Sudden battery drain

[airmaxx]

I'm experiencing a sudden battery drain, after 1-2 weeks of running just fine. Any idea what could be the cause, or how I can debug it?

[lanrat]

What does your battery voltage graph look like?

The battery % is calculated based on the voltage, and every battery will be a little different. I calibrated this for the batteries I had for testing, but yours may have a different charge/discharge curve.

[airmaxx]

Here's the voltage graph for the same time period.

[elratt0r]

Could you also show the boots for the same timerange?

Assumptions: a) the esp is booting more often than the assumed sleep interval b) the battery you're using has a very "non-linear" discharge curve

About the battery: I am using a LiPo, 3.7V, 1200mAh (from Eckstein, ZB07002) and that gives the following discharge curve. You see a steep(er) drop at around 3.45V, that's more or less the cut-off voltage for my battery.

[airmaxx]

How can I see the boots? Are they published via MQTT to Home Assistant?

My battery is specified as Rated voltage: 3.7V, capacity: 3000mAh, 11.1Wh, it shouldn't start discharging so fast at around 4.0V.

I'll keep an eye on it, it's currently approaching 4V again (after about 10 days of use).

[elratt0r]

Okay, with that specs it's really odd that it empties that rapid.

For comparsion: Mine usually lasts ~1 month, with sleep cycle set to 20 minutes and a weekday speed up between 6 and 9 o'clock to 5 min intervals.

Yes, it's published via MQTT under homeassistant/sensor/<hostname>/boot/state -> {"boot": <value>}

You should also be able to see it in HA:

[klopyrev]

I want to report that I'm seeing the exact same issue. It runs fine for a while. At some point, several days or weeks later, the battery suddenly drains.

I have an identical Inkplate 10 in another room running an older version of the code, and I see no such problem.

[airmaxx]

Interesting that it happens on different levels. At the moment I'm waiting to see the next event. Did you check the temperature at the moment of the battery discharge? It was higher at the time of the battery discharge.

[elratt0r]

@klopyrev Could you provide the commit hash the old version is running on?

@airmaxx Good catch! My assumption would be, that the ESP32 is not going to sleep (anymore)

Is any of you using the feature to modify the sleep times via MQTT? (homeplate/homeplate/activity/run -> {"action": "hass", "refresh": "300"})

[klopyrev]

@elratt0r I checked the version date printed in the main.cpp file, and I think both my devices have been running the new version (I forgot I reflashed the older device). So, good chance it's not new code.

I just reflashed both devices again to make sure. I'll monitor to see what happens.

[hillbillypolenta]

Also experiencing this. Looking at a few cycles wild guess there may be an issue with the voltage calibration as a pretty typical LiPo discharge curve can be observed over the span of (in my use case) 1 month before it drops off a cliff at a reported 3.8-3.9 volts or so. After recovering it reported 3.2-ish volts.

Unfortunately I was not logging boot reason but I have turned all of those on now and will see if there's anything to report there (with reference to the configurable sleep duration feature which I am using)

Using the 3000mAh battery included with the Inkplate 10

[lanrat]

I know this may be difficult, but getting the device's serial logs while it tries to go into sleep would help. However this would need to be done while not providing any power over USB in order to not affect the battery readings. This could be done by using a USB cable that has the VCC line cut, or an external UART connected directly to the UART pins on the ESP32.

[klopyrev]

This is from using an external UART adapter connected to the ESP32 TX / RX pins.

I can leave this setup connected to my PC until the issue reoccurs.

13:19:33.884 > ets Jul 29 2019 12:21:46 13:19:33.885 > 13:19:33.885 > rst:0x1 (POWERON_RESET),boot:0x12 (SPI_FAST_FLASH_BOOT) 13:19:33.885 > configsip: 0, SPIWP:0xee 13:19:33.885 > clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00 13:19:33.885 > mode:DIO, clock div:2 13:19:33.885 > load:0x3fff0030,len:1184 13:19:33.885 > load:0x40078000,len:13192 13:19:33.885 > load:0x40080400,len:3028 13:19:33.885 > entry 0x400805e4 13:19:34.889 > 13:19:34.889 > 13:19:34.889 > [SETUP] starting, version(Nov 20 2024, 08:47:13) boot: 0 13:19:34.931 > ESP Chip information: 13:19:34.931 > This is esp32 chip with 2 CPU core(s), WiFi/BT/BLE, silicon revision 3, 4MB external flash 13:19:34.931 > Minimum free heap size: 4415055 bytes 13:19:34.931 > Heap summary for capabilities 0x00000C06: 13:19:34.931 > Totals: 13:19:34.931 > free 0 allocated 0 min_free 0 largest_free_block 0 13:19:34.931 > [ACTIVITY] startActivity(1) put into queue 13:19:35.042 > [STATUS] Boot 1 normal 13:19:37.163 > [SETUP] Battery: 100% (4.24v) 13:19:37.741 > [SETUP] starting button task 13:19:37.741 > [INPUT] Input monitoring not supported by this platform 13:19:37.741 > [SETUP] starting time task 13:19:37.741 > [TIME] Internal Clock and RTC differ by -1732126773 seconds. local(3) RTC(1732126776) 13:19:37.741 > [TIME] local time (3) 19:00:03 Wednesday, December 31 1969 13:19:37.741 > [TIME] ERROR: RTC time is too far in past. RTC likely has wrong value! 13:19:37.741 > [SETUP] starting WiFi task 13:19:37.741 > [WIFI] Connecting... 13:19:37.741 > [SETUP] starting OTA task 13:19:37.741 > [SETUP] starting MQTT task 13:19:37.741 > [MQTT] starting MQTT 13:19:37.741 > [SETUP] starting sleep task 13:19:37.741 > [SETUP] starting activities task 13:19:37.741 > [ACTIVITY] starting activity: 1 13:19:37.741 > [SLEEP] delaying sleep for 15 seconds 13:19:37.852 > [WIFI] Connected to AP successfully! 13:19:37.853 > [WIFI] IP address: 192.168.4.38 13:19:37.868 > [STATUS] WiFi connected 13:19:37.932 > [WIFI] Connected: 192.168.4.38 13:19:37.964 > [TIME] Syncing RTC to NTP 13:19:37.993 > [MQTT] Connecting to MQTT... 13:19:37.993 > [OTA] OTA ready 13:19:37.993 > [SLEEP] delaying sleep for 20 seconds 13:19:37.993 > [STATUS] Downloading image... 13:19:38.025 > [MQTT] Connected to MQTT. 13:19:38.025 > [MQTT] Session present: 0 13:19:38.025 > [MQTT] Subscribing at QoS 2, packetId: 1 13:19:38.025 > [MQTT] Sending MQTT Config 13:19:38.025 > [MQTT] Subscribe acknowledged: packetId: 1 qos: 2 13:19:38.072 > [MQTT] Connected 13:19:39.931 > [IMAGE] Downloading image: http://192.168.4.3:5000/2/ 13:19:39.931 > [ 5506][E][WiFiClient.cpp:313] setSocketOption(): fail on -1, errno: 9, "Bad file number" 13:19:39.931 > [TIME] Internal clock was adjusted by 1732126776 seconds 13:19:39.931 > [TIME] Internal RTC was adjusted by 1 seconds 13:19:39.931 > [TIME] NTP UNIX time Epoch(1732126779) 13:19:39.931 > [TIME] Timezone offset: (-18000) -5 hours 13:19:39.931 > [TIME] synced local UNIX time Epoch(1732126779) 13:19:39 Wednesday, November 20 2024 13:19:40.265 > [IMAGE] Download done 13:19:40.282 > [STATUS] Rendering image... 13:19:45.257 > [IMAGE] Image render ready 13:19:45.258 > [IMAGE] displaying.... 13:19:47.283 > [IMAGE] displaying done. 13:19:47.740 > [MQTT] sending status update 13:19:47.741 > [MQTT] Sending MQTT State: [homeassistant/sensor/homeplate2/boot/state] {"boot":1,"activity_count":1,"boot_reason":"normal","sleep_duration":1200} 13:19:47.932 > [MQTT] Sending MQTT State: [homeassistant/sensor/homeplate2/wifi_signal/state] {"signal":-45} 13:19:47.948 > [MQTT] Sending MQTT State: [homeassistant/sensor/homeplate2/temperature/state] {"temperature":24} 13:19:47.981 > [MQTT] Sending MQTT State: [homeassistant/sensor/homeplate2/battery/state] {"voltage":4.2,"battery":100} 13:19:58.752 > [ACTIVITY] startActivity(0) put into queue 13:19:58.752 > [ACTIVITY] starting activity: 0 13:19:58.752 > [SLEEP] prepping for sleep 13:19:58.752 > [MQTT] Stopping and disconnecting... 13:19:58.752 > [MQTT] Disconnected from MQTT. 13:19:58.752 > [WIFI] Stopping and disconnecting... 13:19:58.752 > [WIFI] Disconnected from WiFi access point 13:19:58.752 > [WIFI] WiFi lost connection. Reason: 8 13:19:58.895 > [SLEEP] entering sleep for 1200 seconds (20 min)

[lanrat]

@klopyrev Thanks for the logs.

It looks like its reporting the battery is at 100% when you took these. is that is correct?

Additionally, do you know if it stayed in sleep mode for the 20 minutes? or did it wake early for anything?

[klopyrev]

Yeah, the device was fully charged at the beginning.

Here are the logs from a few hours: https://pastebin.com/fADMkRPj

[lanrat]

I don't see anything out or the ordinary in the logs. And your reported battery drain looks normal too.

[airmaxx]

It happend again. This time i logged "boot count", yet it does not show anything. Sleep duration is set in config.cpp - I do not use MQTT for it.

[lanrat]

@airmaxx , it may be unrelated to the battery drain issue, but it looks like your homeplate crashed towards the end of the boot count graph.

I'm thinking on how to go about debugging this.

[klopyrev]

I'm still running with my inkplate connected to my PC with the external UART adapter. I'll keep it running until the issue happens. It make take some time. 2 of the times the drain happened it was after about a week. 1 of the times it was after about a month.

[airmaxx]

it looks like your homeplate crashed towards the end of the

I am not 100% sure about this - I pressed the button to (re?)start Homeplate at that time because I thought it was already crashed - but it seems to have continued to run while losing a lot of battery. I have now restarted it with only 35% battery to see if the behaviour is related to battery level/voltage.

[lanrat]

From a fresh boot, the boot count counter starts at 0, and is incremented every time the homepalte wakes from sleep. It only resets to 0 when doing a fresh boot (not waking from sleep).

[elratt0r]

Could you provide your sanitized config.h and also config.cpp?

More confusing I find, that the sudden battery drain does not seem to correlate with the changed sleep interval, neither with a raise in boot cycles.

[klopyrev]

My Inkplate which I had connected to the PC for the last several weeks finally reached the point where the battery suddenly drained. Unfortunately, the only thing that was saved in the logs was a very long list of null symbols, more than 20000 of them. I'm going to try again, but this time save everything into a file rather than using the history of the console.

[klopyrev]

I got the logs captured this time: https://public.klopyrev.net/minicom.cap

[elratt0r]

I got the logs captured this time: https://public.klopyrev.net/minicom.cap

27939:[2025-01-04 15:34:20] CORRUPT HEAP: Bad head at 0x3ffdf970. Expected 0xabba1234 got 0x3ffdab48
* First crash, but recovered
31908:[2025-01-05 11:10:29] CORRUPT HEAP: Bad head at 0x3ffdf970. Expected 0xabba1234 got 0x3ffdab48
* Second crash, but recovered
35118:[2025-01-06 03:03:07] [  2779][E][WiFiClient.cpp:313] setSocketOption(): fail on -1, errno: 9, "Bad file number"
* Next crash, probably while downloading the png
35243:[2025-01-06 03:23:19] [  2817][E][WiFiClient.cpp:313] setSocketOption(): fail on -1, errno: 9, "Bad file number"
* From here on it seems to be missing WiFi or a deadlock in the starting activity:
35351:[2025-01-06 03:43:31] [ACTIVITY] starting activity: 1
35352:[2025-01-06 03:43:31] [SLEEP] delaying sleep for 15 seconds
35353:[2025-01-06 03:43:31] ets Jul 29 2019 12:21:46
35354:[2025-01-06 03:43:31] 
35355:[2025-01-06 03:43:31] rst:0x1 (POWERON_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
...
35590:[2025-01-06 03:43:46] [SETUP] Battery: 16% (3.53v)
...
35603:[2025-01-06 03:43:46] [SETUP] starting activities task
35604:[2025-01-06 03:43:46] [ACTIVITY] starting activity: 1
* what the hell did the CPU did in this 90mins???
35605:[2025-01-06 05:11:43] ets Jul 29 2019 12:21:46
...
35626:[2025-01-06 05:11:45] [SETUP] Battery: 0% (3.31v)

Well, odd for sure. The crashes when downloading the PNG I also encountered with mine, usually when screenshotter and plate collide in file access (rare but happened).

Thank you very much for your effort! I'll see if I can at least track down the heap problems.

[huggenknubbel]

i have the same problem. but i don't know how to get the logs. One thing i noticed: maybe it has todo with wifi connectivity? beceause i faced this issue much more if i set the IP to DHCP instead to a fixed one.

[lanrat]

I just updated the base Inkplate library from 10.0 to 10.1. I don't know if this will have an effect, but I'm monitoring my Inkplate to see.

[lanrat]

Since there have been no updates on this in a while I'm going to close this issue. Please feel free to comment if its still occurring on the latest code and I'll reopen and revisit.