search

tttapa / ESP8266

Documentation and help with the ESP8266 chip/boards/modules
GNU General Public License v3.0
661 stars 283 forks source link

The GPIO pins are 5V tolerant #61

Open Bartvelp opened 5 years ago

Bartvelp commented 5 years ago

Hi,

In your hardware guide you indicate the io pins are not 5v tolerant, while they are. Source: ba0sh1.com

gmacario commented 5 years ago

,,

Il dom 30 giu 2019, 20:30 Bartvelp notifications@github.com ha scritto:

Hi,

In your hardware guide you indicate the io pins are not 5v tolerant, while they are. Source: ba0sh1.com https://www.ba0sh1.com/blog/2016/08/03/is-esp8266-io-really-5v-tolerant/

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/tttapa/ESP8266/issues/61?email_source=notifications&email_token=AAASLLWR7UJWTJB5YMJUZHLP5D3VTA5CNFSM4H4NE5D2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4G4QFKYA, or mute the thread https://github.com/notifications/unsubscribe-auth/AAASLLQXFEMF5H47PHW2S2DP5D3VTANCNFSM4H4NE5DQ .

Tech-TX commented 5 years ago

Just a bit of follow-up documentation: here's the article from Hackaday where they opened the question of 5V tolerance of the GPIOs, and the CEO of Espressif responded with a Facebook update that said yes, the GPIO pins are 5V tolerant: https://hackaday.com/2016/07/28/ask-hackaday-is-the-esp8266-5v-tolerant/#comment-3122361 The Facebook comment from Teo Swee Ann was “I can reply officially here: it is 5V tolerant at the IO. while the supply voltage is at 3.3V.”

The test done by ba0sh1 (above) is also linked in the Hackaday article.

The module power is NOT 5V tolerant, as the flash chip may die with 5V on it's power pin. That suggests that GPIO6-11 on ESP-12 modules and DevKits (SC, SK, S0-S3) are probably also not 5V tolerant as they connect directly to the flash chip. The RST pin on the module is not 5V tolerant, either.

Mohammad271091 commented 5 years ago

can you guys plz answer my issue #66 about GET requests ...

Buzzardbait commented 5 years ago

In general, I always use level translators on the I/O pins of the NodeMCU SPI except for one pin. You cannot use a level translator on the SPI SS (slave select pin) because it will lock up the NodeMCU). I've also used the I2C pins without level translators and they work ok. Or at least they havn't blown up yet. :-)

mk-pmb commented 4 years ago

I'm all for at least mentioning the CEO claim, albeit with a warning that other technical parameters might deviate from specs when used with out-of-spec inputs, as mentioned in the Hackaday comments. (Random FUD thought: Will the excess voltage applied to your specific specimen boost antenna strength beyond FCC limits?)

Also a preliminary warning: When asked about the opening post's article, one of my advisors on electronics said the experiment seems to have some serious flaws we'll need to discuss, so you might want to ask your own advisor before relying on that report. (I'll update this warning as soon as I understand the concerns.)

Update: The concerns were mostly a lot of ideas what kinds of implementation of the hardware could cause what kinds of (potentially random, potentially delayed) malfunctions, with the overall gist that

  1. When you exceed the specs limits, the chip is "allowed to" deviate from known behavior in almost any way imaginable (electrical as well as stuff like broken timers, sensors, RNG, even variying with ambient temperature or unknown factors), and will probably "choose" whatever will be worst for your project, at the worst time. :-P
  2. Experimental results from one production batch may not apply to another.
  3. Only trust what's in the datasheet.

So maybe we should clarify what "5V tolerant" means. Ideally it would mean "Will work as usual, reliably." A lesser claim like "Won't burn within the first four hours" might be way easier to establish.