reading on every update()? about responsiveanalogread

[brianjmurrell]

Hrm. Battery aside, updating -- actually calling analogRead() I would guess -- on every loop iteration affects other things. For example, on an ESP8266, it causes the WiFi module to go non-responsive.

If I call .update() in every loop() cycle, my module only responds to ping for a few seconds after reboot and then stops. If I only call .update() once per second, this does not happen.

So, while I can surely add such a delay to all of my sketches, I think it needs to be reconsidered if this library really ought to be doing an analogRead() as frequently as every loop iteration or whether every user of this library needs to build delay into their sketches or whether this library should have some built in, user settable, sampling delay/rate.

[mikegleasonjr]

I guess it is you who decide when you call .update(). Examples are provided with an update in the loop but ultimately you choose.

[brianjmurrell]

Of course, I, and everyone else could write .update() delays/sampling into their main loop()s.

Do you not think doing so right inside the library so that .update() can just be called on every loop iteration is a useful feature of the library? I can most certainly push a PR if you think it's useful enough to merge it.

[mikegleasonjr]

Not maintainer just a library user. But in my opinion it belongs in userland (deciding when to update).

[dxinteractive]

As mikegleasonjr said, the library is only called every loop if you call it every loop. It's deliberately up to you as a dev when to call update, because the frequency at which you call it will affect some of the parameters you'll use.