They can clamp the input and/or the reference! My inputs are often capacitor+current source ramps, and my references are often dacs. They don't like being coupled to one another. And sometimes one signal feeds more than one comparator, so think about the consequences of that one.
John
Becxause it demands that both of the comparator inputs be fed from zero-impedance sources that can drive several milliamperes.
John
John Larkin a écrit :
Sure. OTOH I wonder what would be the consequences of zenering the input transistors EB junction and how much reverse voltage they can accept.
And you don't have an anti-parallel diodes pair, but 2 in series.
Hello,
I think
- most important property of a OPA: linearity, use for analog signals
- most important property of a comparator: High slew rate, use for digital output
Sometimes I use an OPA as comparator, because some glitches are filtered, because the lower slew-rate. So I need no additional filtering.
Regards
Wolfgang Weinmann
--
Both Analog and Maxim have sub-ns comparators whose inputs can be railed any way without zenering, so it's strange that the "drop-in" replacement for the 9690 has clamp diodes. I could live with a maximum differential voltage spec, but they didn't give me that choice.
They leak seriously at about 1.3 volts at high temp, and that's often a nasty.
John
Demands a zero-impedance source? Bollocks. It demands you keep the inputs within a diode drop of one another.
Tim
Which reduces precision and increases jitter by almost an order of magnitude, and requires that things like DAC outputs be divided down and usually rebuffered. If you don't mind too much, I'll buy something else.
Hey, if I can keep the inputs close enough, I don't need the comparator at all!
John
Ohmygod. I thought you were talking about protection diodes against the supply rails, with which I can't see anything wrong. But a closer look at the DS shows the "Differential Input Clamp Voltage 1.7V"
Yes, that looks very dumb. So much that I could have prototyped something with such a part and wondered why it wouldn't work, and then would have found the comparator's inputs shorted. Well, perhaps I would have stumbled on that line in the DS, but if not I probably would have chucked out the "faulty" part and put in a new one only to find the same fault.
Thanks for pointing out somthing that I didn't know existed.
robert
Right. We are not limited by what they call it.
John
Have something to add? Share your thoughts — no account required.
Ask the community — no account required