Update: Much has been solved. For remaining questions, see
https://forum.allaboutcircuits.com/...ith-unipolar-power-supply.167724/post-1485109

original question below:
=================================

This isn't working.
http://tinyurl.com/vbjl5x7

 

This isn't working.
http://tinyurl.com/vbjl5x7

View attachment 200872

And how do you expect it to work?
You have both inputs at 0V so the output is ambiguous, determined by circuit offsets.
You need the (-) input slightly positive to have 0V at the output with 0V input.

 

You have both inputs at 0V so the output is ambiguous, determined by circuit offsets.
You need the (-) input slightly positive to have 0V at the output with 0V input.

How to make a comparator that switches when input goes over zero volts?
Or some other kind of over-zero detector?

 

How to make a comparator that switches at zero volts?
Or switches when both inputs are equal?

By definition you can't.
You need some finite voltage difference to cause the switch.
Otherwise what is the output when inputs are equal, high or low?
It can't be both.

 

i corrected my above misstatement.
i want the output to go low when input = 0V, stay high when input goes over 0V.
unanswered here: https://forum.allaboutcircuits.com/threads/comparator-referenced-to-ground.165850/

 

go high when input goes over 0V.

How much over 0V?

 

Not sure. Just above a typical noise floor. i know, no such thing as typical! Sorry i don't have more detail at the moment. 1 uV? 1 mV?

Whatever's easy with about 5 parts or less.

 

Opamps/comparators (other than expensive ones) typically have an input offset of a few mV, so a threshold of, say, 10mV should be readily achievable with an LM393 comparator or similar.

 

Thx, @Alec_t . Can you show the simplest practical schematic, i'm having trouble finding an LM393 circuit for the simple case.

 

Not sure. Just above a typical noise floor. i know, no such thing as typical! Sorry i don't have more detail at the moment. 1 uV? 1 mV?

Whatever's easy with about 5 parts or less.

I think your specs are going to cause you problems. Don't know of any comparators that will have an input offset voltage low enough. Signal coupling (from the air) will be an issue, so shielding will be required.

 

I think your specs are going to cause you problems. Don't know of any comparators that will have an input offset voltage low enough. Signal coupling (from the air) will be an issue, so shielding will be required.

i'm not that picky. Whatever's the lowest default offset i can get with common parts, no special circuitry to reduce offset, and no shielding.

What's lowest-parts-count way with LM393? I think this one from the LM393 datasheet will only give a pulse on zero crossings, rather than staying high when the input is over 0V which is what i need. And it's for bipoloar power, when i need unipolar.

 

And.....let’s not forget about hysteresis.
All circuits and signals have noise, and as you get close to the threshold, the output will oscillate.

the question you must ask is how much noise does the compared signal have? Or are willing to tolerate?
And no. It won’t be zero.

 

i want the output to go low when input = 0V, stay high when input goes over 0V.

When the difference between the non-inverting and inverting input is positive, the output will be HIGH.

With a single supply, you need a comparator with a common mode input range that includes ground.

 

And.....let’s not forget about hysteresis.
All circuits and signals have noise, and as you get close to the threshold, the output will oscillate.
the question you must ask is how much noise does the compared signal have? Or are willing to tolerate?
And no. It won’t be zero.

i assume that a resistor (hopefully just one!) will control my hysteresis, correct? If so, can you show the circuit? I can tune it on the bench.

When the difference between the non-inverting and inverting input is positive, the output will be HIGH.
With a single supply, you need a comparator with a common mode input range that includes ground.

So, not the LM393? Can you suggest a device? thx

 

So, not the LM393?

LM393 includes ground in the input range:

 

Here's the offset voltage spec:



A conservative design would use the worst case spec, plus or minus 9mV.

You could add circuitry to adjust the offset closer to 0V, but you might need a negative supply and a DVM accurate enough to measure it.

 

Here's the offset voltage spec:
View attachment 200911
View attachment 200912

A conservative design would use the worst case spec, plus or minus 9mV.

So, about 20 mV?

You could add circuitry to adjust the offset closer to 0V

i can go with the OTS offset.

DVM accurate enough to measure it.

Can't i just view the oscillation on a scope?

 

So, about 20 mV?

It depends. If the input offset voltage was 9V and you wanted a difference of 20mV to cause the comparator to change states, it might change with only 11mV.

Can you share the circuitry to adjust hysteresis? Or is that what you mean?

Offset voltage and hysteresis are different things. Hysteresis introduces positive feedback to stop oscillations at the trip voltage.

Can't i just view the oscillation on a scope?

Don't know of any scopes that can measure uV or mV accurately.

 

it might change with only 11mV.

No matter for my usage. OTS behavior is fine. All i need is to tweak hysteresis to eliminate oscillation.

Don't know of any scopes that can measure uV or mV accurately.

not trying to measure, just to trim hysteresis R until oscillation stops. A scope can display oscillation, correct?

Can i find the circuit someplace?
thx

 

Can i find the circuit someplace

From the TI datasheet:


You can also get it from National Semiconductor Linear Applications handbooks. They'll also have offset null circuits for opamps/comparators that don't have dedicated null pins.