Replies: 8 comments 36 replies
-
The |
Beta Was this translation helpful? Give feedback.
-
There are few more quirks in the device. When you unplug and plug the device, it does not seem to complete the boot the first time. Second unplug/plug action boots the device but the display does not get turned on. When you execute a restart from GUI or command line then the display gets initialized and turns on. I wonder what is going on. I cannot see any serial port output either. |
Beta Was this translation helpful? Give feedback.
-
Well for a start I managed to come up with this to be able to monitor my off grid solar system and EV charging status. Next I will try to add a Buzzer for alarms and possibly a proximity sensor so that display gets turned on only when somebody is near. Probably the ESP32 based device will be useful as the next device to make the PV generation graphs etc on this small screen. Still need to figure out the weird initialization issue of the display. ie. When powered up display does not get initialized. Once wifi is connected and restart command is given GUI/console the display gets initialized. |
Beta Was this translation helpful? Give feedback.
-
I need a solution for the boot/initialization issue. Here are the symptoms:
The log looks like this:
The log messages are same except for the following message which does not appear after doing
Also the CFG log entries are different F7 vs F6. I do not know what it means.
What could be getting wrong here? I am not sure whether Display Descriptor has anything to do with it but then it works on the reboot.
|
Beta Was this translation helpful? Give feedback.
-
Weblog 4 shows that SPI message comes before the delay starts
|
Beta Was this translation helpful? Give feedback.
-
In the original firmware, the screen comes up immediately when power is connected. Then it shows a progress bar for few seconds and comes to the weather display. I guess there must be a significant delay between screen initialization and WiFi attempt. I had a second identical device and decided to flash that unit also and check. The results were identical. Since I have minimal config on the second unit, I can clearly see timing wise that during
During power up also I can feel that there is a pause at the same message but hard to compare the timings except from the local time stamps from the terminal program that I use (Termite). You can about two seconds between the
During power up also there is a pause visibly but the timestamp does not get printed by the terminal program. However if you go by the timestamp printed by Tasmota the delay seems to be comparable:
Is there a way to power up without WiFi? |
Beta Was this translation helpful? Give feedback.
-
I have this Display module from Aliexpress. Is this the same kind that you have? I was thinking about connecting this to a Wemos D1 Mini to see whether same error occurs. |
Beta Was this translation helpful? Give feedback.
-
I got the "pro" ESP32 version of the device working, based on the info in this thread, and also Spotted 3 Github projects about the device, including from the original creator: Like posted earlier here, I used the ST7789 display descriptor sans the touch part, and this template:
The ESP32 version is sensitive to touching the top of the box, and Doing a rudimentary RGB color test worked, after the invert: With the OG variant of the ESP32 and no PSRAM, there are strict limitations to how much headroom for fun, but a lot more than the ESP8266 version. I had no trouble running HASPmota with a small Berry program to plot the touch input levels on screen, without the heap shrinking to a catastrophic level. Arguably, the price is a lot higher than the hardware cost, as the device is pretty simple. It makes very much sense to not have a temperature sensor inside the box, which gets noticeably warmer in use. As the version I got also has the RTC chip position unpopulated, it does not make much sense with words like "weather clock" in the name, but who cares.... The creator was agreeable about making a version with ESP32-S3 and PSRAM, but you could also replace the ESP32-WROOM-32E module with a similar ESP32-S3 based one, meaning that the product price would then be for the plastic box, a small display and yet another USB-C cable. Of course, you could also stick with the standard firmware, if you like that solution where the features are not "too bad". |
Beta Was this translation helpful? Give feedback.
-
I purchased a GeekMagic SmallTV look alike device from Aliexpress:
https://vi.aliexpress.com/item/1005006895576912.html
It also has ESP8266 like the original SmallTV device but the PCB is different. It has a different PCB.
There is a built in CH340 USB-TTL chip also. So flashing this device is a breeze.
Here is what I have done so far:
Flashed Tasmota 14.1.0 display.bin
Configured the GPIO's based on following discussion about GeekMagic SmallTV:
https://community.home-assistant.io/t/installing-esphome-on-geekmagic-smart-weather-clock-smalltv-pro/618029/3?filter=summary
UPDATE 2024/07/23 : There was an issue with Display Initialization during Power Up due to not setting
GPIO 2
asDisplay Rst
. See posts below for investigation of the issueUPDATE 2024/07/25: Controlling brightness was not possible with GPIO Setting
Brightness
and commandDisplayDimmer
was not possible. Solution was to set GPIO5 toPWM_i
and useDimmer
commandhttps://github.com/arendst/Tasmota/blob/development/tasmota/displaydesc/ST7789_display.ini
Rule3 :H,ST7789,240,240,16,SPI,1,*,*,*,*,*,*,*,40 :S,2,1,3,0,80,30 :I 01,A0 11,A0 3A,81,55 36,81,00 21,80 13,80 29,A0 :o,28 :O,29 :A,2A,2B,2C :R,36 :0,C0,00,50,00 :1,A0,50,00,01 :2,00,00,00,02 :3,60,00,00,03 :i,20,21 #
UPDATE 2024/07/25: Controlling brightness was not possible with GPIO Setting
Brightness
and commandDisplayDimmer
was not possible. Solution was to set GPIO5 toPWM_i
and useDimmer
commandObservation:
Displaymode 1 showed the Time Stamp but whole screen gets utilized upside down, meaning that I have to keep the device upside down for text to appear correctly.
Definitely my Display Descriptor may have a flaw.
Need help to get it corrected.
Here is the device descriptor from the original file.
Beta Was this translation helpful? Give feedback.
All reactions