So, about that Kentli battery…

It’s been a while since I’ve posted about the Kentli PH5 battery, which is a Li-ion cell with an integrated 1.5-volt regulator, wrapped up in an AA-sized package. Although I haven’t written much about its performance yet, that doesn’t mean I haven’t been doing work on it. In fact, I’m sure I have never put so much work into a single blog post before!

The full analysis of the battery’s performance is not fully complete, but I’ll reveal some details of my test setup and what I’m currently working on:

Analysis

I’m doing a much more thorough analysis of this battery than I have done with any other one on this blog. I have created a second bq27541 fuel gauge board, but with the explicit goal of measuring the voltage, current, passed charge (mAh) and temperature of a given DC-DC converter. This way, I can measure the input and output of the DC-DC converter simultaneously, greatly enhancing the data I can collect.

These are the data points/attributes I am currently collecting:

  • Battery voltage sag at high load currents
  • Battery capacity over different load currents (it’s not constant!)
  • DC-DC efficiency, both at different load currents but also over a single discharge cycle
  • Temperature rise of the DC-DC converter at different loads, and also over a single discharge cycle
  • Changes in battery capacity and internal resistance over many charge cycles

I want to be as thorough as possible with my measurements, mostly because nobody else has done a detailed performance review of this rather unusual battery, but also partially because I want to challenge myself and see how much of a “real engineer” I can be (#JustHobbyistThings). ๐Ÿ˜›

Advertisements

Reading out HDQ-equipped battery fuel gauges with a serial port

Battery fuel gauges are the unsung hero of the battery world. There’s more to it than just measuring the voltage on the battery terminals,. These little chips are microcontrollers (tiny computers, essentially) that sit inside the battery pack and keep tabs on the battery’s performance for the life of that battery pack.

Texas Instruments makes battery fuel gauges that are small enough to fit in the circuitry of a cell phone, and one of the most common ones that uses this technology are iPhone batteries. These batteries use a single-wire interface called HDQ (which stands for High-Speed Data Queue). It may sound similar to Dallas Semiconductors’ 1-Wire protocol, but the two are completely different and incompatible with each other.

Protocol details

The HDQ protocol can be emulated with a serial port and a little bit of external circuitry. The protocol can be emulated with a serial port at 57600 baud with 8 data bits, no parity bit and 2 stop bits. Because this is a bi-directional bus, an open-drain configuration is needed. Most TTL serial ports are not open-drain, so some circuitry is required to do this. TI’s application note suggests using a CMOS inverter and an N-channel MOSFET along with a 1 kOhm pull-up resistor, but this can be cut down with a 74HC07 open-drain buffer and pull-up resistor.

[EDIT: June 13, 2015 – Corrected schematic]

The HDQ protocol uses a short pulse to indicate a logic 1, with a longer pulse to indicate a logic 0. The data is sent LSB (least significant byte) first, with a 7-bit address and an eighth bit to indicate if the operation is a read or write (0 is read, 1 is write). If it is a read operation, the fuel gauge will respond with one byte of data. As you might think, this is a very slow means of communication; the typical bus speed is 5-7 kilobits per second, but the actual usable throughput will be less than this.

The hack in this is that the bit timing can be made by sending a specially crafted UART byte that meets the timing specifications. Each bit takes up one byte of UART buffer memory, with 24 bytes being enough to perform an HDQ read (the first 8 bytes are echoed back to the PC and need to be ignored by the software). TI’s application note goes into this with a bit more detail.

Windows HDQ utility

HDQ utility icon, in all its pixelated glory.

HDQ utility icon, in all its pixelated glory.

I have written a small Windows program that will read out the battery’s main data, identify as a certain iPhone battery model (most iPhone batteries are supported), and save a copy of this data to a text file for safekeeping. This program requires the National Instruments LabWindows/CVI Runtime library to run, since I whipped this program up with the first available IDE on my college PC.

fdd82eef8d

Screenshot of HDQ Utility version 0.96

The source code is not yet available (translation: I’m too ashamed of my programming skills to share it with others); however, a Windows executable is available for download below.

You will need to download the National Instruments LabWindows/CVI Runtime to run this program.

Current version (0.96): https://www.dropbox.com/s/pf0vszgfei7s8ly/HDQ%20Utility%200.96.zip?dl=0

Version 0.95: https://www.dropbox.com/s/7xdurbh9qibdftl/HDQ%20Utility%200.95.zip?dl=0
Version 0.9: https://www.dropbox.com/s/cd3esa5us6elfgr/HDQ%20Utility.zip?dl=0

Contributions are always accepted! Email me if you would like to send in a battery for me to analyze, or you can buy me a coffee through PayPal:


[EDIT – July 28, 2016] Welp, looks like the PayPal button’s broken (or was it never working to begin with…?). If you’d like to send anything to me, just give me a shout at ginbot86@gmail.com!

[EDIT – August 2, 2016] Whoops, looks like I never had the button working in the first place. Hopefully it works this time.

 

So Phone Me Maybe: A list of iPhone/iPad batteries with gas gauge functionality

Looking for my HDQ Utility to read out your own batteries? Click here!

UPDATE: Turns out the iPhone 3G and 3GS do have gas gauges! I will add them to my list as I find out more about them.

Each iPhone generation since the iPhone 4 iPhone 3G uses a TI gas gauge and uses the HDQ bus (iOS refers to this as the SWI [single-wire interface]) to communicate with the outside world. For more information about the HDQ protocol, click here.

I’ve noticed that many of the iPhone 5S and 5C batteries that can be purchased online are reusing iPhone 4 circuits, which will cause a significant decrease in gauge accuracy (proper parameters need to be programmed into the gas gauge, and that information is chemistry dependent), and the protection circuits in the iPhone 4 battery PCB will kick into overvoltage protection mode at 4.25 volts, less than the 4.3 volts that the iPhone 5 (and newer) batteries need to charge fully.

Because I have been unable to find a list of information of each battery generation, I’m making one myself. Because nobody else has dug this deep into the fuel gauges that the iPhone uses, I have to get this information experimentally (that is, by buying various batteries from online shops; the iPhone 5S battery has been very difficult to get, besides the fake ones I mentioned earlier).

So far I’m in need of an iPhone 3G (not the 3GS) battery, as well as all iPad batteries (or, if you have my program on hand, what model the battery is intended for, the fuel gauge device (eg. bq27541, bq27545), firmware version and designed capacity.

Model Gas Gauge Firmware Designed Capacity Default Unseal Key? Comments
iPhone 3G bq27541 ? ? Yes (0x36720414) Need to acquire one of these.
iPhone 3GS bq27541 1.17 1200 mAh Yes (0x36720414) Limited feature set. My utility will throw “No response” errors when reading this battery.
iPhone 4 bq27541 1.25 1420 mAh Yes (0x36720414)
iPhone 4S bq27541 1.35 1430 mAh Yes (0x36720414)
iPhone 5 bq27545 3.10 1430 mAh No (0x52695035) Many thanks to Yann B. for finding the unseal key!
iPhone 5S bq27545 3.10 1550 mAh No (0x84966864)
iPhone 5C bq27545 3.10 1500 mAh No (0x84966864)
iPhone 6 sn27545-A4 (note 4) 5.02 1751 mAh No (0x65441236)
iPhone 6 Plus sn27545-A4 (note 4) 5.02 2855 mAh No (0x18794977)
iPhone 6S sn27546-A5 (note 5) 6.01 1690 mAh No (0x90375994)
iPhone 6S Plus sn27546-A5 (note 5) 6.01 2725 mAh No (0x11022669)
iPhone SE Unrecognized (note 6) (A1141/0x1141) 1.03 1560 mAh No (unknown) (See note 6)
Apple Watch (38mm) sn27545-A4 5.02 235 mAh No (0x09130978)
Apple Watch (42mm) sn27545-A4 5.02 245 mAh No (unknown) If anyone has one that reads “FULL ACCESS” in my program, please send it to me! ๐Ÿ™‚
iPad (3rd gen) bq27541 1.35 11560 mAh Yes (0x36720414)

Notes:

  1. All known iPhone battery models use custom firmware, so not all of the features that the mainstream gas gauge models use are available. For example, none of these gauges will calculate the battery’s State of Health percentage (it is basically the percentage of the battery’s full charge capacity (it degrades with use) versus its designed capacity.
  2. The iPhone 5C’s battery label indicates a designed capacity of 1510 mAh, but the battery I’ve received indicates a capacity of 1550 mAh. As I have only been able to get one of these batteries that seem to be genuine, I will need to get more batteries of this type to confirm that this information is correct.
  3. The iPhone 5’s battery label indicates a designed capacity of 1440 mAh, but the fuel gauge reports 1430 mAh. The 5S battery reports 1550 mAh, but is labeled 1560 mAh. The 5C reports 1500 mAh, but is labeled 1510 mAh.
  4. The iPhone 6 and 6 Plus use a special firmware that is identified in TI’s battery software (except the very latest releases where such data was removed), and it has a very extensive feature set, and a lot of data logging features.
  5. The iPhone 6S/6S Plus use a firmware version similar to the iPhone 6/6 Plus, but with a newer chip and some features trimmed out. I’mย reasonably confident that the chip is an sn27546-A5 but have no idea if it’s the official part designator.
  6. The iPhone SE battery seems to have a unique custom chip, but has gone back to a DFN-based package (similar to bq27541) rather than a BGA like the bq27545/546. It is marked “A1141” and does not respond to my HDQ adapter, only the official TI EV2300/EV2400. I have only one in my possession, so I am not 100% sure whether this is true for this series of batteries.

Tearing down and modifying the Mars RPB60 power bank

A while ago I mentioned purchasing a very cheap battery pack that obviously didn’t live up to expectations. However, I didn’t get a chance to write about a more capable power bank, the Mars RPB60. It was branded as a SoundLogic/XT power bank, and holds 4400 mAh of battery capacity, with two USB ports (one labeled for 1 amp and 2.1 amps) and a micro-USB power input.

The power bank

The power bank from the outside looks pretty nondescript, with two USB ports, a micro-USB input, a button and four blue LEDs. Initially it seemed that there wouldn’t be any easy way to open up the casing without damaging it, so I tried to pry away the plastic covers at the ends. Doing so revealed plastic plates held in with small Phillips screws. Disassembly from that point was a cinch.

Removing covers reveals hidden screws

Removing covers reveals hidden screws

The PCB portion of the pack is of a stacked design. The two halves are connected with a set of small pin headers, with one side being the main DC-DC converter and USB output, and the other being the “gas gauge” and charging circuitry. The reason I put the phrase ‘gas gauge’ in quotes is that it’s only going by pure voltage thresholds, making it inaccurate when under load (like charging a phone and tablet, for example).

2014-01-05 00.25.02The main microcontroller is an unmarked 14-pin SOIC (likely an OTP-based PIC clone) and a TP4056 Li-Ion charging IC. The DC-DC converter is a DFN package that I couldn’t find any data on, but from what I can tell it integrates the DC-DC converter control circuitry and the switching MOSFETs.

2014-01-05 00.26.41

Blurry photo of microcontroller/charger PCB, taken with a potato for a camera. ๐Ÿ˜›

2014-01-05 00.27.16 Cells

The cells themselves seemed to be of good quality, but the bastards at Mars decided to black out the branding and model number of the cells! However, I was unphased at their attempt to cover up what cells they were using. With a careful cleaning with flux remover and some Kimwipes, I found that this pack uses ATL INR18650 cells with a DW01-based protection circuit. These cells hold 2200 mAh each, and the INR prefix means that these are high-power cells intended to provide heavy output currents. This is desirable as a 10 watt load like an iPad would definitely put heavy strain on the batteries. Considering the good cells they used, I don’t understand why they’d want to hide the markings on the cells (and in a half-assed way too!)…

The initial capacity was at about 4800 mAh (greater than the rated capacity ๐Ÿ˜€ ) with an average ESR of about 63 milliOhms. However, after a dozen charge-discharge cycles, the capacity has decreased to 4630 mAh and has an average internal resistance of 190 milliOhms. I’ve got a feeling charge cycle endurance may be an issue with this battery pack. Time will tell…

Gas gauge hacking

Since this battery pack didn’t have the gas gauge capabilities I wanted (voltage threshold-based gauging isn’t enough!), I decided to put in my own. I built a small bq27541-V200 gauge board with an external thermistor and current-sense resistor, using the breakout board itself to hold all the SMD passive components required for the gauge to function. The thermistor is taped to the cells to get an accurate temperature reading, and the current-sense resistor is attached in series between the cell’s negative terminal and the negative contact of the protection circuit.

Gas gauge chip added

Gas gauge chip added

This is where the hacking happens. I connected the I2C lines to the left USB port’s data lines. The voltage divider used for Apple devices is very high-resistance and makes for good I2C pullup resistors. The device still appears as a normal device charger but works just fine when the I2C signals are hidden behind the USB lines.

Quirks

However, the design for my gauge is definitely not the best one. I noticed that with heavy use (and not even one full discharge cycle), the gauge had reset 4 times without me knowing. Of course, I’m not expecting great performance from this gauge since it’s extremely susceptible to EMI (long wires looping around are just asking for trouble ๐Ÿ™‚ ) in its current state. Given how I basically hacked this together in a matter of a few hours, it works well enough. Next up is to go into Altium Designer and make a proper gas gauge board with good EMI and RFI mitigation (and perhaps sell them on Tindie; the hobbyist community needs better gas gauges and stop being so paranoid about Li-Ion batteries).

Further testing showed that certain phones put pulses on the USB lines which has occasionally caused the bq27541 to crash and reset as well.

Additionally, I’ve noticed that the DC-DC converter circuit is quite inefficient. It has 5-7 mA of quiescent current draw and has about 60-80% efficiency. At a full charge, it will take only one month before all the charge is drained from the cells and the protection circuit disconnects the cells.

Future plans

Since this battery pack has a nicely built casing, I intend to gut the battery pack, design new PCBs inside with good DC-DC conversion, an Impedance Track-based fuel gauge, and an onboard microcontroller with some battery-logging capabilities (perhaps to an EEPROM or an SPI Flash ROM), accessible through the micro-USB port. I also plan to use some higher-capacity cells, like the 3400 mAh Panasonic NCR18650B.

If not, then at least I want to replace the microcontroller with one that will read the bq27541’s state of charge readings and display them on the LED bar graph.

Looking inside an iPhone 4/4S battery

A classmate of mine had a couple broken iPhones that he ‘relieved’ of their batteries and let me take a look at them. Being the curious type I peeled away the outer layers of tape to reveal the protection circuit. I spotted a current sense resistor, andย  that got me thinking…

… can it be? Yes, I found a bq27541 fuel gauge chip inside the battery! After fooling around with the battery, I found out that the battery is using the HDQ interface.

iphone battery pinoutThe HDQ bus, which stands for ‘High-speed Data Queue’, is a single-wire communications bus used by TI fuel gauges. It’s similar to Maxim’s 1-Wire protocol but runs with different protocols and timing. It operates at 7 kilobits per second (so much for ‘high speed’ right? ๐Ÿ˜› ) and a refresh of the data memory in the TI software can take almost half a minute. However, it’s good enough for occasional polling (like every minute or so) since it’s unlikely that the gauge will be read from every second.

The bq27541(labeled BQ 7541) in the iPhone battery runs an unusual firmware version. It’s running version 1.35 and doesn’t match with any release on TI’s website. The gas gauge is sealed so initially it seems like gaining access to the Data Flash memory would be impossible. However, in non-Apple fashion, the gauge’s passwords are left at the default; 0x36720414 and 0xFFFFFFFF for the unseal and full-access keys, respectively (and it’s not the first time Apple’s done this!). Since the firmware version is unknown, I told bqEVSW to treat the chip as if it were the bq27541-V200. I then saved only the calibration, capacity, resistance and lifetime data.

Updating the firmware over HDQ was a nightmare. It took over a dozen tries for each of the two batteries I had, and the update process took 45 minutes (!) to update the bq27541 to the V200 firmware. At one point, it seemed as if I bricked the chip, but a power-on reset of the chip by shorting the cell very quickly ๐Ÿ˜€ sent the device into ROM mode (ie. firmware-update mode). From there I used bqCONFIG to update the firmware, and it was successful! Now I could use GaugeStudio to interface with the gauge rather than the unsightly bqEVSW software.

bq27541 updated to version 2.00

bq27541 updated to version 2.00

Given how long it took for me to update the firmware of the gauge, I have doubts that iPhones will update their batteries’ firmware in-system. Hell, the iPhone OS ignores the bq27541’s State of Charge readings and substitutes its own. Nice going, Apple!

Now to start going through cell phone recycling bins to pull out dead iPhone batteries for their gauges…