Op amps as comparators

And that is why I think it is a good application note. With as many reasons as they give to support their first statement of not using opamps as comparators, they ultimately conclude that, when properly designed, that it can work and can be "a proper engineering decision." There is more than one way to design any circuit.
 

steveb

Joined Jul 3, 2008
2,436
I particularly liked the conclusion section:
I basically agree with the conclusion too - at least in principle.

There is a complication with realistically applying the conclusion of the app-note in the modern work environment. I've noticed that many companies take too many shortcuts in the engineering process. For at least 20 years now, there has been a trend to speed up the design process by skipping critical breadboarding and testing steps. I've seen this at companies I've worked at and know it exists at other companies by talking to friends. There are various things that drive this including economics, use of surface mount components and the availability of more sophisticated simulation tools. More than once I've seen a design go from schematic to PCB design with no breadboarding, then if the PCB seems to work it goes to production without adequate environmental and stress testing.

All this really bothers me on a gut level, and I can't say I agree with it, but it seems to be something some of us have to deal with. The process can work, but the engineer has more limited options in the design phase. Basically, no uncertainties can exist at any decision point; whether it be the choice of the component or the circuit topology.

In order for the above process to work, reliable simulation models need to be in place, and components whose performance is well-understood for the application should be used. In any kind of situation like the above, it is much harder to say that using opamps as comparators is a proper engineering decision. As the app-note points out, the spice models and quoted specifications do not always allow reliable prediction of the performance of an opamp in comparator applications. They recommend thorough experimental testing on multiple components to qualify the component, much the way the manufacturer would do when they put a "comparator" label on the device. This type of testing is very expensive in time, resources and money. The component would be tested over all input voltage ranges in various circuits with variations over the full temperature range. Testing of noise margins is also important, and there must be more that I can't think of off the top of my head. This is all fine as long as you have the opportunity to do it and have factored in the additional testing cost into the "engineering decision". Of course, once this is done, we are really debating semantics, because the "opamp" is a now a "comparator" by any engineer's standard.
 

Thread Starter

millwood

Joined Dec 31, 1969
0
And that is why I think it is a good application note. With as many reasons as they give to support their first statement of not using opamps as comparators, they ultimately conclude that, when properly designed, that it can work and can be "a proper engineering decision." There is more than one way to design any circuit.
that's basically true for anything in life. while it is certainly not graceful to race it, but you CAN race a buick roadmaster, and in the class of 0-60mph while hauling 8 people and a 50" TV, the buick will likely beat out a 911 hands down.

the key to any good circuits is less in the parts used, or the topology used. it is in the person who designs the circuits.

so any blanket statement that opamps cannot function as comparators or comparators cannot function as opamps is just a reflection of the person making that statement, not a reflection of the parts' inability to perform those functions, intended or otherwise.
 

steveb

Joined Jul 3, 2008
2,436
the key to any good circuits is less in the parts used, or the topology used. it is in the person who designs the circuits.
I would submit that the people who designed the parts, and the people who invented the common topologies are among the best designers the human race has produced. Certainly, the modern circuit designer is standing on their shoulders.

There is a long history of engineering invention showing the importance of creativity. There is also a long history of engineering failures showing the importance of caution. One thing I've learned is that arrogance is the root of a great many failures, even among very talented people.
 

beenthere

Joined Apr 20, 2004
15,819
Before this gets too strange - did anyone actually make this assertion? -
so any blanket statement that opamps cannot function as comparators or comparators cannot function as opamps is just a reflection of the person making that statement, not a reflection of the parts' inability to perform those functions, intended or otherwise.
This begs expansion -
the key to any good circuits is less in the parts used, or the topology used. it is in the person who designs the circuits.
Is it possible to have some specific example of this?
 
Top