Voltage dividers

Duckman · Post by **Duckman** » 21 Jun 2012, 02:29

I see how some classic designs use different approaches to bias his opamp and 10K, 47K, 100K or 1M are, more or less, the most used values. Since the vast mayority of VD's are two equal resistors, all of them give us 1/2 of V+ (usually 4.5v)
At least, I understand that part.

My question is: those different resistor values are related to each opamp specs or the companies just use what they have on hand?
How about DC, AC (a blank page for me) and noise?

DrNomis · Post by **DrNomis** » 21 Jun 2012, 02:55

Duckman wrote:I see how some classic designs use different approaches to bias his opamp and 10K, 47K, 100K or 1M are, more or less, the most used values. Since the vast mayority of VD's are two equal resistors, all of them give us 1/2 of V+ (usually 4.5v)
At least, I understand that part.
My question is: those different resistor values are related to each opamp specs or the companies just use what they have on hand?
How about DC, AC (a blank page for me) and noise?

The two resistor voltage-divider is one of the most basic circuit building blocks in Electronics, personally I really don't think the Op-Amp specs have any relation at all to the resistor values, the typical input impedance of an Op-Amp is something like 10M, so you can use any value from 10k to 1M, the only proviso is that the resistors need to be equal in value in order to bias the Op-Amp inputs/output to half the supply voltage so that the Op-Amp can produce the largest output-swing, noise is minimized by putting a capacitor in parallel with one of the resistors so that it bypasses the noise straight to circuit-ground....

juanro · Post by **juanro** » 21 Jun 2012, 03:11

If biasing an opamp, input current is the parameter to look at: a higher input current opamp with too high R's would deviate from 1/2 Vcc. Also, noise is important: the higher the R value, the higher the noise. So I would say: use the lower value R's you can afford, without draining too much current (say: using two 1K for the divider in a 9v circuit, you'll be wasting 4,5mA just for the biasing)
In other cases, like biasing the top FET in a mu-amp (very little "input current" for the reverse-biased gate) you can use 1M for the divider, but then noise (hiss) could be an issue. I seem to remember somewhere it recomended to use lower valued R's (10K, maybe) and then a 1M "in series" from the divisor to the gate (that resistor will not have a DC drop across it so it will generate very little -if any- Johnson noise).

Juanro

Duckman · Post by **Duckman** » 21 Jun 2012, 21:42

I read somewhere "1K per volt" as a a voltage divider "law"...

Any logic?
The MXR Distortion+ and DOD 250 share an almost identical topology, but MXR uses 1M to bias the 741 and DOD uses 20K, so they're working with a different current flow, right?
I'm guessing if that detail have something to do with the tonal caracter (again, very similar)