Cannot do flash read reliably (EUB-95)

[ammaree]

Checklist

• Checked the issue tracker for similar issues to ensure this is not a duplicate
• Read the documentation to confirm the issue is not addressed there and your configuration is set correctly
• Tested with the latest version to ensure the issue hasn't been fixed

How often does this bug occurs?

always

Expected behavior

I expect to be able to read the flash from the target device back back into a file but cannot.

Actual behavior (suspected bug)

Instead it starts and read <5% of a 4MB flash before failing

Error logs or terminal output

esptool.py v4.10.0
Found 2 serial ports
Serial port /dev/cu.usbmodem94A990397EE81
Connecting....
Detecting chip type... ESP32
Chip is ESP32-D0WDR2-V3 (revision v3.1)
Features: WiFi, BT, Dual Core, 240MHz, VRef calibration in efuse, Coding Scheme None
Crystal is 40MHz
MAC: 8c:4f:00:a5:0b:0c
Uploading stub...
Running stub...
Stub running...
Configuring flash size...
4096 (0 %)
A fatal error occurred: Corrupt data, expected 0x1000 bytes but received 0xf08 bytes

Steps to reproduce the behavior

Build latest version 1.2.0 and flash to latest hardware ESP-Prog-2
Connect ESP32 WROOM based target device with 4MB flash

execute correct command:
esptool.py read_flash 0 0x400000 output.bin

Have tried with --no-stub option; and
At speeds 115200, 230400, 256000, 512000, 460800 and 921600;
Fails at all speeds, with or without --no-stub

Commands write_flash works perfectly at all speeds up to 921600

Commands erase_flash, flash_id, flash_status and monitor works perfectly

Any suggestions?

Project release version

v1.2.0

System architecture

ARM 64-bit (Apple M1/M2, Raspberry Pi 4/5)

Operating system

MacOS

Operating system version

macOS Tahoe 26.1

Shell

ZSH

Additional context

No response

[Dzarda7]

Hello @ammaree, I tried running read-flash on MacOS (sequoia 15.6.1, do not have the same version as you), Linux (Fedora), with the same target board you use and was not able to reproduce the issue. Are you sure you are using latest ESP-USB-BRIDGE firmware? Is it possible for you to connect to UART debug pins of the ESP-Prog2 or use normal ESP32-S3 and try to look what is on the UART port? Or maybe if you can provide some info how I can reliably reproduce it, it would be highly appreciated.

[Dzarda7]

Okay, tested on Tahoe 26.0, also works.

[ammaree]

100% sure using latest updated v1.1.0 esp-usb-bridge repo.
All other commands work reliably, including write(flashing) at 921600.

Changed to a 40cm 10Gbps certified cable to ESP-Prog-2, 12.5cm long Alibaba "Dupont" wires to dev board. Now can read at 230400 but at 460800 fails consistently, sometimes immediately.

When using an ESP-Prog (FT2232) adapter with same Dupont wires and old MicroUSB Type-A cable it also works at 230400 reliably, nothing more.

Unable to connect a second ESP-Prog-2 as the target device, RX, TX, EN and IO0 pins not exposed.

So I can only think that the problem lies with the quality of Dupont cables?

Any suggestions?

[Dzarda7]

First of all latest version is v1.2.0, but I believe you meant that. Second, so you saying now that you can read flash at 230400, but higher baudrate fails right? It might be HW issue, but I can try to reproduce it and take a look if there is not any issue with the firmware (buffer overflow or something).

[ammaree]

Sorry, type, meant v1.2.0.

Considering exact same symptom with original ESP-Prog (with no firmware) I am confident the problem lies with the Dupont wires, problem now is to find something genuine quality.

[Dzarda7]

Yes, I believe it might be HW issue which cannot be resolved by us. If you are okay, I will check it and close the issue if I won't find any SW issue with that.

[Dzarda7]

Okay, I believe this is the HW issue with high baudrate. So I am closing this.