ESP32-Paxcounter
ESP32-Paxcounter copied to clipboard
cyberman54
Hardware support for T-Beam V1.2 with AXP2101
Please add support for the new version of the Lilygo T-Beam V1.2 with the new power management chip AXP2101.
Hi, thanks for your reply! Here are the links: https://www.amazon.de/LILYGO-Entwicklungsboard-Meshtastic-Mainboards-Soldered/dp/B0B454TB6S https://github.com/Xinyuan-LilyGO/LilyGO-T-Beam https://github.com/Xinyuan-LilyGO/LilyGo-LoRa-Series AXP2101 chip: https://github.com/meshtastic/firmware/blob/7bb281d5c569870c27cdd65bcd14507148e4ca9b/src/Power.cpp#L516
In all your links, i can't find a schematic or specs for v1.2, it all renders as v1.1?
Yes, they hid this fact pretty well, unfortunately. By changing the PMU from AXP192 to AXP2102 they have changed the versioning from 1.1 to 1.2. Please see the screenshot.
@XAct-6510 i checked the Xpowerslib example, and found that there is some refactoring of power.cpp needed. Could you do this, and make a PR?
Unfortunately I'm not a programmer, I'm a networking guy... But I can provide you one of my 5 T-Beam V1.2 if that helps.
@XAct-6510 not necessary, because if i would code it, you could test it on your board. But currently i have no time to code this.
As mentioned I'm not a programmer, but maybe this can be used? https://github.com/meshtastic/firmware/blob/master/src/Power.cpp The T-Beam V1.2 comes preinstalled with the Meshtastic firmware, which obviously works with the new hardware revision.
i saw that, but would stick to the example code of XPowersLib, because Meshtastic has more complexity. But again, you or someone else would be welcome to show up with a completely refactored power.cpp
Just for completeness, the schematics of the 1.2 version of the T-beam: https://github.com/Xinyuan-LilyGO/LilyGo-LoRa-Series/tree/master/schematic
Also they split off the older boards, so here are all the LilyGo LoRa series schematics: https://github.com/Xinyuan-LilyGO/LilyGo-LoRa-Series/tree/v0.0.1/schematic
I am looking into this. First I am ironing out some issues with AXP192 (my PRs have just been merged) and then I will look into the AXP2101.
I have both versions here on my desk, but since those T-beam boards are quite expensive these days, I wanted to make sure I knew how it all should work.
I am looking into this. First I am ironing out some issues with AXP192 (my PRs have just been merged) and then I will look into the AXP2101.
I have both versions here on my desk, but since those T-beam boards are quite expensive these days, I wanted to make sure I knew how it all should work.
@TD-er I could do some testing. Received a T-Beam "Supreme" with AXP2101 this week. I finalized a hal file with matching pinout already, build paxcounter code for S3 and flashed it successfully on the T-Supreme. The board seems to be alive, because i can see the USB port triggers on reset. But nothing else happenes. Probably becaue AXP2101 is missing proper initialization, thus the board components are probably not powered.
I have been reading into the existing libraries. There is even an example of how to make it runtime detectable. This does require some changes like keeping a pointer to the base class. See: https://github.com/lewisxhe/XPowersLib/blob/master/examples/XPowersLibInterface_Example/XPowersLibInterface_Example.ino
Still I'm puzzling a bit on how to make this as fool proof as possible as wrong settings for power chips like these can for sure ruin your device or stuff connected to it.
The last 2 weeks have been extremely unproductive for me, as apparently Covid is still lurking around and even though I've had it 3 times already, it still hits you every time. At least I'm already feeling better now, so I will have another look at this code.
Support for T-Beam v1.2 is now implemented in branch feature-asp2101
Does anyone who reads this own this board, and could test, please? Thanks. /cc @XAct-6510
For those willing to test, please double or triple check these lines first: https://github.com/cyberman54/ESP32-Paxcounter/commit/33486cc4ad51ce7a0e6ddeda970e1356fbacd842 I only can test it later this evening.
Can't get it to work yet - after flashing the new build the device hangs on boot:
rst:0x1 (POWERON_RESET),boot:0x12 (SPI_FAST_FLASH_BOOT) configsip: 0, SPIWP:0xee clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00 mode:DIO, clock div:2 load:0x3fff0030,len:1184 load:0x40078000,len:13232 load:0x40080400,len:3028 entry 0x400805e4 I (17) src/configmanager.cpp: Loading device configuration from NVRAM... I (20) src/configmanager.cpp: Runtime configuration v3.6.2 loaded I (21) src/main.cpp: Starting paxcounter_bb0aff61 v3.6.2 (runmode=0 / restarts=0 ) I (26) src/main.cpp: code build date: 1719913557 I (31) src/main.cpp: This is ESP32 chip with 2 CPU cores, WiFi/BT/BLE, silicon r evision 3, 4MB external Flash I (40) src/main.cpp: Internal Total heap 225948, internal Free Heap 198120 I (47) src/main.cpp: ChipRevision 3, Cpu Freq 240, SDK Version v4.4.5 I (53) src/main.cpp: Flash Size 4194304, Flash Speed 40000000 I (59) src/main.cpp: Wifi/BT software coexist version 1.2.0 I (64) src/main.cpp: Wifi STA MAC: B0:B2:1C:5F:59:24 I (68) src/main.cpp: IBM LMIC version 1.6.1468577746 I (73) src/main.cpp: Arduino LMIC version 4.2.0.1 I (78) src/lorawan.cpp: DevEUI: 70B3D57ED00657F9 I (82) src/main.cpp: TinyGPS+ version 1.1.0 I (86) C:/Users/GIS-Team/.platformio/packages/[email protected] 11.230801/cores/esp32/esp32-hal-i2c.c: Initialising I2C Master: sda=21 scl=22 fr eq=100000 I (102) .pio/libdeps/usb/XPowersLib/src/XPowersCommon.tpp: SDA:21 SCL:22 W (107) C:/Users/GIS-Team/.platformio/packages/[email protected] 011.230801/libraries/Wire/src/Wire.cpp: Bus already started in Master Mode. E (121) C:/Users/GIS-Team/.platformio/packages/[email protected] 011.230801/libraries/Wire/src/Wire.cpp: NULL TX buffer pointer E (133) C:/Users/GIS-Team/.platformio/packages/[email protected] 011.230801/libraries/Wire/src/Wire.cpp: NULL TX buffer pointer I (146) .pio/libdeps/usb/XPowersLib/src/XPowersAXP2101.tpp: ~XPowersAXP2101
Looks like the i2c bus did not initialize for some reason. Thus, there is no communication with the AXP2101 chip, so it does not power up all board circuits.
I can reproduce this with a T-Beam with AXP192 chip and single i2c bus.
The reason is that there is a conflict in power.cpp by the already opened i2c bus by i2c_init() from main.cpp. The code needs some refactoring here, e.g. using a pointer to the i2c object.
i committed a hot fix, it should work (please re-test), but yet to be discussed if this is a straight forward approach for boards with two i2c buses. /cc @XAct-6510
I have applied the hotfix and it is now working. I'm still running some tests, but it seems to be running fine.