At this point I read that ESP-03's from factory don't respond to AT commands, so I proceeded to flash the latest firmware from Electrodragon (https://drive.google.com/folderview?id= ... sp=sharing). First, I additionally tied GPIO0 to GND, then used esptool. With esptool I could read_mac and get flash_id, but only when GPIO0 was grounded.
I flashed like this:
#./esptool.py write_flash 0x000000 New-AI-v0.9.5.0_AT_Firmware.bin
After flashing, and removing GPIO0 to GND link, I could see a new access point come up. The name changed slightly from before. As before, I can see bootloader output on booting at 75k baud, but I no longer see anything from the rom program. Or maybe the rom program isn't running, but then I do see the new access point, so probably it is running. Needless to say I still can't communicate over UART with the module, in fact it's even worse that before
Currently I am working on building a firmware from source using the IoT SDK to see if that will change anything...
esp03madness wrote:I read that ESP-03's from factory don't respond to AT commands
Well, mine were already flashed with an AT firmware, I just had to choose the correct terminal settings: 9600, 8, 1, N, CR+LF. I wasn't much interested in using those commands though, so I soon flashed a NodeMCU Lua firmware, using NodeMCU's flasher, and it went seamlessly. You might want to try that path to test your setup before going the hassle of building a new firmware.
You don't mention power supply which is a frequent cause of failure. Check that yours can provide enough current. Use an external one if you can, and connect the serial adapter to Tx, Rx, and ground only.
I also found a pinout (on Pete Scargill's blog AFAIR) which I have slightly corrected:
...and I also wrote down the following notes as a foolproof setup:
To bring ESP-03 into firmware update mode from power down without having to wiggle the reset pin:
Connect GPIO0 and GPIO15 to GND, CH_PD to VCC before powering up the module. (GPIO0 can then be left floating or up.)
Then launch tool to flash a .bin file.
I remember trouble when CH_PD was pulled directly up, and adding a resistor in the 1~10k range solved it.
I hope that'll help.