Universal battery

[blueskull]

Can anyone suggest me an easy to obtain universal battery pack for my next project?
Something like BL5C, easy to obtain (genuine ones), with protection, used by other projects and has off the shelf connectors available from major distributors.

I'm looking for 2000mAh+, which BL5C is only rated 1020mAh.

It must has built-in fuel gauge! Like Sony InfoLITHIUM, but it must have a comm protocol that has been reverse engineered.

[tautech]

Do you want a flat pack ?

16650 18650 cells not suitable ?

[tautech]

These look interesting:
https://www.aliexpress.com/item/Free-shipping-3-7V-3000mAH-655565-PLIB-polymer-lithium-ion-Li-ion-battery-for-dvr-GPS/32613641699.html?spm=2114.01010208.8.14.dQII0d

[tautech]

Or if you need more grunt:
https://www.aliexpress.com/item/3-7V-10000mAH-40120130-LIIB-polymer-lithium-ion-battery-cell-Li-ion-battery-for-onda-sanei/32608668904.html?spm=2114.01010208.8.4.dQII0d

[tautech]

Might not be industry standard but what about iPhone batteries:
https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20170213165252&SearchText=iPhone+6+battery

Sockets/terminals:
https://www.aliexpress.com/wholesale?catId=0&initiative_id=&SearchText=iPhone+6+battery+terminal

[tautech]

Might not be industry standard but what about iPhone batteries:
https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20170213165252&SearchText=iPhone+6+battery

I mean a battery that is swappable without having to open the case and expose the PCB.
Like a swappable phone/camera/camcorder battery.

Something like these with a bit more grunt ?
https://www.aliexpress.com/item/2pcs-1800mAh-Li-50B-D-LI92-Li-50B-D-LI92-Li-ion-camera-Battery-USB-Dual/32603362240.html?spm=2114.01010208.3.55.a9iEZk&ws_ab_test=searchweb0_0,searchweb201602_4_10065_10000073_10068_10501_10000074_10503_10000032_119_10000030_10000026_10000023_10000069_10000068_10060_10062_10056_10055_10000062_10054_10000063_10059_10099_10000020_10000013_10103_10102_10000016_10096_10000056_10000059_10052_10053_10107_10050_10106_10051_10000097_10000094_10000053_10000007_10000050_10084_10083_10000100_10080_10000047_10082_10081_10110_10111_10112_10113_10114_10115_10000041_10000044_10078_10079_10000038_10073_10000035_10122_10121-10503_10501,searchweb201603_7,afswitch_3,ppcSwitch_3,single_sort_2_default&btsid=e9a4d473-2fae-432d-b64a-155974adc4b3

Sorry for the massive URL.

[tautech]

Nikon EN-EL23:
http://www.ebay.com/itm/2x-EN-EL23-ENEL23-Li-on-Battery-Charger-for-Nikon-COOLPIX-P600-Camera-High-Q/321718531958?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D41375%26meid%3Dd42707110b0d491ea57d76def0cbadef%26pid%3D100005%26rk%3D6%26rkt%3D6%26sd%3D221576260561

[DBecker]

Have you considered a Samsung phone battery?  The S3 and S5 were very popular phones, with a vigorous aftermarket of replacement parts including the battery contact assembly.  The contact assembly is quite easy to use in your own design, with a serpentine spring allowing loose tolerances on the enclosure.

[BravoV]

Look for Panasonic's camera battery, top quality and good availability even the cameras are discontinued.

[NiHaoMike]

If you're not running a particularly high current, you can get a very good idea of the charge state by measuring the voltage.

I presume you're using a general purpose embedded system like a Raspberry Pi Zero? There's a lot of room for improvement - I once had a Dell handheld computer that would easily get 2 hours playing video with the stock 900mAh battery, many times that playing music. Some iPods from 10 years ago get 7-10 hours runtime playing FLAC with a tiny 450mAh or so battery. An iPod of that era only has an 80MHz ARM CPU, so an ARM M4 would probably be the best modern equivalent.

[Siwastaja]

Considering the recent history of Samsung, I rather not touch their product either.

This is just stupid. Samsung has one of the best li-ion R&D and products in the world.

It's as brand-name and as reliable as it can humanly get.

Every manufacturer have had issues, both technical and philosophical. For example, I could find a lot of reasons why you should boycott Sony. At least their scandal of exploding li-ion cells about 10 years ago was about cells actually manufactured by them (unlike the recent Samsung issues, which weren't related to Samsung SDI at all)! Still, Sony is one of the best in li-ion - as is Samsung SDI. Or Panasonic, even if they produced some crappy capacitors decades ago.

[NiHaoMike]

So you're basically looking to build (much of) the DAC from scratch. Quite ambitious when a PCM1792A (built on an antiquated 180nm or so process) uses on the order of 250mW, and that's including all the logic needed to accept I2S.

There hasn't been much demand for standalone low power HD DACs until USB-C smartphones appeared on the market. And those are almost certainly very tiny BGAs in interest of minimizing size. (Maybe we'll be lucky to find one in a QFN package instead?)

Maybe look into doing a "direct DAC to headphone" design? It's basically just a Delta Sigma stage that switches between a positive and negative reference voltage, with a tiny LC or RC filter on the output. More or less a very scaled down version of a digital amplifier. (At least some USB-C DAC chips do actually implement it that way.)

[TerraHertz]

For a product with user-swapping of batteries, then battery lowest unit cost, robustness, availability and convenience is top priority.

So why not use plain 18650s? 'Bare' ones, ie no built-in over-discharge protection.
Include slots for two or three (or four) batteries, with completely independent battery monitoring, charge and drain circuits for each slot, so there's no need for a user to keep track of which batteries are in what state. All converters transfer power to a common system power rail. Also can insert variable numbers of batteries, one to four as needed/available. Just change one, to all of them, when needed. Can change one at a time without interrupting play.

You know LiIons have a quite linear voltage to charge state relationship, so no special charge tracking is needed? Just measure the voltage.
However this is why your battery slots can't be just wired in parallel. LiIons *can* be run in parallel, but not if users can swap batteries. Because they will plug in batteries in highly different charge states, with destructive results.

[NiHaoMike]

Multibit DACs could be run direct into headphones. In fact, I did that with the PCM1792A for testing purposes. (Terminate outputs to ground via 10 ohm or so resistors, connect headphones between differential outputs.) The sound quality was very good at a very usable volume, albeit in mono. It's too bad that very few headphones are differential as that would make it easy to do a "direct drive" configuration from a differential DAC.

The 1792A is a current output DAC so it was not being used optimally in that configuration. Hence why a proper application has an I/V converter stage - 2 opamps per channel as it's differential. The output driver is another opamp, a pretty high power one in my case as I wanted multiple headphone outputs. The only other opamp used is low power, low offset for active offset correction to prevent a very noticeable click when the output relay engages or disengages. +-5V is plenty for this application in my experience, could go lower if the opamps will work that low.

As for paralleling multiple cells, you could put PTCs between them to limit the current if the user mixes cells of very different charge states.

[TerraHertz]

Li-ion batteries do need fuel gauge to accurately measure remaining energy and battery life. Voltage measurement is highly unreliable when temperature changes, which could happen if you carry one of these in your outer pocket.

Look at that chart again. How much is the difference over *reasonable* room temps? Even down to zero deg C. Not much.
So, just include a temperature sensor and adjust power estimates.
Bear in mind that your alternative is to include full charge input-output tracking for each battery pack IN THE PACK, which is going to add a lot to your system cost, and be annoying for users who want to acquire multiple battery packs. And will STILL have the same kinds of temperature inaccuracies. Probably more in fact, due to no processing power.

Protection is absolutely needed in swappable batteries.

Depends what you mean by 'protection'. If you mean overdischarge protection, no, with swappable batteries that is better put in the device. Ditto charging control, if the device is able to charge batteries.
If you mean 'protection against accidentally shorting out fully charged batteries loose in your pocket', this seems to be something some people worry about. Which is why we don't have many nice products using bare 18650 cells. And I think that is stupid; another example of nanny-State.  We *FINALLY* have pretty good batteries, but now we're supposed to accept that because of being good, they are too dangerous to use? Screw that. For people who can't keep batteries stored sensibly on their own, the solution is to sell convenient spare-battery storage holsters.

As for paralleling multiple cells, you could put PTCs between them to limit the current if the user mixes cells of very different charge states.

Won't work well. Which cell node are you going to use as supply source? Also, letting the cells self-equalize messes with the user's idea of which cells he just put in fully charged, and which he needs to change next. Plus if you allow enough current flow to self-equalize it's not an optimal charging profile for the cells.
I still think it's better to have completely independent switcher supplies, one per cell. Check Aliexpress to see how cheap little single cell LiIon switchmode supplies are.

eg
1-6V to 5V buck boost DC DC converter Voltage Board for portable Solar Charger.    2 for US $4.99
https://www.aliexpress.com/item/2pcs-1-6V-to-5V-buck-boost-DC-DC-converter-Voltage-Board-for-portable-Solar-Charger/32662597171.html

TP4056 Charge And Discharge Protection Module Over Current Over Voltage Protection 18650 MicroUSB   5 for US $3.73
https://www.aliexpress.com/item/5pcs-lot-TP4056-Charge-And-Discharge-Protection-Module-Over-Current-Over-Voltage-Protection-18650-MicroUSB/32455748917.html

Edit to add:

Couple of examples of 9gag posts demonstrating how many people feel nostalgic for a 'nice simple music player with replaceable batteries' -

  http://9gag.com/gag/aopwq12
  How many of you owned one of these?

http://9gag.com/gag/aopwq12#cs_comment_id=c_147558621096335412
My comment:
Replaceable battery and NO DRM bullshit! Perfect. Mine died eventually, then I bought two Thomson 6677. Similar thing, but even better. Has a uSD slot, and I still use them.
If I could find something like these, using an 18650 cell for months of operation, a bigger screen, but still acting as just a simple flash drive with no stupid over-complication, no DRM, I'd be very happy.
--------------------

20170206
http://9gag.com/gag/aLDK5EM
Admit it, you had one of these

http://9gag.com/gag/aLDK5EM#cs_comment_id=c_148637817914938223
My comment:
Had 2 of these, then a Samsung thing that looked great but was DRM-ridden junk. Next bought 2 Thompson 6677. Splendid, still use. uSD slot FTW! But sick of using non-rechargable AAAs. Alternative similar ones I can find all use pathetic inbuilt batteries - only last a day between charges, can't swap them, and they eventually fail.
The Thompson model is no longer available and mine will eventually die. And then...

Currently starting a project to design a better one, with replaceable bare 18650 Liion battery, slightly better screen, uSD only (no internal flash to fail), built-in storage compartment for many uSDs, WAV and MP3 play, record to WAV with ONE BUTTON press (not bloody 8-button sequence so hard to get right), external mic option, plain disk filesystem, still zero DRM crap. With 18650 the play time should be MONTHS before need to charge or change battery. Open hardware & source, 3D printed body.

Fnck the corporations who refuse to get it right.

Who would like this?
------------------------------------------

Blueskull, your project interests me because I too wish there was something better than the current commercial offerings. But I don't mean 'better' in terms of audio fidelity. I mean usability. Here's a note I made back in 2013:

20130330

A Sensible music player
-----------------------
Prompted by the unavailability of small, replaceable battery MP3 players like the two Thomson 6677 I bought years ago.

Features of the Thomson 6677
 * One battery. Can be alkaline rechargable, or standard. ie voltage range 1.2V or 1.5V
 * Plays music for days on a single battery.
 * In standby (off, but still responds to 'on' button) the battery charge lasts for months.
 * Has a uSD card slot. (But can only take up to 2GB size cards.)
 * USB plug. (USB ver 2)
 * Appears as standard disk drives. One for internal mem, one for the uSD card.
 * 1GB Internal flash mem.
 * Small low power monochrome LCD screen, with RGB LED backlight.

Flaws:
 * The shape. It has a bulge at rear for the battery, that makes it not sit flat.
 * Tedious button-press sequence to get it into record mode.
 * No VOX record mode.
 * Internal mic is OK, but has no Mic input.
 * Screen text is too small for poor eyesight.
 * Does not use the USB interface for config.

Chips in Thomson 6677:
    Sigmatel STMP3510        L100E-SB6  A03167  0714
    Hynix    HY27UT088G2M    TPCB 820H

See: F:\_camera\20140729_thomson_6677_music_player

These days all players I can find have internal batteries, that do not last even 1 day. Then they take many hours to recharge.
And the battery cannot be easily replaced once it's dead. The short play, downtime during recharge, and non-replaceable battery are all individually completely unacceptable.
And uSD slots are very rare.

Features I want
---------------
* Cheap. Under $30.
* Battery compartment can take anything: 18650A Li-ion, AA, AAA alkaline or NiMh.
  Aims: - Operating battery life of at least a month between recharges, using standard bare 18650.
        - Quick swap with new battery,
        - Recharging: Can recharge internally from USB, but normally battery just swapped out and recharged externally.
        - Can run with no battery, with power via USB port.
        - Battery compartment & lid are solid, long lasting, survive drops and batt leaks.
        - Battery type set in config: [Li-ion, NiCd, Alkaline, Carbon-Zinc] (sets params, including whether rechargable.)
        - Battery management and protection is in the device, not the removable battery.
          Configurable:
             - discharge protection voltage limit,
             - charge protection voltage limit,
             - charge current.
             
* Shape: Flatish. Little thicker than an 18650. So battery compartment to the side.
* Upgradable firmware.
* Open source design of case, circuit, PCB & software.
* Internal flash used mainly for firmware. Some space can be used for audio, but normally use plugin flash to avoid
  wearout of the internal memory.
* uSD flash card slot, taking modern cards. 32GB+  Can play from and record to these.
* USB master port, allowing use of standard USB thumb drives for storage. (Optional, with uSD slot.)
* USB slave plug, for connection to a PC/laptop/etc.
  (Maybe can combine the above, using a 'bare' plastic tongue that can be master or slave. Screw the stupid standards.)

* Built-in microphone.
* Microphone jack, mono.
* Earphone jack, stereo.

* Operation configuration done via PC over USB, or via config text file in the uSD card.
  All config data retained on player in one plain text file, in accessible location.
  Config changes via:
    - Web interface (browser, read html/js on device)
    - Use text editor directly on config file.
    - Load new config file.

* All sound, config and software files are standard filesystem, accessible via USB from a PC.
  NO SPECIAL SOFTWARE or drivers required on the PC.

* No 'proprietry song database' bullshit. Tracks just play from a selected folder in alphabetic order, and
  displayed as their unaltered filenames. Or can select a playlist to run, but the playlist is just a simple
  text file containing sequential list of path-filenames to play.

* Largish screen, colour, pixel driven.
     - Backlight white LEDs, config brightn.
     - Font size adjustable.
     - screen can be config'd to:
         - stay on while device active.
         - stay on while menus in use, otherwise power off.
         - configurable on period at start of new track play.
     - screen on/off and backlight on/off are separate.
     - If scren is touch sensitive, there MUST be a physical, firm-detent LOCK SCREEN slider switch (same as locks any buttons.)

* Plays MP3, WAV, FLAC and others. Must include at least one non-lossy compressed format.

* Recording
    - Recording format: WAV.
  Configs:
    - VOX on/off, trig level, silence rec time.
    - Timestamping format.
    - Break files vs continuous. Max file size.
    - Allocated storage size
    - Storage location: (Internal, uSD or USB, plus folder name.)
    - Overwrite (ring buffer) vs no overwrite.
    - Record ON: <button def> (Can be single button press, quick, or x seconds hold. Or can be menu entered.)
    - Record OFF: <button def> (As above.)
    - State lock: <button def> (Can be any simultaneous button hold/sequence, including the LOCK slider.)
    - External record wired-enable of some kind. Perhaps using the earphone jack configurable as a switch or logic input?

 Note: The Thomson 6677 seemed to be designed to be difficult to get into record mode. Whether this was intentional to
 meet some legislation related to recording conversations, who knows (or cares.) It's bullshit.
 To go from off to recording took:
 1. Slide 'Power-hold' button left momentarily to turn unit on.
 2. Wait for several seconds while it boots up.
 3. Button 'M' to get into 1st menu.
 4. Button 'M' again to switch to menu with 'recording' as 2nd item.
 5. Button '-' to bump selection down to recording.
 6. Button '>>' to select 'recording'. Now 'record' is 1st item, 'play back' is 2nd.
 7. Button '>>' again to select record. Now says 'voice recording ready'.
 8. Button 'red dot' to begin recording. Now pretty much any button will stop recording.
 9. Slide 'Power-hold' button right into 'hold' position to lock unit.

 That's ridiculous, and makes starting record in an urgent situation impossible.
 Needs ONE button that starts recording immediately, even from powered off condition.
 Together with a 'lock' slider that works with the unit on or off, to prevent accidental accidental startups
 or drop out from recording/play, that's all that's needed.

* Internal RTC & calendar (for timestamping recordings) operates from internal coin cell or supercap
  when main battery removed & no USB power.
  If coincell: Several years runtime. If supercap, allows 5 minutes to change battery.
---------------------------------

This 'project' (ha ha, as if I'll ever actually do it) is on the list of reasons I bought a 3D printer. Because it seems NO ONE is ever going to do a small music player that takes 18650 batteries. or if they do, it will be DRM-poisoned. So, if I can print the needed case...

Recently I came across this:
 https://www.aliexpress.com/item/DFPlayer-Mini-MP3-Player-Module-MP3-Voice-Module-for-Arduino-DIY-Supporting-TF-Card-and-USB/32691455792.html?spm=2114.13010608.0.0.va4k3D
 DFPlayer Mini MP3 Player Module MP3 Voice Module for Arduino DIY Supporting TF Card and USB Disk  $2.49

I bought a couple. Reversing the circuit seems like a good start. Because I like the KISS principle, and don't care at all about extreme audiophool fidelity, or exotic output drivers.
It just has to sound OK on cheap earbuds, be SMALL, CHEAP, and the battery must last for weeks and be quick-swappable. Sensible UI would be good too. Readable text, but not a large colour screen - ha ha, what for?



Pic is one of my Thompson 6677 players. About as good as they got, before the designs went stupid. The Samsung being a perfect example of insanity. God that was a useless piece of crap.

Edit *again*: Oh, and one more thing. If making a flat case that has the electronics & screen next to one or two 18650 batteries, there should be room to include a storage space for many uSD slivers. They are so tiny...  Something with slots to hold them, and a slip-in space for a label of what's on them. (Or at least numbered slots, with a text file index in internal mem, accessed via the screen.)  Physically a bit like a tiny CD storage rack, except they could be slanted for less depth. Some kind of spring detent on each slot, so the uSDs can't all be sent flying if you fumble opening the cover.

[Kalvin]

Summary on the battery gauge technology:

https://www.digikey.com/en/articles/techzone/2014/jan/fuel-gauge-ics-simplify-li-ion-cell-charge-monitoring
https://www.digikey.com/en/articles/techzone/2013/jan/advanced-ics-simplify-accurate-state-of-charge-measurement-for-lithium-ion-batteries

Probably the coulomb counting with the voltage measurement is the easiest way to get decent accuracy. TI has some advanced battery gauge devices which may offer better accuracy.

[Siwastaja]

I suspect the voltage curves you showed are either bogus, or from some rather exotic chemistry. I have never seen a cell range from 4.2V at full charge to 3.7V when almost empty; typically, the open-circuit voltage range between 100% and, say, 20%, is more like 4.2V to 3.5V.

Temperature is only relevant at relatively high discharge rates, as it doesn't affect the open-circuit voltage, but only the ESR.

Estimating the SoC from voltage only is highly recommended in simple single-cell applications, especially with slow discharge rates, because it is guaranteed to give usable approximations, whereas more complex means, such as charge integration, while very precise in optimum conditions, may easily end up showing totally random shit. Charge integration tends to drift, and realigning it is based on the voltage estimation, anyway. Reliable 5-step battery gauge, which may momentarily show 1-2 steps too little under a peak condition, is intuitive and very useful, compared to a fancy percentage reading showing steady 92% when the battery in reality is at 25%.

And you can average the shit out of the reading to get rid of short dips.

Fancy "fuel gauges" have been buggy since their advent, and still are difficult to get right. Voltage-based estimation works on li-ion because the curve is gradual. Voltage-based on lead acid, NiXX, (or LiFePO4), on the other hand, is almost useless.