Op-amp closed loop gain question

[Dave Boland]

It has been a lot of years since I have done anything with op-amps, but
I seem to recall that some devices have a min. suggested closed loop
gain. I have also seen an app-note that has a gain of about .303
(inverting input). My question is what are the guidelines on:

1. Gain range. Can it go from less than one to 10 without stability
problems. The Fbw is DC to 1000 Hz, so I'm not worried too much about
the AC aspect (unless I should be).

2. My recollection is also that I should not select resistor values much
below 10K for either Rf or Ri. I don't see this as a problem now (fixed
gain), but doesn't it present a problem with variable gain designs?

3. Finally, I believe that using a pot. as Rf is frowned on because of
the noise of the wiper. If I need to ad some trim capability to the
design, should I do this with a pot connected to Ri?

Thanks for your time and any helpful answers.

Dave,

 

[John Larkin]

It has been a lot of years since I have done anything with op-amps, but
I seem to recall that some devices have a min. suggested closed loop
gain. I have also seen an app-note that has a gain of about .303
(inverting input). My question is what are the guidelines on:

1. Gain range. Can it go from less than one to 10 without stability
problems. The Fbw is DC to 1000 Hz, so I'm not worried too much about
the AC aspect (unless I should be).

Most opamps are stable for any inverting or non-inverting gain. A very
few are "undercompensated" and don't like low closed-loop gains, but
their datasheet will say so clearly.

2. My recollection is also that I should not select resistor values much
below 10K for either Rf or Ri. I don't see this as a problem now (fixed
gain), but doesn't it present a problem with variable gain designs?

If the R's are very low, they'll gobble signal current, but no harm
otherwise. If they are *very high*, the resulting frequency rolloffs
with stray capacitance could reduce amp stability.

3. Finally, I believe that using a pot. as Rf is frowned on because of
the noise of the wiper. If I need to ad some trim capability to the
design, should I do this with a pot connected to Ri?

Nothing wrong with a pot. Wipers aren't very noisy these days, and
only while in motion at that. And it doesn't matter whether it's in
the Rf or Ri path.


Since these statements are unarguably true, I expact lots of arguments
and flames.

John

 

[Tim Wescott]

Most opamps are stable for any inverting or non-inverting gain. A very
few are "undercompensated" and don't like low closed-loop gains, but
their datasheet will say so clearly.




If the R's are very low, they'll gobble signal current, but no harm
otherwise. If they are *very high*, the resulting frequency rolloffs
with stray capacitance could reduce amp stability.





Nothing wrong with a pot. Wipers aren't very noisy these days, and
only while in motion at that. And it doesn't matter whether it's in
the Rf or Ri path.


Since these statements are unarguably true, I expact lots of arguments
and flames.

John

"Expect" is spelled with an 'e' in the middle. Otherwise I totally
agree. Is this enough of a flame?

 

[John Popelish]

It has been a lot of years since I have done anything with op-amps, but
I seem to recall that some devices have a min. suggested closed loop
gain. I have also seen an app-note that has a gain of about .303
(inverting input). My question is what are the guidelines on:

1. Gain range. Can it go from less than one to 10 without stability
problems. The Fbw is DC to 1000 Hz, so I'm not worried too much about
the AC aspect (unless I should be).

Minimum gain less than 1 is not a problem for a unity gain stable
opamp (an opamp that can handle a direct connection from output to -
input).

2. My recollection is also that I should not select resistor values much
below 10K for either Rf or Ri. I don't see this as a problem now (fixed
gain), but doesn't it present a problem with variable gain designs?

You need to keep the feedback resistor current well below the maximum
current rating of the output. You also want ot keep the resistor
signal currents much higher than the opamp bias currents, if possible,
so that the input to feedback divider is dominated by signal current
not opamp imperfections.

3. Finally, I believe that using a pot. as Rf is frowned on because of
the noise of the wiper. If I need to ad some trim capability to the
design, should I do this with a pot connected to Ri?

Connect the wiper to the opamp input, and the input and feedback
resistors to the ends. That way, the only current passing through the
wiper is the opamp bias current. You also get a bigger gain swing for
a given pot resistance, since swinging the wiper moves resistance from
feedback to input or vice versa, instead of changing just one of these
resistances.

Say you wanted a gain of .1 to 10. With a 10k input resistor and a
10k feedback resistor connected to the ends of a 100k pot, the wiper
to the - input, you would have 110k input to 10k output at one extreme
(gain of less than .1) and an input of 10k, feedback of 110k (gain
greater than 10). Remember that most pots have a 10% or 20% overall
tolerance.

 

[John Larkin]

"Expect" is spelled with an 'e' in the middle. Otherwise I totally
agree. Is this enough of a flame?

Excellent work. I can see that you have a bright career ahead as a
spell checker.

John

 

[Tim Wescott]

Excellent work. I can see that you have a bright career ahead as a
spell checker.

John

Oddly enough my spelling is fairly poor. I can catch the odd misplaced
vowel, and I am (hopefully) good at choosing the correct word from the
list when the spell checker tags one for me, but that's about as far as
I go. Left to my own devices I can make some pretty embarrassing mistakes.

 

[John Larkin]

Oddly enough my spelling is fairly poor. I can catch the odd misplaced
vowel, and I am (hopefully) good at choosing the correct word from the
list when the spell checker tags one for me, but that's about as far as
I go. Left to my own devices I can make some pretty embarrassing mistakes.

Lately I find that I'm also misspelling via substitution of words that
are close in sound to the one intended. I don't bother to spell check
ng posts because so many of the words aren't in the dictionary, and
I'm not getting paid anyhow.

John

typed with a chicken nugget in my hand.

 

[Jim Thompson]

On Mon, 26 Apr 2004 14:39:45 -0700, John Larkin

[snip]

Lately I find that I'm also misspelling via substitution of words that
are close in sound to the one intended. I don't bother to spell check
ng posts because so many of the words aren't in the dictionary, and
I'm not getting paid anyhow.

John

typed with a chicken nugget in my hand.

I *try* to always spell-check since, with my klutzy typing, I tend to
skip letters :-(

...Jim Thompson