Bangle.js 3: GPS?

[gfwilliams]

As I mentioned in https://github.com/orgs/espruino/discussions/7341#discussioncomment-13959214 I'm looking a little more seriously at a Bangle.js 3 now (but we're still 18 months away at least!).

One big question is: is GPS worth it?

If I add GPS I want it to work well, so I'd probably use a Ublox chip. Battery life with GPS should be around 5x better by the look of it, and the fix should be more accurate.

It will add to the price (I'm still trying to get a better idea how much), but the new watch will be more expensive anyway as I'm having to design it from scratch, and will be going for higher quality materials and more memory.

The GPS will also likely push the physical size of the watch up too, but again because I'm hoping to make this one so that it's not glued shut (and the screen is bigger) it will be a larger watch than Bangle.js 2 anyway (although I hope it can be a bit thinner).

Any thoughts? How many of you use GPS? Personally I do use the GPS for walks/runs, and recently even used it for working out lap times on a track day, so I think I'd miss it (especially in 'recorder'). But if I'm being honest, running is the only time I couldn't have just used by phone's GPS as I don't take it with me.

[yerpj]

I own both Bangle.js 1 and 2, and to be honest, I've never used the GPS. I've always preferred to use a smartphone for anything related to geolocation. If Bangle.js 3 comes out without GPS, that won't stop me from buying it.

[fanoush]

I would remove the (optical) heart rate sensor too. For same reason as GPS - it is not good enough and may bring the cost and size down. However OTOH if the watch will be expensive anyway then the HW may not add that much to it, so each feature included may get more buyers. Or maybe other sensors could be more useful instead - ECG, body temperature.

For running if one cannot take phone then maybe some other device could add GPS and/or HR sensor over BLE (ring, chest strap). If you would add GPS to something like Puck.js that could solve the GPS issue too.

[gfwilliams]

Interesting - lots of thumbs ups. Do a lot of people feel the same way about the HR sensor?

Honestly it has been a bit of a pain. I like having it and the ability to see my heart rate, but we routinely get people grumbling about the accuracy in different scenarios and over the years I've spent a lot of time on it.

[thyttan]

My thumb is more about @fanoush delibering on some different tradeoffs, not in support for or against hrm exclusion.

But also the idea to integrate more with eg smart rings/separate small computing device (puck.js advanced/ pebble core concept). Like is already possible with hrm chest straps.

Taking the hrm chest strap as an example though, I like it as a complement to the included one and not a replacement. For now at least.

[fanoush]

But also the idea to integrate more with eg smart rings/separate small computing device (puck.js advanced/ pebble core concept). Like is already possible with hrm chest straps.

Those cheap (~US$15) aliexpress Colmi rings should be supported in GadgetBridge since this https://codeberg.org/Freeyourgadget/Gadgetbridge/pulls/3896 Too bad I did not try it as I am not thar much interested in HRM but I have some of those rings so this reminds me I should try it when running. I hope having it tightly fit on finger could help with accuracy. They report HRM rate over BLE so either one could connect it to Bangle or even use it directly from phone in GB instead of Bangle HRM.
Here is one shop I bought from about a year ago https://www.aliexpress.com/item/1005007167358142.html if anyone wants to try. They are (or were) also hackable and could possibly even run Espruino but that is not the point here, for HRM they can be used as is.
EDIT: here is some discussion about them too https://www.reddit.com/r/Gadgetbridge/comments/1iq75xj/colmi_rings/
EDIT2: here is list of currently supported Colmi rings https://gadgetbridge.org/gadgets/wearables/yawell/

I take HRM as something that is there because we repurposed cheap fitness smartwatches that have it as marketing bullet point, not that is is that much useful and actually working and being accurate. When running the wrist is probably the worst place to have the sensor. But if people want it and it is cheap and there is nothing better to put there instead then so be it :-)

[sonicityV]

Honestly ive never used the watch's gps as my use for the watch has been mostly as a watch with notifications, sleep tracking and step counting.
I dont think its worth the $30 usd extra of just chip costs (based off the SAM-M8Q-0)

If the swapable back idea goes through, then it might just be worth making a "fitness" back with both an hrm and a gps module.

[sir-indy]

I'm glad you suggested this, I was thinking it would be a good idea too! I'm aware it adds a lot of technical complexity, but as someone who doesn't use the GPS or the HRM I could happily live without them.

My personal use case for any smart watch is as a watch first, then for notifications, then for helpful apps (timer, counters, step counter, etc.).

[gfwilliams]

Thanks - I like the idea of a 'fitness back', but given the quantities of Bangles I ship and the R&D costs associated with it, I can't imagine it's ever going to make enough sense that I would personally develop it.

[RKBoss6]

That brings an interesting point though. Is it possible to have a customizable option, so when ordering the Bangle.js 3 you can either select the model with GPS or WiFi (as an example), or the base one, with HRM? Would it be worth it?

[gfwilliams]

I think I mentioned elsewhere, but GPS is potentially an issue because for it to work well enough you have to put the aerial somewhere. But you could in theory have a Bangle.js add-in back with WiFi (maybe LoRa) that made the Bangle a bit thicker by adding a second PCB.

I'm not sure if that'd be something it was worth me doing personally (it'd depend on demand), but for instance I might supply the CAD files for 3D printing a bigger back, as well as the PCB outlines so anyone could easily make their own versions.

[thyttan]

I admittedly don't use GPS on my watch that much nowadays. I can't help feel like it still makes sense though. Having it leaves the door open to more types of apps. Especially if the rest of the hardware is more capable in terms of computing and screen real estate.

Personally I'll most likely be buying the Bangle.js 3 regardless of how it shapes up. But I would not have bought my original Bangle.js 2 in the first place if it didn't have GPS.

For me the more important tradeoff is probably size vs function, and not money vs function. I prefer a not too bulky smartwatch, but if it would be as thin as or thinner than Bangle.js 2 either way I'd probably say include GPS.

As @fanoush says maybe it could be moved to a separate companion puck.js device a la the unreleased Pebble core, https://youtu.be/Pwq89K6RBTI?si=nJqhEjR_lNmBDVWQ&t=171 .

I think the new pebbles won't have GPS so it could be a differentiator there as well. Maybe you could offer PebbleOS versions with a GPS app included, making it the first pebbleOS watch with GPS.

My thoughts as of now.

[gfwilliams]

don't use GPS on my watch that much nowadays ... But I would not have bought my original Bangle.js 2 in the first place if it didn't have GPS.

Thanks for the honestly there! I think that's my concern - that maybe it's not used much, but it is something that makes people interested in it.

Having a small, programmable, wrist-mounted device with GPS does feel like something that's not really offered elsewhere. Even if I don't actually use it much I am excited about what I could do with it.

Without it and HRM, it's just a touchscreen display on your wrist.

[RKBoss6]

I agree with size; most modern watches are bulky, and not at all something I would wear daily. The bangle.js 2 is a perfect size, so keeping it light and small would be quite valuable for me.

[nxdefiant]

I use the GPS and I think it is one of the features that separates the Bangle from its alternatives like the pinetime.

[gfwilliams]

Just to put this out there (I don't know if it's feasible or not yet).

What if there wasn't GPS but there was WiFi instead? It seems a lot of users would value WiFi, and while I don't think an ESP32 makes sense as the main processor due to its power usage, we could use one for WiFi or even potentially share the IO pins for the display so that apps could have the choice of running on the ESP32 or the nRF52840.

[fanoush]

If you stay with Nordic there is companion nRF7002 wifi chip from nordic that can do 5ghz and is low power. It also comes bundled on some modules like this one https://www.aliexpress.com/item/1005005542789484.html However some recent esp32 can do 5ghz too(?) and some are finally better at low power https://github.com/espressif/esp-idf/tree/master/examples/bluetooth/nimble/power_save#typical-current-consumption-with-management-enabled (with external 32KHz XTAL added)

[gfwilliams]

My concern with nRF7002 is it'll require Nordic's new SDK I think, so a complete Bluetooth rewrite for me. It's a pretty huge change.

Recent ESP32 should be fine for 5ghz and it opens up quite a few new possibilities. There are a ton of toolchains for ESP32 (IDF, Arduino, MicroPython, etc) so it's possible someone could write a watch app in Python for instance and have basically complete control of the ESP32 for their app - which avoids having to worry about abstraction and interaction with Bangle.js itself.

• imagine being able to just boot up a Flipper Zero firmware on it, but when that's not running still having several weeks battery life.
• or run one of the ESP32-based emulators to emulate a game console, PC, etc...

[fanoush]

There are a ton of toolchains for ESP32 (IDF, Arduino, MicroPython, etc) so it's possible someone could write a watch app in Python for instance and have basically complete control of the ESP32 for their app - which avoids having to worry about abstraction and interaction with Bangle.js itself.

That is advantage but also adds complexity and fragmentation. But I am all for it :-)

My concern with nRF7002 is it'll require Nordic's new SDK I think

well sooner or later this should probably happen so that would be good opportunity. I think new port of bluetooth.c from scratch could give a shorter cleaner code and maybe it wouldn't be so much work compared to the rest. This would also allow you to go with new nrf54 family which has lower power consumption. also newer BLE features like LE audio could be useful

[jeonlab]

I don't really want a wifi on my watch. It consumes a lot of power, only works when wifi is available, and hard to enter passcode for a new wifi, etc.
I made a GPS watch using ESP32 (both wifi and bluetooth available) (https://jeonlab.wordpress.com/2022/09/25/gps-watch-v3-1-watch/), but used wifi only to upload golf course data at home which can be done through programming port.
So, my vote for the wifi is NO.

[RKBoss6]

WiFi isn't really valuable to me, seems like i'd always keep it off. I can get all data that needs wifi from my phone, with minimal software modification.

[thyttan]

A comment spotted re no GPS on the new Pebbles on the Verge's article on the new watch design:

[gfwilliams]

:) Interesting - thanks! Although it looks like nobody else actually voted that comment up as it's right at the bottom?

I totally see Pebble's point of view - they never had GPS (afaik?) so there'd have to be a big engineering effort to add that into Pebble OS - plus the cost increase in each device.

It looks like Eric's taking a very sensible view - just putting in what people loved about Pebble, and not much else.

Most people wanting GPS are mainly wanting it for running? But then they're also the people that are going to be disappointed that the HRM isn't as good as the one on a Garmin/Apple watch costing 4x as much, and half of them will be running in a city with high buildings around where the GPS struggles anyway. Maybe it's better to look at Pebble, and to stick with what we're good at.

[fanoush]

Most people wanting GPS are mainly wanting it for running? But then they're also the people that are going to be disappointed that the HRM

Not sure about that. GPS can be also used for hiking (or geocaching) where you don't care about HRM at all. also you may use something else for HRM and still appreciate GPS even when running.

[gfwilliams]

... I do use GPS for hiking myself, but if I'm honest I've always got a bag with supplies (and my phone) in, and even now it's possible to configure Gadgetbridge so that it can use the phone's GPS instead of the watch GPS, so it wouldn't really affect my experience.

[lightnin]

I like the idea of offloading as much as possible, whenever possible, to the phone, to keep the watch as slim, light, inexpensive, and running as long per charge as possible.

[pinq-]

I understood that, if you want good GPS, it needs a lot of money and good hardware. I have bangle 2 and I tried running software, but it never worked as well as my sport watch. So every time I do sports I change to sports watch. So really, no need for gps, because I think you can't get it good without huge effort.
Also HR sensors in watches are...well.. not good. I use heart rate belt and it differs a lot even with my sport watch HR sensor. So no need for gps and hr.
LoRa or wifi would be cool or something else, that is not common sensor. 16g acceleration sensor and gyro would be cool ^^

[gfwilliams]

Thanks - yes, I think it's safe to say that whatever I do for HRM and GPS it's not going to be as good as a £400 watch by a company with a huge R&D budget.

It's a hard one because I think there are a lot of quiet people on the forum that do use and value GPS and HRM, but at the same most of what I get as feedback is complaints when they don't work perfectly.

... but if I took the approach of not shipping things people complained about, I'd never ship anything :)

[pinq-]

Yeah, that's true. You would make bias conclusion, if you only read complaints. On the other hand, you can get some idea, what are most wanted sensors with what are most used/downloaded apps. Also, if something gets a lot of complaints, it is also widely used.
And from business point of view, more is more. More sells better than cheap price. You can have sales pitch " now with better gps" :D

[lightnin]

To me the value of Bangle JS is its open source hackability. The fact that I can write my own app or watch face and share it with a larger community is the killer app. In this it is somewhat like LEGO: the value is in what you can make with it . That's the domain in which I think Bangle can compete in and win.

[jeonlab]

It seems many people don't use GPS on Bangle.js, but I use it a lot for golfing. I really hope the next version to have the GPS. I don't really care about HRM though.
Also, after I modified my Bangle.js 2 for 4 additional buttons and piezo buzzer, I'm happy that I did the modification. Additional buttons for custom inputs and the buzzer are very useful.

[gfwilliams]

Thanks! And yes, I totally agree about the buttons and ability to make sound - both something I'll be trying to include.

[bobrippling]

I use GPS a lot but for the BangleJS2 I carry my phone as a backup on races as I can't rely on the Bangle getting a signal. I'd love to not have to carry my phone on runs or rides! I use the GPS daily too, it's the main feature after BLE I make use of.

HRM I would use, but I found that the one on the Bangle wasn't accurate, or I had it setup wrong, so I don't use it at the moment.

[mbattistini]

For me, the lack of GPS wouldn't be a problem. I used it a little at the beginning, but mainly due to battery issues, I always used my smartphone when I needed it and I don't miss it.
I've never used Wi-Fi, though.
Today, however, waterproofing would be essential for me.

[bobrippling]

+1 for waterproofing

[gfwilliams]

Don't worry, waterproofing is essential for me too!

[WhelanB]

While I could never get the GPS working correctly on my first BangleJS, I've found the GPS on the replacement I bought to be a lot more reliable and have been using it a lot lately mounted to my bike handlebars with Gipy running to follow gpx routes. Especially if the BangleJS 3 is going to have more colours on screen which could display more detailed maps I'd love if it could still have GPS, but understand also that for example my phone can supply coords as a substitute

[baldursigurds]

I would definitly want gps. The more accurate, the better :)

[dashavoo]

My Bangle2 just died so the reason I saw this was looking for any updates on Bangle3.

I wouldn't have bought a Bangle2 without GPS and HRM, and I wouldn't buy a Bangle3 without them. I bought the watch for running in the first place. The main thing I would wish with the GPS is faster time to lock - even with aGPS I sometimes give up. I don't take my phone with me when I run. I hope having GPS wouldn't make the watch too much bigger than the Bangle2 - it's pretty much the perfect size at the moment. The only features I use more than GPS on Bangle2 are telling the time and the vibration alarm.

I have no interest in WiFi in a watch. I'd keep it turned off.

If we are comparing things to the new pebbles, I also have no interest in a microphone. In fact, I think it's an antifeature. There are enough microphones around me all day.

[kingcharles1666]

Exactly my view. A faster fix would make GPS more useful. I give up on it too if it doesn't fix within half a minute

[gfwilliams]

The GPS lock can be quite fast now if you've uploaded AGPS in the last 12 hours or so, but the AGPS data on Bangle.js 2 only lasts 24 hours. A different radio could improve on that - the ublox ones were good for 7 days

[numerist]

GPS can be flaky, on a watch more so. Not interested. Also not interested in Pebble. Bad management killed it in 2016, and once is enough for me.

[RKBoss6]

Honestly, I haven't used the GPS at all since I got my watch (3 months ago). I just don't see the need of it if it's paired to my phone, and can get the location from there. In my mind, I prefer a long battery life over GPS....

That being said, if we do add the GPS, we might also be able to squeeze a bigger battery in there as well. Even though the watch advertises up to a month battery life, realistically, it's just not possible, and if you do manage it, you can't use it like it's the smartwatch you expect. Currently I charge my Bangle every 5 days or so. It would be a quality-of-life upgrade, that also would contribute to longer life with GPS on.

Nice that you're designing the 3 from scratch! It's pretty cool when it's not just open-source software, but right now open-source hardware, in a sense, where the whole community can come together and discuss about what all should be in the next generation of the Bangle, creating something amazing in the process!

[RKBoss6]

In my opinion, HRM should be something to keep. Most smartwatches have an HRM, and that's one of the reasons I use it as a daily driver. With the bangle.js 3 already more expensive that 1 and 2, having that, and then paying more just to get a smart ring or another heart rate detector feels quite underhanded to me. I use the HRM every day, and I don't mind that it's slightly inaccurate. Maybe a high quality sensor might combat a few of these issues, but I'd say definitely keep it in...

[gfwilliams]

Thanks! I didn't think I'd really be in a position to do something entirely custom, but it looks possible. It's still a pretty huge expense, and I'd have to commit to an initial order of 10,000 watches which is obviously a lot of money. Realistically what's going to happen I think is I'll pay for the R&D out of my own pocket, but to actually get the 10,000pcs order done I'll have to do preorders again, so if not enough people want one then it may not happen.

I think realistically we will have an HRM - because all the smartwatches have them, you can just get an HRM on a flex PCB, with the correct rear glass, for not that much money.

It seems GPS is another matter though because fitting the aerial in the case and tuning it correctly can be very tricky, especially in a round watch. I'm still on the fence about it - it really shines if you're a runner, but if you have your phone with you everywhere (and it's Android, not iPhone) then as you say GPS seems a bit pointless.

I am hopeful we'll be able to get a bigger battery as at minimum there is newer technology and the round case should give us some extra volume for it, but also with the 'bigger back' option that's suggested it should be possible to put a massively bigger battery in there if needed.

[RKBoss6]

Alright, thanks! Having a round watch would be nice - are you planning on the screen being round as well? Also, just an fyi, technically you can setup an iPhone to push the location to the bangle using a location shortcut as well.
As for the 10000 orders, have you considered advertising? Not like a big company that focuses on getting sales, but just to pique interest in it...