Modifying a Tile (BLE Tracker) for an Electric Bike

[BAM5]

Hey all,

I'll just preface this with my skillset. I love to disassemble, diagnose, and repair devices & circuits. I have not designed a circuit although I understand EE to a decent extent to where I believe I could with a bit of help.

Alright, so GF is moving to a large metro area where having an electric bike is super helpful, and also super desirable. So she wanted to put a tile inside the metal frame of the bike (which she quickly realized was a bad idea when I mentioned signal strength,) but was excited about the prospect of my idea when I brought it up, which was to take the tile and it's 1 off battery and power it from the E-Bike's battery.
This way any would be thief would help with getting themselves caught as long as they keep it charged as they use it. Muahahah

So my idea is
Disassemble and find space inside the e-bike's battery pack to place the tile's PCB & additional components
Devise a way to step down E-Bike Battery (henceforth EBB) Voltage (13S li-ion [54.6V max]) & ideally charge a small dedicated rechargeable li-ion cell for the tile. This way if the EBB hits the low voltage cut off the tile will still be able to be powered for a year or 2.

From my experience, this is my plan so far
I'll need a buck regulator placed post EBB BMS that steps down the max 54.6V to 4.1V which will then power an added Li-Ion BMS (henceforth TileBMS) connected to the dedicated cell for the tile. There will then be another buck regulator that will step down the dedicated cell to 3V which is what the tile expects (Originally powered by a 3V CR2025) The Tile is extremely power efficient (typically lasts one year on 150mah) so all components can be super low powered which should make them small.

My biggest constraint is volume. I have a considerably small space to work with so all these components will need to be small and I am not sure how to search for breakout boards by size. Also I am not even sure if such small breakout boards exist. Theoretically they should. I've done some searching but haven't come across any.

So I was hoping someone with more experience could help me with what I should do from here now that I have the theoretical plan, and procurement of the devices I'd need.
Also, I'd love to hear your thoughts on this project.

Thanks for any help you can offer!

[jeduffy]

Not really sure what help you're looking for, you seem to have a decent handle on things, but one thing I would watch out for is keeping a li-ion charged for a long time, you typically want to keep them around 3.8V, otherwise they start to puff up pretty quickly (on the order of weeks, depending on the exact cell).  Other thing would be disassembing the battery, presumably it's waterproofed in some way, and waterproofing is very easy to mess up if you aren't particularly careful.

[jeduffy]

Also just realized, if you're using that little power (17uA average), a linear regulator may be more efficient, depending on the regulator you use, and will definately be available in very small sizes.  Something like this https://www.digikey.com/product-detail/en/ablic-u-s-a-inc/S-1142B39I-E6T1U/1662-1352-6-ND/6601952 may work, (designed for 50V max, but has an absolute max of 60V, just make sure you've got some good filter caps).

[nali]

Be aware that those "key finder" type things don't actually track! They are simply just a BLE beacon typically serving proximity and push notifications for the bleep; any "tracking" as such is done by the phone app.

So the only useful function you could hope for is a proximity alert so if the bike goes out of range from your phone you'll get an alert. Once the bike is out of BLE range that's it unless you get within whatever the range is again.

[BAM5]

I have a handle on the concepts, but no experience with actual design. What I suppose I'm looking for help for is ic recommendations, or a good method to find ICs that fit my constraints. And also a review of my design when I finish it. So far I've just been googling and filtering through a single manufacturer's catalog at a time since some distribution sites (mouser) don't support the parameters I am constrained by.

So far I've found this frigg'n tiny IC (TI TPS62740) which fulfills my requirements (90% efficiency at 10ua) but it's sooo small I don't know if I'll be able to utilize it. Does JLCPCB support trace spacing that small? Would I even be able to solder it to the board with my fx888d & Cheapo rework station? 

Yeah, I was wondering if having a li-ion batt constantly charged at 4.2V would have adverse effect on the cell. I was thinking I'd drop it down to 4V or something. Would that mitigate issues? Or would I need to have it like 60% charged at 3.8V?

I already disassembled the batt to see what kind of space I'm working with. Not much for waterproofing except for channels in the plastic that would prevent water from getting to far in unless the battery is flipped around multiple times.

I thought the thing about linear regulators is that their efficiency is determined by the difference between the two voltages you're utilizing. So no matter what 4V to 3V will always run at 75% efficiency (and increase in efficiency as the battery voltage drops). Also, don't linear regulators have an inherent voltage drop? Would that be my limiting factor for efficeincy with a linear regulator (Max efficiency would be BattV = TargetV[3V in my case]+LinearVDrop) Or am I misunderstanding something about Linear regulators?

Also, I assume current draw goes up significantly when the speaker is used. Maybe 5ma? 10ma? 50ma? I'm not even sure. But I have to be prepared for that as well.

Be aware that those "key finder" type things don't actually track! They are simply just a BLE beacon typically serving proximity and push notifications for the bleep; any "tracking" as such is done by the phone app.

So the only useful function you could hope for is a proximity alert so if the bike goes out of range from your phone you'll get an alert. Once the bike is out of BLE range that's it unless you get within whatever the range is again.

You are correct! Except for one small oversight. With Tile (and similar products) the way it works is that their phone app forms a network of bluetooth connectivity. Anyone with the app installed on their phone will be an extension of your detection range. So in the background the app will report the location of any tile within its range to the tile servers, which will then get relay'd back to you. I personally don't own any tiles but have the app installed just to help out anyone who has misplaced their stuff. I dunno if I've actually helped anyone or not by doing this since the app doesn't notify you of other people's tiles within range, but I like to think I help.

[NiHaoMike]

LTC3639 would be near perfect for your use.

I really think that some implementation of embedded security (e.g. challenge/response with the key) in the inverter would be a very effective security measure that doesn't cost much to implement. Could even be implemented so that trying to walk the bicycle with it locked would generate a voltage that activates a crowbar circuit and makes it very hard to move.

[BAM5]

I really think that some implementation of embedded security (e.g. challenge/response with the key) in the inverter would be a very effective security measure that doesn't cost much to implement. Could even be implemented so that trying to walk the bicycle with it locked would generate a voltage that activates a crowbar circuit and makes it very hard to move.

Hah! I like the way you think.
Though I'm not modifying it that much, nor do I have the time to do so. 

LTC3639 would be near perfect for your use.

After looking into the ltc3639 that looks like a solution for going direct from EBB pack to the tile, but it doesn't seem as if it would be very efficient when I look at the datasheet graph. Looks around 25-40% efficiency at 100ua which is 5x what I'd expect it to use 99% of the time. Is there something I'm missing? Some configuration that would make it higher efficiency at such low load?

[NiHaoMike]

After looking into the ltc3639 that looks like a solution for going direct from EBB pack to the tile, but it doesn't seem as if it would be very efficient when I look at the datasheet graph. Looks around 25-40% efficiency at 100ua which is 5x what I'd expect it to use 99% of the time. Is there something I'm missing? Some configuration that would make it higher efficiency at such low load?

Since you already plan to have a backup battery as well, set the current limit so the regulator would operate at its most efficient point, then use some micropower circuit (e.g. comparator) to switch the regulator on and off.

[BAM5]

Unfortunately I'm not quite sure what you're saying. My guess is use the regulator you listed to power a battery charging ic, run the tile psu off of the battery, and then some circuitry to cycle the battery charging? I'm not sure what I could use to do the... Oh hmmm... Ok, so maybe I can put the circuitry that disables the regulator after the regulator (Running on 3V or so), and defaultly the regulator is enabled just in case the battery goes dead, that way the regulator comes on automatically and only gets shut off when charging is full. I'd need some sort of flip flop or latch that would disable the regulator once battery is full, and then kick it back on when the battery hits a certain voltage.

Ugggh, I get the concepts behind this but doing up a circuit and finding ICs that do what I want how I want is proving frustrating.

[NiHaoMike]

Just set it for 4.2V and add a diode in series with the inductor, then you won't need a separate charger chip. Then have a separate comparator (micropower supply monitors work well) with a trigger point around 4.1V switching the regulator on and off.

[nali]

You are correct! Except for one small oversight. With Tile (and similar products) the way it works is that their phone app forms a network of bluetooth connectivity. Anyone with the app installed on their phone will be an extension of your detection range. So in the background the app will report the location of any tile within its range to the tile servers, which will then get relay'd back to you. I personally don't own any tiles but have the app installed just to help out anyone who has misplaced their stuff. I dunno if I've actually helped anyone or not by doing this since the app doesn't notify you of other people's tiles within range, but I like to think I help.

Ah OK... I missed that, it makes the concept a bit more interesting. I'd looked at a couple of similar products a year or two ago with a view to integrate or adapt them and they were pretty basic using cheap Chinese fob hardware.