ESP8266 Community Forum

Thanks all. Not the answer I was hoping for however it is the answer I was expecting.

Bergenheis wrote:Data sheet says Vcc + 0.3V, so 3.6V. You would probably get away with it if you are just experimenting or making a temporary project, but reliability and longevity would be affected.

There's a difference between allowable input voltage, and allowable voltages on the IO pins.

Vcc + 0.3V means you can have a voltage on the IO pins up to 0.3V above Vcc.

Vcc itself can be 3.0 ~ 3.6V (from the datasheet). If Vcc is 3.0V and you stick 3.6V on an IO pin, current will flow from the IO pin to Vcc due to the input protection diodes (3.6V - 3.0V = 0.6V which is enough to forward bias the diode)



Just something to be wary of. It's not usually explicitly explained in any datasheets what it means.

Barnabybear wrote:Hi, thats a bit to high. I have used lipos but with a diode in series, dropping 0.7V takes it down to 3.5V and that has been fine with the work I've done.
This might not be ideal if your going to use deep sleep as the current draw through the diode will drop below the 0.7V threshold and the voltage will rise back up to 4.2V during sleep.

Interesting point - the forward diode drop will fall as current tends to zero but then there's no stress on the ESP8266 with zero current going into it! The reason 3.6V is the quoted absolute maximum is because the on-chip regulation for internal 1.8V regions would be dissipating too much at operational currents. It's extremely unlikely that there are any semiconductor junctions or capacitors that would break down at 4.2V - besides, if any "breaking-down" were to happen, only a few uA would flow before the forward drop was re-established. Very self-limiting.

Out of interest, with a high-impedance DMM, I just measured vdd at 3.75V when the device is in deep sleep while connected to 4.2V via a 1N4001 diode. The current into vdd was 65uA - about right I think

I have tried many different solutions with regulators and diodes with a LiPo battery but have found it very unstable. I am stumped as to how others seem to have success with a diode. Problem I have is that it just crashes/resets when using a diode. I have tried added all manor of caps, 470uf + ceramic for decoupling but still no luck whereas connecting straight to the LiPo works fine.

I have a cheepo oscilloscope, and when using a 1N4001 diode the voltage only reads around 2-2.5v across the vcc/gnd of the ESP-12. I just don't understand what I am doing wrong.