Comparator, 50mV-100mV

[xwhatsit]

Hello all,

Please forgive what to many will be a simple response, but I'm a software guy and this primitive hardware stuff is new to me

I'm trying to detect a pulse which is about 100mV. Well, it rises to 100mV, then trails off quickly as the 4pF capacitor that it was generated from discharges.

I'm figuring I can use a comparator, which a Vref of about 70mV, then have it output at a level that my microcontroller can read it.

Can comparators read those sorts of voltage levels? Or do I need to run it through an amplifier first?

Many thanks.

 

[eem2am]

search for "rail-to-rail" comparators in digikey or google.

you may have to use a +/gnd/- power supply, and use the comparator from that.......because then your 100mV is then not down at the lower rail.

I wonder if the LM339 could do it.?

beware you know difference between "open collector" types and "differential output" types, (eg totem pole output

 

[poor mystic]

I think any decent comparator (not an op-amp) will work fine.

 

[xwhatsit]

Thanks for the responses.

I'd prefer not to use a +/gnd/- power supply as I'm going to be powering from USB, so only 5V available.

What parameter should I be looking for with respect to `sensing near ground', as the LM339 write-up mentioned? I understand rail-to-rail inputs (I think), but these don't necessarily mean it can distinguish between 50mV and 100mV above the lower rail, does it?

If it's too hard to do single-rail with an application like this I may just have to amplify the signal, transistors are cheap.

 

[poor mystic]

When I looked at the data sheet I found a 339 will sense right down to ground. Check it out!

 

[Laplace]

That 4 pF capacitor catches my attention. Even with a relatively high resistance such as 10K the time constant is only just 40 nanoseconds. Yet the response time of the LM339 is 1300 nsec. But if you selected a comparator that was 100 times faster, then you could expect oscillation on the trailing edge of the discharge curve without using hysteresis. So why is it 4 pF?

 

[xwhatsit]

I don't have a choice regarding the 4pF -- it's simply the capacitance produced by a key-switch when your finger presses on it.

So you're saying that the pulse will be too short to reliably switch the comparator?

 

[xwhatsit]

Actually I tell a lie. Just looked at it on the scope again and the pulse remains above 100mV for at least 80us. That should be long enough, shouldn't it?

I'm using the capacitor to differentiate the rise rate of the input pulse.