QUERY · ISSUE

ESP8266: PWM glitches when doing DHT measure

openby EddieParisopened 2021-02-21updated 2022-03-30

port-esp8266

When using a PWM channel in parallel of a DHT measure, there can be some glitches on the PWM channel at the measure time. This is because the DHT drivers locks interrupts for accurate protocol timing handling, and the PWM driver relies on these interrupts.

CANDIDATE · PULL REQUEST

ESP8266: Always run pwm timer to avoid glitches from full pwm to any …

closedby EddieParisopened 2018-02-23updated 2018-03-05

proposed resolution of #3631

Was simpler that I thought, just always run the timer even if no "active" PWM (non 0 or non full) is present. The glitch after full PWM disapeared, no cross effect on other channel introduced.

Assessment

Q Qwen · medium

Both items address PWM glitches on ESP8266, but they describe different root causes and solutions. Issue #6945 describes glitches caused by the DHT driver disabling interrupts during measurement, which disrupts PWM's interrupt-driven operation. PR #3633 addresses a different issue (#3631) where PWM glitches occur when switching between duty cycles (specifically when no active PWM is present), by keeping the timer always running. While both involve PWM glitches on ESP8266, the underlying mechanisms are distinct: one is about interrupt masking from DHT, the other is about timer state management. PR #3633 does not resolve the DHT-related interrupt issue described in #6945.

Suggested action

Keep both items open as they address different problems. The PR may provide some context for PWM stability but does not solve the DHT interrupt masking issue.

ESP8266: PWM glitches when doing DHT measure

ESP8266: Always run pwm timer to avoid glitches from full pwm to any …

Keyboard