[sdiy] unused comparators

Magnus Danielson magnus at rubidium.dyndns.org
Sun Nov 3 12:50:58 CET 2013

On 11/03/2013 05:39 AM, David G Dixon wrote:
>> On 11/03/2013 04:55 AM, David G Dixon wrote:
>>> Yes (based on what Magnus says below), the more I think 
>> about it the 
>>> more it seems dumb to ground the inputs of an unused 
>> comparator, since 
>>> the output state will be indeterminate.  Better to connect 
>> the rails 
>>> to the respective inputs, or one rail and ground at least, and to 
>>> arrange it so that the input doesn't consume current.
>> Beware, railing them may not be a guarantee for stability. 
>> That is a feature of a particular design. Another feature may 
>> be that the valid input range may include the rail... such 
>> that railing both inputs is just as dumb as grounding them. 
>> When you don't know better, park them at distinct levels from 
>> each other!
>> Just did a quick search for "comparator rail to rail input" 
>> and found a bunch of them.
>> Cheers,
>> Magnus
> So, what you're saying is: comparators suck!  Yeah, I concur.  In any case,
> if I did have to leave one unused (which I don't), I would hook the inputs
> up to +5V and GND.  Since, I'm running some TTL off them anyway, I have a 5V
> source in the circuit, so yeah.  I wouldn't hook it up to the rails.
No, I'm not saying that they suck. Actually, when they do not have an
input range of rail-to-rail they suck, but these days they do handle that.
What I am saying is that you need to understand the tool, and handle it
with care. I give advice which may apply generally, but is also saying
that you might use it differently, if you read the fine-print in the
manual (datasheet). My general recommendation will however be wise to
follow, as it allows you to drop in a different comparator and it will
still be stable in this case.

Learn to use the tool wisely, and here are my thoughts about what that
means for comparators.


More information about the Synth-diy mailing list