototype.647/

One of the goals was lowest power consumption. For this I used this LCD display.
https://futurlec.com/LCDDisp.shtml

and ICL7211A display driver IC.

The total average current draw for the whole project is about 3μA.

Wow. Amazing. I'll have a look and reply with comments soon.

Thank you very much!

 

Yep, the Grand Seiko Spring Drive is a huge inspiration for this project. Its like I want to create a Spring Drive v2, but now with a display!

Regards.

The display is the easy part. The electro-mechanical generator design of the Spring Drive is fantastic. The regulator electromagnet, rotor and electrical power system are micro-miniature marvels of how they convert the spring PE to electrical energy KE.
This conversion is the key to success.
Good luck.

 

The display is the easy part. The electro-mechanical generator design of the Spring Drive is fantastic. The regulator electromagnet, rotor and electrical power system are micro-miniature marvels of how they convert the spring PE to electrical energy KE.
This conversion is the key to success.
Good luck.
View attachment 292893

For now I'll be happy to only prove the concept. I'll leave possible miniaturization and mass production for the experts

 

For now I'll be happy to only prove the concept. I'll leave possible miniaturization and mass production for the experts

There's not much to prove. We know it's possible to generate the required power from the main spring, the devil is in the details of how to duplicate it.

 

There are energy harvesting ICs such as LTC3105 and LTC3108.

 

There are energy harvesting ICs such as LTC3105 and LTC3108.

Amazing, thanks for sharing!

So, if I understood correctly, a single LTC3108 IC will work as a capacitor + "charge control", in a sense that it will take inputs as low as 20mV, accumulate according to a certain charge time (like 10 seconds for 50mV) and then output/discharge a certain voltage and current (3.3V for example, configurable).

So I would plug some energy source at one end and "harvest" a fixed charge and current at the other end after a certain number of seconds. And I see that there is a pin (PGD) to indicate that its ready to "discharge". And pins (VSTORE) for an extra large capacitor!

Its nice to discover that those things even exist and are very affordable even from a mass production perspective. I'm only somewhat used to Arduino / Raspberry PI that are like "brute force" software based approach for prototyping. But for this project, since the energy requirements, I'll try to stick to very basic components like those you shared.

Sorry for the basic terms, I'm trying to use pure logic since I don't have any jargon in electronics (yet).

Thanks again!

 

You are slowly getting there.

Firstly, you want a clock circuit that can run on the lowest amount of current. I have demonstrated one that takes as little as 3μA.
Now you need to do the power and energy analysis.
3μA x 3.3V = 10μW

Find a power source that can supply 20μW to take into account efficiency losses.
If it can generate 1mA @ 20mV then you're getting somewhere.

Next you have to consider energy requirements.
If the energy source is intermittent, then you have to work out the average energy delivered and match that with the average energy required by your clock.

 

You are slowly getting there.

Firstly, you want a clock circuit that can run on the lowest amount of current. I have demonstrated one that takes as little as 3μA.
Now you need to do the power and energy analysis.
3μA x 3.3V = 10μW

Find a power source that can supply 20μW to take into account efficiency losses.
If it can generate 1mA @ 20mV then you're getting somewhere.

Next you have to consider energy requirements.
If the energy source is intermittent, then you have to work out the average energy delivered and match that with the average energy required by your clock.

Thanks!

Just to organize the thoughts before I start drafting and buying components:

1- Spring barrel (for now) would provide mechanical energy. A single good mainspring can store around 1300 millijoules, but multiple barrels could be used in future iterations;
2- The mechanical energy will be converted to electric energy using some kind of micro generator/dynamo. Still to research and understand if there is a commercially available component, that hopefully can produce the minimum input for a LTC3108 IC;
3- Energy output from the generator will input to a LTC3108 (or a more appropriate harvester if it is the case);
4- LTC3108 will output to a circuit that will consume the energy appropriately. For now, everything points to a "burstable" circuit with little to none "standby" power needs. This circuit should also be able to stop the movement from the spring barrel somehow when no more energy is needed (this could be tricky);
5- Another hypothetical "rotational" component will be connected to the circuit and signal it for every 360 turn (or something configurable). This component will also be connected by gears to the appropriate mechanical watch gear that indicates a full minute has passed;
6- E-papper / e-ink display connected to the circuit, the smallest commercially available, black and white only;
7- Small flash memory connected to the circuit to store the current time (hh24:mi) or total minutes after midnight. Two bytes should by enough, so the flash memory could be overkill, another simpler storage (to research) should sufice;
8- Reset button connected to the circuit;
9- Circuit logic (programmed and stored in another IC?) to add 1 to the total number of minutes stored in the memory, every time the component #5 signals and partially refresh the display to reflect the time, to consume the least amount of energy possible. Also logic to reset when #8 is pressed.

Something like this to start.

Thanks again for taking the time to provide advice to a somewhat crazy project.

Regards.

 

Well, you seem you have everything worked out except the one most critical piece (the generator.) I would have looked into that first to see if it is even feasible.

 

Well, you seem you have everything worked out except the one most critical piece (the generator.) I would have looked into that first to see if it is even feasible.

True...

But at least now I have some clarity about the project, including possible impediments/impossibilities, after the community contributions.

For the motor / generator, I'll try something like this: https://loja.fabricadebolso.com.br/produto/micro-motor-super-alta-velocidade-53000-rpm.html

In any case, as far as I understand, a generator could be a couple of fixed magnets rotating near a copper coil. So it would be feasible to build it by hand. However I rather use commercially available parts because I'm not looking for miniaturization for a V1, only to prove the concept.

Thanks for sharing!

 

I don’t think you are going to be able to spin that fast enough to generate useful power.

 

I don’t think you are going to be able to spin that fast enough to generate useful power.

Yeah, I'll have to test it. An alternative is using a motor with reduction gear like this one, as long as I have torque: https://www.robocore.net/motor-motoredutor/micro-motor-com-caixa-de-reducao-6v-1000rpm

So it might be better to add another component to the project, a gear set to be able to fine tune torque / speed.

Regards.