Large Feedback Resistance in MFB Bandpass Filter

Discussion in 'Homework Help' started by SkiBum326, Jul 15, 2014.

May 16, 2014
33
0
Hey Guys,

Quick question about MFB bandpass filters. I attached an image of the particular circuit I have a question about.

The book I'm reading states that R2 is usually less than 100kohms to avoid problems with input offset and bias current.

What problems could arise? I thought that amplification of either could be what the book implied, but that shouldn't be an issue since the target signal, the input offset and bias current would all be amplified by the same factor.

Thanks,

Austin

File size:
1.4 MB
Views:
41
2. crutschow Expert

Mar 14, 2008
16,576
4,478
Don't quite understand what you mean by
That's precisely the problem. You typically don't want a DC offset at the op amp output. You only want the AC signal.

May 16, 2014
33
0
When they discuss limiting R2, I assume that's to limit the voltage gain. However, I'm unclear why that matters. The input offset voltage and bias current will be there regardless. Furthermore, they will be amplified in proportion to the target signal, so why should it matter if they are amplified by a factor of 2 (completely arbitrary value) or a factor of 100, since the target signal is amplified by the same factor.

4. crutschow Expert

Mar 14, 2008
16,576
4,478
The purpose of changing R2 is not to change the voltage gain. True, the voltage offset is amplified by the voltage gain. But the input bias current generates a voltage across R2 equal to the current times the resistor value which, in turn, is multiplied by the voltage gain, adding further to the output offset voltage. That's why you don't want the resistance of R2 to be too high. Of course, if you reduce the value of R2 then all the other component values must be changed proportionally to maintain the same filter response.

SkiBum326 likes this.
5. MrAl Distinguished Member

Jun 17, 2014
3,758
791
Hi,

Typically you dont want to let any of the resistors get too much above 100k if you can help it, but that's mainly to keep the impedance of each node down to a reasonable level in order to keep injected noise down low.

Why they even mention input offset with respect to R2 with this circuit is puzzling. That is usually a problem when there is DC coupling. When there is DC coupling the offset (which is also DC) gets amplified, but with this circuit the offset does NOT get amplified. This is because there is no DC voltage divider effect between R2 and some other resistor (like the input resistor in many circuits).

With a zero volt input and DC coupling, the offset is amplified by the gain of the circuit because the output Vout must reach a higher (or lower) level to obtain a null input for DC. With this circuit, when we input zero volts (or any DC value) the circuit settles with a DC output voltage equal to the input offset, but no more assuming a reasonably high input resistance. So even with an op amp with a 10mv DC input offset, the output will go to plus or minus 10mv and that's about it. That will be the same with R2 being 1k, 10k, 100k, or even 500k, unless the input impedance happens to be comparable to R2, which it usually is not because it's usually very very high. In fact, with a input resistance of 500k and feedback R2 resistor of 500k the output would only go to plus or minus 20mv with a 10mv input offset.

My guess here is that they applied a general principle of op amp circuits to a circuit where that principle was not really applicable.
It was probably just a small oversight.

glene77is_1444393601 likes this.