ESP8266 Community Forum

Hi,
I'm having a strange problem that seems to be consistent across all the modules I try (4 from two different sellers) so I think the problem must be me! These all have the two LEDs, the blue TX LED and the red power LED.

Basically, the board only prints anything on the serial line when it starts up after a reset, or when I touch the CH_PD pin to ground then back to 3.3V. There's no response to AT commands at all, whether typed into the terminal or sent from an Arduino.



Plus what is really weird is that the serial prints at 76.8K Baud! It cannot be read at either 115.2K or 57.6K which is very strange and I can't find any instance of this happening on other sites. I even wrote a program that stepped through all Bauds in 400b increments starting from 2400 all the way to 150K to see if they'd get any response and there was none.

I'm using a typical FTDI Converter with a separate +3.3V supply pumping out around ~145mA when this thing is just sitting there. When I pull down the CH_PD or RESET pins, the current draw drops to ~45mA.

I've consulted about a million tutorials at this point and since I can get data from it at least at the start I know my serial lines are hooked up to the correct pins. Does anybody have any ideas?

Thanks - this is driving me crazy I want to play this soooo badly.

Ok this may interest people but I've found a solution.

I've tried this chip with 4 completely different USB to serial conversion boards, all based on FTDI chips but they all have a few different pin outs, components, etc.. I sell a lot of these so I have many different kinds laying around anyway, all from China in one form or another.

Ultimately, the behavior was ALWAYs the same (as described above) until I read on a hackaday comment here: http://hackaday.com/2014/09/06/the-curr ... velopment/ something to the effect that "logic input has high value of 3.6V" which got me thinking..how can I hit 3.6V if I'm running it off 3.3V as instructed? The article even warns that 5V will nuke the chip, which doesn't actually seem to be the case with mine. I have several dozen laying around so I'm not too terribly concerned about experimenting but I don't want to be consuming these like matches so at any rate I was hesitant..

Well I tried powering it with 5.0V and guess what..it worked. So far it's been running a little while and I haven't noticed it melting down or anything but it actually works now. It responds to my AT commands..soo go figured.

I can't really say why this worked, and frankly wouldn't advise it other than a last resort but it's certainly interesting to finally have a solution!

I did the same thing at first, but if you touch the esp8266 you'll notice it's burning hot when running on 5v.
Also my flash finally stepped out and the module wouldn't boot at all, until I let it cool down for an hour.

What you may be experiencing is a weak power supply, where do the 3v3 come from?
It seems an USB<>UART does not provide enough power for the modules, except for bootup.
Maybe after bootup the wifi hardware is engaged, resulting in a spike in drawn power?

Try running the module off constant 3v3 with at least 800ma (more than USB normally allows), then it should work just fine and keep the module cool.

Although "Absolute Maximum Ratings" for the chip do not appear anywhere in the leaked documents, the chip "Operating Voltage" is specified between 3.0V and 3.6V (3.3V typical).

Running the chip on a 5.0V power supply may have unpredictable effects, and those foolish enough to do this need to know that it may even be dangerous if heat is high enough to put some nearby material on fire, or if a short-circuit is drawing excessive enough current to melt the supply wires if no protection is present.

However, all theses cheap modules suffer from several problems. Among them is the lack of proper decoupling capacitor on the power supply: generally, only a single 100 nF capacitor is provided on board, although the datasheet application diagram recommends another 10 µF capacitor in parallel. This is even more adequate if you use long wires for the power supply: these will act like series inductors and will prevent the voltage from dipping when a sudden current is required, like when turning on the RF. The additional recommended 10 µF decoupling capacitor will provide the momentary power to compensate for this drop and will prevent the chip to perform a brown-out reset.

Any large enough value capacitor will do, but it must be soldered as close as possible to the chip, or on the header if not. A small 10 µF 0603 MLCC (Mutli-Layer Ceramic Capacitor) is OK.