ADC1 operation affecting GPIO27 - chip issue?

[yugguy]

I've got GPIO27 configured as a digital input (with both an external pull-up and internal pull-up enabled), but when I perform ADC1 conversion operations, IO27 glitches 'low' for about 6us. With a 10K pull-up, the glitch is about 1V (drops to ~2.3V). With a weaker pull-up, the glitch is low enough to hit a logic low threshold and I get a spurious interrupt.

No other pins are affected. Has anyone observed interaction between ADC and GPIO27? It feels like a chip bug. U can use a strong pull-up as a work-around, but would like to understand what's going on.

[Darmann]

I found the same adc1 side effect on gpi39 and gpi36. I think there is a relationship between adc1 operation and the hall effect sensor.
In my program I do adc1_ch0 read (20ms), and every time I will get gpio interrupt on gpi39 and gpi36. My solution: move gpio interrupts away from gpi39 and gpi36.
In normal polling mode, hopefully there will be no side effect.

[WiFive]

36&39 is a known issue, check eco workaround PDF. Not sure if anyone else has reported 27.

[ESP_Sprite]

Wifive: Yep, conformed by the hardware guys, but GPIO27 is also new to them. If someone runs into this, please post the source code so we can double-check this.

[kuhatovuk]

Hello, same thing with ADC1_CHANNEL_2 and IO25 internally pulled up. ADC reads trigger its LOW edge interrupt.