After how long do i expect the module to wear out?

[meneldor]

I have created 4 metering stations in my rooms measuring temp, co2, etc. 2 years ago. Recently i started observing some strange behavior in 3 of them - uart returning error checksums in the one reading from mhz19b sensor, other is crashing or cannot set time using the sntp lib, the third is something else..
I havent flashed them for at least 2 years and they were working flawlessly. I used idf v4.1 and the boards are geekreit esp32 wroom.
I have no other explanation except the boards components or the module to start to wear out.

[PeterR]

or your psu and/or other electrical design is poor
A digital 1 is a 1 and a digital 0 is, well 0.
Line driving is another matter entirely.

ESD is also another issue. ESD can kill slowly such that many do not believe that ESD is an issue. How did you protect the ESP communication interfaces?

[ESP_Sprite]

As long as you use them within spec (e.g. not putting 5V on an IO line directly) the chips should last forever. The flash does have a limited lifespan, but unless you write to it extensively it should have a practical lifetime of decades. Perhaps it's indeed a component in your power supply that has become marginal?

[username]

I havent flashed them for at least 2 years and they were working flawlessly. I used idf v4.1 and the boards are geekreit esp32 wroom.

IDF 4.1 was not even out 2 years ago. We have hundreds of modules out in the field more than 3 years old and have never had one come back. It more likely its your design or parts you used that is to blame.

[meneldor]

Thanks for the replies. So its most probably the quality of the Doit esp32 board.
However i have no explanation why the sntp sync started to fail on all of them.

[PeterR]

Most development boards do not contain proper ESD or even line capability. Raw input to pin? Really, what were they thinking - O, disposable or time to market! Iron triangle afraid.
Don't get me wrong, eval boards are essential to agile system development. Just often not suited for production (being wired for cost).
SPRITE is right. 5V will kill as will 2KV (over time) etc.
So its about what's on your line and why do you think that that signal is acceptable to ESP DC/AC published characteristics? Then start thinking about your qualification standards/requirements - which are you're friend if you want a certain level of reliability & essential if you place a product in most markets.
(PS or your PSU etc. Lost count of the number of years spent debugging software with a bad PSU)

An eval board is classed as a component and does not need to comply with (say) CE product standards (ok, might not be allowed to be radioactive, explode etc but you get the point).
Product/system engineering - that's you're add!
EDIT: It is not clear that you're issue is line protection or PSU related. Just quite likely until you measure and test. Line swap is perfectly good for home use (assuming no criticality), that's your value prop. PPS Back in the day 'select on test' was used to work round bad design