Smartwatch 10 Years Ultra-Long Battery Life

This is the last iteration of the battery. Using spot welding and some creative work, the battery pack is done. Picture attached. This is a 1 Ah battery, 30mm x 6mm , which at 10uAh including self-discharge will last about 10 years (10% self-discharge). The battery is non-rechargeable. Future iteration will add a rechargeable battery with solar harvester which will bump up the life indefinitely as the watch needs to be recharged only once every few years and the typical deep-discharge-charge cycle is about 100 times.

I still need to solve the problem of battery connector and housing, as currently the battery is relying on hand-made kapton tape insulation and hot-solder leads.

Working on the battery pack.

The previous design was suboptimal. The new one is getting better. Using spot-welding and some creative kapton tape work, this is the third variant (first one was using built-in tabs, second was cold solder silver epoxy).

Stay tuned for more updates. I will try to make a better pack, this one was just a study.

Excellent news. Code works with NRF52. Hardware works with the old R1 version of the PCB. Not easy but no surprises. The real challenge was to integrate with the display to see debugging messages. I know I make it sound easy but I'm really wondering.

Ready for integration with the fabbed PCB once I get it, somewhere next week.

Pictures of the setup attached at the end.

Great progress today on accelerometer integration.

I'm using the development board from STM, as well as a Teensy as a dev test bed.

Got the single taps, double taps, and axis acceleration, as well as interrupt/wakeup functionality figured out.

Next is to move to NRF52 and test in the watch.

Cheers!

Picture is of the test setup.

For those of you interested in project management (yeah I know, project + management are dirty words for many), we are employing an evolutionary waterfall development model. That said, project management is absolutely crucial in any setting. The alternative is an abandoned project or relegation to the vast graveyard of great but half-baked projects littering the DIY landscape.

Evolutionary model combines the Iterative and Incremental model of software+hardware development life cycles. Delivering the product is an incremental process over time. It is better for software+hardware products that have their feature sets redefined over the development timeline because of user feedback and other factors such as manufacturability, component availability and/or budget constraints. The Evolutionary Waterfall development model divides the development cycle into smaller, incremental waterfall models in which users are able to get access to the product at the end of each cycle. Feedback is provided by the users on the product for the planning stage of the next cycle and the development team responds, often by changing the product, plan or process. Therefore, the product evolves with time. All conventional project management and development models have the disadvantage that the duration of time from start of the project to the delivery time of a solution is very high. Evolutionary Waterfalls solve this problem!

Evolutionary Waterfall models suggest breaking down of work into smaller chunks, prioritizing them and then delivering those chunks to the customer one by one. The number of chunks is huge and is the number of deliveries made to the customer. The main advantage is that the customer’s confidence increases as he constantly gets quantifiable goods or services from the beginning of the project to verify and validate his requirements. Thus, the model allows for changing requirements as well as all work is broken down into maintainable work chunks.

Another advantage of the Evolutionary Waterfall is that at any point in time you have a working product! Yes, it might be a crappy one, and some features are missing, but you do have a working ten-year-on-a-charge smartwatch on your wrist that proves that the original premise is sound and the end product will work!

Here the situation is complicated by the introduction of hardware dimensions, so combining hardware with software in a development is a nice little challenge.

Cheers!

The cold-solder battery design was a failure, in sense that making cold-solder joints is really messy, takes 24 hours for the cold-solder to join mechanically and the joints themselves leak and require extreme dexterity. On top of that the cold-solder is expensive and very hard to work with.

The solution would be to make spot-welded designs from coin-cell batteries. I have ordered a simple and really cheap spot-welder (it's just a battery with two copper leads and a trigger) and a roll of nickel strips. The upside of the spot-welded design is that I can also design very quickly any arbitrary battery configuration, and the profile is only 100 microns thick. The cold-solder required extra thickness for the silver epoxy, so that's that.

This will also allow me to start work on the rechargeable battery pack with a solar harvester. That's a separate mini-project, so I keep adding to the list, mais c'est la vie.

Stay tuned for the battery pack!

Cheers.

Alpha Release 3 is done, will take some time to get it but looks promising.

=========

Big thanks to PCBWay for sponsoring this release! I will post a full review and feedback later once I get the whole thing assembled and tested. PCBWay went out of their way to accommodate the really demanding SMT and budget and we had to work really hard to "meet in the middle". That tiny PCB which is the size of a US quarter dollar cost $300, so that's not a small effort.

=========

I know some people ask about time. This type of development takes a lot of time, more-so that I'm working alone, and in my spare time off-work. I have a full-time job. As I mentioned, usually this takes a large team of about 10 to 20 people and about a year to develop (say, 10-man-years, 20-m-y on the upper edge of the envelope) in a high-pressure cooker environment. This isn't a typical home-brew DIY project you complete with off-shelf components and a soldering iron. I'm trying to achieve this at a tiny fraction of the time and budget. Realistically, I'll be lucky if I get the Beta release by March '24.

Cheers!

I spun off the nano-amp-meter as a separate project.

Enjoy. Cheers!

https://hackaday.io/project/193886-nano-micro-milli-amp-current-meter

The nano-amp-meter is here, pictures of unboxing and device. I tested, and referenced against a commercial nano-amp meter, it works, don't have time now to go deeper but it does work when connected to an oscilloscope for a good and accessible transient analysis. Pics/videos later in a split project.

I'll split this later as a separate project, I got way too much going on now. Also I need to clean up the design and add silk-screen, and make it more user friendly. Right now you need to really know what's going on to use it, might not be very useful for most people.

Cheers!

Project log.

List of do-to items:

1. 1Ah Battery assembly bookend design complete with cold-solder, and wires.

2. Magnetometer code confirmation / test / POC

3. Accelerometer code confirmation / test / POC

4. Case re-design / re-manufacture

5. Order reed contact switch for R3 assembly

6. Nano-amp-meter test/integration

7. Alarm SW finalize

8. Solar micro-charge/Panasonic rechargeable 2020 combo POC

9. Buzzer/steel case integration/test feasibility (sealed case problem) for alarms.

The way this is going, combined with my free available time, this is targeting January 2024 for the Alpha release. This includes R3 and R5 manufacturing and test and integration.

Cheers.