Sigh. No they don't. RP2040 ADC has output resolution of 12 bits with ENOB of about 8.7. MicroPython probably scales those up.
Also, 65535.
This is surprisingly low quality blog post for HN standards.
Sigh. No they don't. RP2040 ADC has output resolution of 12 bits with ENOB of about 8.7. MicroPython probably scales those up.
Also, 65535.
This is surprisingly low quality blog post for HN standards.
https://docs.espressif.com/projects/esp-idf/en/release-v4.4/...
It even has almost the same graph showing the linearity and dead zone, none of this is a surprise. It's just not got the best ADC ever and I've chosen different parts because of it before. Other parts have better ADCs... and that Mega has a far lower sampling rate.
Is there any microcontroller that manages to have a 16-bit ADC? And if so, are all 16 bits actually usable?
>> The RL78/I1E features a 24-bit delta-sigma A/D
-------
That being said: even just 5V at 16-bits implies a LSB step of 76uV, which means you can only get to 16 usable bits if you have significant work on noise-reduction.
Going to 24-usable bits is probably impossible in practice, but it does mean that the 24-bit ADC there won't be the noise problem / bottleneck in your circuit.
------
Honestly, if any EE came to me and asked about more than 12-bits of ADC precision, I'm beginning to think that they don't understand their problem lol. I kinda feel like anything beyond 12 bits is asking for a lot of noise / hardcore analog signal conditioning circuits that most hobbyists really shouldn't be doing in their projects.
> And if so, are all 16 bits actually usable?
If you know for sure that you have white-noise, you can oversample + decimate to reach absurdly accurate levels. Maybe not 24-bits but yeah, you can reach 16-bits with probably 100k oversamples with effectively 1Hz sample rate
It really depends however. If your bandwidth is low enough then you at least have one way to do make those 16th or 18th bits usable.
Old standlone ADCs hit 16-bits pretty reliably.
Nah. Lets look at it from a db perspective... I don't believe that the typical electrical engineer is in the business of circuits accurate to -96db levels of precision.
Those circuits may have used a 16-bit ADC, but there's no way the typical circuit reached anywhere close to that level of precision. The bottom bits would be all noise in the vast majority of use cases. Even a 12-bit ADC takes a degree of effort to keep your errors and noise down.
16-bits means making no mistake in your entire circuit, across all temperatures, across all kinds of usages, such that your ADC line never drifts more than a few dozen microvolts off reality. That's just not typical circuit design. Temperature effects alone will screw you over and prevent you from reaching there.