The future of ESP32 hardware revision(fixing annoying hardware issues?)

[MikeLemon]

Hello,

I've invested quite a lot of my time on the Espressif platform for IOT and general MCU development and there are a few known hardware Issues with the chip that I'd like to know if espressif intends to fix:

1) ADC Impracticality: So we all know the current ADC sucks in short it is advertised as 12Bit but is closer to 5Bit actual resolution with ability to detect approx voltage changes of only 100mV making even the basic arduino uno's atmega328 10Bit ADC a better ADC a few times over, Not to mentions the lack of input voltage reference pin ect...

2) Lack of gpio pins and lack of gpio functionality with most current available pins it is very often that this limits my ability to get a project running just because there are not enough GPIO's

these are the major ones for what comes into my head currently

Just wondering if Espressif is working on integrating an actual functional ADC and if there is a plan to add more IO pins(with more full functionality) into the next revision of the chip? and if so when is it estimated to be out?

I was also wondering where is it possible to obtain the latest most relevant news and updates from Espressif?

Another question that doesn't deserve opening a thread for:
I'm currently developing a few libraries (With the ESP-IDF) for common hardware that I plan to open source in C but since they are quite complex in functionality making em in C++ would be much more convenient so I was just wondering:
How much "Against the stream" would doing that be?

I barely see any C++ libraries for the ESP32 and sins I'm putting quite a lot of work into them I'd like to make sure they are as "future proof" as can be.

[WiFive]

Maybe in esp32-s3 which does have more pins but probably not esp32

[MikeLemon]

I hope Espressif does in the next few month I'd just be such a shame for me to switch manufacturer and drop off all gathered code snippets and knowledge of the ESP just because of this bottle necks.

[WiFive]

What chip would you switch to instead of adding an external adc?

[bobolink]

What chip would you switch to instead of adding an external adc?

Exactly.
That’s a good question. Maybe it’s too difficult to put a good ADC a microcontroller.

[MikeLemon]

I'm sure of it and I realy like the ESP32 as a whole otherwise I wouldn't spent so much of my time learning about it, but the ADC is still quite unusable and bad for sensor readings in comparison to other MCU's (which is a key in miniaturization and cost saving of products) and it's mostly about the IO count and functionality I assume It's because Espressif is still a relatively a young player in this field.

Qualcomm and NXP seem to offer solutions (but still below Espressif in terms in functionality and cost) for wifi integrated MCUs didn't dive too deep into there sins I'm still optimistic about the ESP progress.

[PeterR]

Qualcomm and NXP seem to offer solutions (but still below Espressif in terms in functionality and cost)

Exactly, 'you pays your money....'

So (basically) finding a silicon sweet spot is very tricky! Also very tricky for anyone else to review that trade without knowing your ram/flash etc requirements etc.
A SPI/I2C GPIO chip is cheap but clearly only good for lower frequency readings/updates.
I usually end up with a GPIO chip (& on an PIC/STM/NXP etc) and often a V&T env chip but the last one is more the managers choice than mine ....
ESP ADC does suck for sure. I am not sure that anyone uses a uC ADC for 'signal processing' however.
You did not mention your ADC Use/Case. I would be more worried about blocking/latency on the ESP for any signal processing & regardless of external or internal ADC.
Happy architecting!

EDIT: I will add that often I give my younger engineers a test bench kitting job and then some days latter I wonder why I don't have a purchase request. Typically the issue is that kit cost is $100 and the engineer is trying to get that down more towards $50. 50$ saved is quite a lot of beers to anyone of course. I try to be gentle when explaining that we are only producing one or two benches. The fact that the engineer is getting $25 p/h & charge out is $100 p/h is probably best not mentioned but on the latter the value proposition is clearer.
So that is a long way to question your NRC and UPC and volumes.

[Jara777]

Quite agree with the issue on ADC, I have almost done the project but the ADC reading resolution cannot be ignored. I am about to investigate on esp32-s2 board, anyone who has the development experience know if the ADC is improved or having the same problem?