TL072 minimum supply

[newbrain]

Perhaps I should been clearer that this is a single-rail job and it's only comparing a something passing a voltage at about 50% VCC so the design isn't picky - by, er, design.

Depending on the actual requirements for a precise threshold and the amount of hysteresis needed (you are adding hysteresis, aren't you?), did you consider the use of a single gate Schmitt trigger (74xxx1G14 family)?
They are small, cheap and require no external passives, just a decoupling cap.

[marcdraco]

Depending on the actual requirements for a precise threshold and the amount of hysteresis needed (you are adding hysteresis, aren't you?), did you consider the use of a single gate Schmitt trigger (74xxx1G14 family)?
They are small, cheap and require no external passives, just a decoupling cap.

I sure did @newbrain - and amazingly I actually found a 6-gate 40-series in my lolly box from decades ago! I also found an LM358 that I thought I had around the same time - embedded in my foot. (Note to self, put the top back on the IC box when the cat is around.  ) Never ceases to amaze me the stuff I've bought and put to one side for another day.

Right now this is only at a very early stage of proof of concept and there are other ways I can achieve what I'm up to. But the point about being simple by design also dictates what's needed. This part is a one-shot following an integrator driven by a constant current source so the voltage continues to rise (until it the current source maxes out) but by that point the rest has triggered so...

Adding a Schmitt isn't necessarily a bad idea but I can't see any real use mathematically but as math isn't my strong suit I'm willing to listen to advice from those wiser (and probably younger) than me.

Development is a strange process and looking at some new libraries last night made me think that I'm may be using the wrong MCU for any final design. Speed is a factor here - not for the pure electronics as that's sloth-like but for the display which would benefit from a much faster SPI bus. I've worked with a parallel-bus board on a 328p and managed to get some fairly impressive speed by tweaking the Adafruit library to use 8-bit where only 8 bits are needed, but using the SAMD21 over SPI the results are a little disappointing considering the chip is clocked at 3x faster.

This little fella (the software, not the chap behind it) looks very interesting and seems to support a lot of board/MCU combinations. His demo video (which, of course, I can't find now!) really made me sit up and take notice.

https://www.arduinolibraries.info/libraries/gfx-library-for-arduino

[Zero999]

Note that LM358 output doesn't swing very close to VCC, so a pullup resistor may be necessary.

Just run the LM358 off 5V and its output voltage will be a at least 1.1V less the supply voltage, giving 3.9V out maximum, which an MCU run off 3.3V will tolerate very well, as the current flowing through the ESD protection diodes will be tiny.

[iMo]

I was using TL072 with a 3.3V MCU (bluepill), 5V and -3V power.
The TL072 output went from 0-3.5V as I can remember.
You may put a ~1k resistor in series with the TL072 output - that will limit the current when the top clamping diode (inside the mcu input) goes on.

[Zero999]

I was using TL072 with a 3.3V MCU (bluepill), 5V and -3V power.
The TL072 output went from 0-3.5V as I can remember.
You may put a ~1k resistor in series with the TL072 output - that will limit the current when the top clamping diode (inside the mcu input) goes on.

Yes, 1.5V less than the supply is more typical. 3.9V (1.1V less than the supply) is the highest you'll get, at high temperatures, when the base-emitter voltages of the output transistors are lowest.

It will probably be fine without a resistor, but 1k is a decent enough value, just in case.