Page 1 of 1

Garbage on only first lines of booloader output

Posted: Wed Oct 23, 2019 4:41 am
by commando_j
Has anyone seen this before? It only happens on one of my two ESP32 boards. I don't think it was happening before, at least I never noticed it (I'm sure I would have) Is this a partially corrupted bootloader? (Ignore the date codes that were generated). If it was a baud rate issue then you wouldn't see the partial June date for example. Is there a way to fix this?

Code: Select all

[10-23 00:22:54.8]ets Jun  ’‚Š6 00:22§ÓSH¨
[10-23 00:22:55.2]áets Jun  ’‚Š² 00:22:57
[10-23 00:22:55.2]C!®.—‚ÂÅ0 (RTCÕ%E5}IMQ¥,¬ë‹é0x13 (SPIeST_FLASH_Bꊕj
[10-23 00:22:55.2]configsié 0, SPIWP:•e
[10-23 00:22:55.2]clk_drv:0Á0,q_drv:0x00‰}drv:0x00,csú‘rv:0x00,hd_d.—‚ÂÁ0,wp_drv:ÂÁ0
[10-23 00:22:55.2]mode:DIO,±½ck div:2
[10-23 00:22:55.2]ë+‘é0x3fff0018‰«¹é4
[10-23 00:22:55.2]load'ÂÍfff001c,«¹:6344
[10-23 00:22:55.2]load:ÂÑ0078000,len:LL“¢
[10-23 00:22:55.2]load:0x40‚¢‚0,len:6648C¡–‹Éå0x4008076SHøI (29) boot: ESP-IDF v3.3 2nd stage bootloader
[10-23 00:22:55.2]I (29) boot: compile time 14:46:47
[10-23 00:22:55.2]I (29) boot: Enabling RNG early entropy source...
[10-23 00:22:55.2]I (33) boot: SPI Speed      : 40MHz
[10-23 00:22:55.2]I (37) boot: SPI Mode       : DIO
[10-23 00:22:55.2]I (41) boot: SPI Flash Size : 4MB

Re: Garbage on only first lines of booloader output

Posted: Thu Oct 24, 2019 1:11 am
by commando_j
Well I can see how I may have caused the issue. With old age comes bad eyes. The GND pin that I thought I connected my logic analyzer to was actually the CMD pin of the ESP32 flash on my dev kit. Now I know why I was seeing crazy garbage on my SPI pins!
SPI is better now, but I still have the garbage as shown above.
From what I read, this is output from the 1st bootloader in ROM, not the external flash. Is this true? If so, I assume it can't be fixed.

Thanks.

Re: Garbage on only first lines of booloader output

Posted: Thu Oct 24, 2019 2:27 am
by WiFive
If you have a logic analyzer, check the bit width, the baud rate is probably slightly off, possibly due to unstable Crystal startup and frequency estimation.