Use this forum to chat about hardware specific topics for the ESP8266 (peripherals, memory, clocks, JTAG, programming)

User avatar
By ArnieO
#22340
iothing wrote:Hey, I can now also add to the not working list: ferrite core around the leads to the esp module.

The problem with inductive loads is not when they run, but if they are powered off. My module always reboots when the relay switches off. To reduce EMI I put an snubber on the relais C and NO terminals so that the peak at the switchoff is reduced. I use a X2 capacitor .23uF in Series with a 56Ohm 3W resistor. This mitigates the problem with crackling/popping sounds in the stereo and relay contact burnoff/welding together. Moving the esp away did not help because it is coupled to the relais board by the wires - I guess.

Actually cutting the RST trace helped most so far, but the problem is: if the module crashes, it takes ages to reboot and as the rst is not there anymore I cannot reboot it via the arduino (heartbeat check). Also how can RST be pulled up when the trace is cut? Someone wrote there is an internal pullup - is this really the case?! What pins have internal pullups on the esp01? Can the pullup be controlled by the software?


Yes, this looks/sounds like conducted EMI, and there are strong indications that the RST line has an unfortunate layout on the module, making it sensitive.

Some very experienced users here have reported not having seen these problems at all, while others report very persistent problems. So maybe there are production batch differences? One thing that it could be interesting to check is whether the internal pullup on the RST line is sufficiently strong.

A hypothesis could be that it is too weak, and thereby causes the problem if the module layout is poor. But in that case, hardwiring RST to VCC shuold help; apparently it does not.

The ESPs are reportedly very sensitive to Vcc stability. Have you looked over your Vcc/GND layout to eliminate any ground loops you may have?
User avatar
By ArnieO
#22348
iothing wrote:To eleminate any ground loops is it sufficient/advised to connect GND to the earth wire aka mains earth?

In some cases that could create a ground loop, not fix one.
The idea is to avoid loops by grounding all modules / equipments from one common point, in a star configuration. Ground loops can lead to electomagnetic noise in the environment creating currents in your circuitry by induction. You avoid that by breaking the loop(s).
If you google "Ground loop" there is quite a bit of information out there.-
User avatar
By kenn
#22376
iothing wrote:What I tried without success:
1. Connecting GND to mains ground
2. Connecting RST to VCC by jumper wire
3. Connecting RST to VCC via 5kOhm resistor
3. Connecting RST to GND via 100nF capacitor and
4. Removing the RST pin
5. Cutting the RST trace near to the chip
6. Running the esp off of a battery pack
7. Running the esp off of different psu's

I also have a 220uF cap between VCC and GND (near the PSU - problem?) and GPIO0 as well as CH_PD are pulled up with 5kOhm resistors.


I'm too lazy to read upthread. Have you tried a snubber diode across the relay coil? That eats up the reverse spike generated by the collapsing field when the relay is switched off.