Notes on Opamp’s Gain

Esteban Rodriguez-Marek
April 16, 2011

1 Ideal Opamp
The ideal opamp, whose model is shown in Figure 1, has the following characteristics:

AOL = ∞ (1)
Rin = ∞ (2)
Rout = 0 (3)

Note that the AOL value is the open-loop gain, i.e. the gain when the device has no feedback connected.

Figure 1: Circuit model for the ideal opamp. Note that AOL = ∞ and, thus, v− = v+ .

This gain is unchangeable, as it is the gain of the device itself. If you design your own opamp (using
transistors), then you can design for AOL . If you are handed a device, its AOL will be stated in the data
sheet.
Adding feedback will modify the gain, as shown in class. The new gain will depend on the configu-
ration. For the inverting and non-inverting configurations, respectively, the new gain will be:

RF
Av,inv = − (4)
RI
RF
Av,non−inv = 1+ (5)
RI
Note that adding feedback implies we no longer will be able to have an open loop from input to
output, i.e. now we have closed the loop. Thus, the gains in equations 4 and 5 really are closed-loop
gains, or ACL . It is worth mentioning that even if we have a closed loop configuration, the output is

1
still

vout = AOL (v+ − v− ) (6)

vout
=⇒ v+ − v− = (7)
AOL
vout
= (8)
∞
= 0 (9)

2 Nonidealities
There are three parameters we can include in the study of the non-ideal opamp. The open-loop gain,
the input impedance, and the output impedance.

2.1 Non-ideal open-loop gain

For the purposes of this discussion, let’s keep the impedances ideal, i.e. Rin is open and Rout is a short.
When we assume a finite AOL , i.e. AOL 6= ∞, no longer is v+ = v− , instead, we have

vout = AOL (v+ − v− ) (10)

vout
=⇒ v− = v+ − (11)
AOL
This results in the closed-loop gain (i.e. the gain for either the inverting or non-inverting configura-
tions) being modified. In the text it is shown how the new closed-loop gains decrease by a factor
 
1 RF
1+ 1+ (12)
AOL RI

where the second part is denoted the gain error. Thus, the new closed loop gains for the inverting and
non-inverting configurations, respectively, are

− RRFI
ACL,inv =   (13)
RF
1 + A1OL 1 + RI

1 + RRFI
ACL,noninv =   (14)
RF
1 + A1OL 1 + RI

Note that the gains converge to those of the ideal case as AOL → ∞.

2.2 Non-Ideal input and output impedances

Now let’s assume that we have neither Rin = ∞ nor Rout = 0, i.e. the input impedance is no longer
an open and the output impedance is no longer a short. Our goal is to investigate what will happen
with the closed-loop gain with the addition of non-ideal impedances. Note that we are modifying are
the open-loop input and output impedances, i.e. we are making our opamp model more closely resemble
reality. Let’s study the gain changes by deriving the gain of the the non-inverting configuration amplifier,
as shown in Figure 2. Note that vin = v+ and thus

2
Figure 2: Circuit modeling the non-ideal amplifier (i.e. finite AOL and Rin and non-zero Rout ) connected
in non-inverting configuration.

vout
v− = vin − (15)
AOL
Applying KCL at the inverting node we have
v− vin − v− vout − v−
= + (16)
RI Rin RF
 
1 1 1 vin vout
=⇒ v− + + = + (17)
RI Rin RF Rin RF
 
Rin RF + RF RI + RI Rin vin RF + vout Rin
=⇒ v− = (18)
RI Rin RF Rin RF
Applying some mathematical manipulation we arrive to
RI RF RI Rin
v− = vin + vout (19)
Rin RF + RI RF + RI Rin Rin RF + RI RF + RI Rin
Now we apply KCL at the output node and we get
AOL (vin − v− ) − vout vout − v−
= (20)
Rout RF
AOL vin AOL v− vout vout v−
=⇒ − − = − (21)
Rout R Rout RF R
 out   F 
1 AOL 1 1 AOL
=⇒ v− − = vout + − vin (22)
RF Rout R Rout R
   F   out 
Rout − AOL RF Rout + RF AOL RF
=⇒ v− = vout − vin (23)
RF Rout RF Rout RF Rout
The denominator cancels out, and after rearranging we get
Rout + RF AOL RF
v− = vout − vin (24)
Rout − AOL RF Rout − AOL RF

3
We now equate equations 24 and 19 and solve for the gain, i.e. vout /vin :

Rout + RF AOL RF RI RF RI Rin

vout − vin = vin + vout
Rout − AOL RF Rout − AOL RF Rin RF + RI RF + RI Rin Rin RF + RI RF + RI Rin
(25)
Combining terms
 
Rout + RF RI Rin
vout − =
Rout − AOL RF Rin RF + RI RF + RI Rin
 
RI RF AOL RF
vin + (26)
Rin RF + RI RF + RI Rin Rout − AOL RF

Add the fractions inside the parenthesis to get

Rout RI RF − AOL RI RF 2 + AOL RF 2 Rin + AOL RF 2 RI + AOL RI RF Rin

 
vout =
(Rout − AOL RF )(Rin RF + RI RF + RI Rin )
Ro Rin RF + Rin RF 2 + Ro RI RF + RI RF 2 + Rout RI Rin + RI RF Rin − Rout RI Rin + AOL RI RF Rin
 
vin (27)
(Rout − AOL RF )(Rin RF + RI RF + RI Rin )

Canceling out the denominators and the quantities with opposing signs, dividing both sides by RF , and
solving for the gain, we get
vout Rout RI + AOL RF Rin + AOL RI Rin
= (28)
vin Rout Rin + RF Rin + Rout RI + RI RF + RI Rin + AOL RI Rin
Rout RI + AOL Rin (RI + RF )
= (29)
Rout Rin + RF Rin + Rout RI + RI RF + RI Rin (1 + AOL )
Rout RI
AOL Rin
+ RI + RF
= Rout (30)
AOL
+ AOL + AOL Rin + AROLI RRFin + AROL
RF Rout RI I
(1 + AOL )
RF Rout
1+ RI
+ AOL Rin
= Rout RF Rout RF 1
(31)
AOL RI
+ RI AOL
+ AOL Rin
+ AOL Rin
+ AOL
+1

which leads to the final expression

RF Rout
vout 1+ RI
+ AOL Rin
=   (32)
vin 1+ 1
1+ RF
+ RF
+ Rout
+ Rout
AOL RI Rin RI Rin

(Note: I omitted details. You can take a couple of minutes to find them. Please do not ask me about
the intermediate steps. I am sure you can find them yourselves. However, in transcribing these notes I
may have made a mistake. Please let me know if you find one.)
Let’s make the following observations about the gain in equation 32:

1. When looking at the case when Rin → ∞ and Rout → 0, it is easy to see that the gain converges
to
vout 1 + RRFI
=   (33)
vin 1 + 1 1 + RF AOL RI

i.e. the gain when we only considered AOL to be non-ideal (equation 14).

4
 2. The gain when we assume the output impedance Rout is ideal, i.e. Rout = 0, is
vout 1 + RF
=  RI  (34)
vin 1+ 1
1 + RRFI + RF
AOL Rin

3. The gain when we assume the input impedance is ideal, i.e. Rin = ∞, is
vout 1 + RRFI
=   (35)
vin 1+ 1
1 + RRFI + Rout
AOL RI

4. Note that the gain has a new term in the denominator, namely
Rout
(36)
AOL Rin
It is not hard to dismiss this term as negligible, as typically the numerator is small and the
denominator is large.
Rout
5. Regardless of the term AOL Rin
in the numerator, we can observe that the gain error in equation 12
has been increased to  
1 RF RF Rout Rout
1+ + + + (37)
AOL RI Rin RI Rin
Thus, we conclude that non-idealities cause a reduction from the nominal gain. The reduction
depends on the values of Rin , Rout and, of course, AOL .
Example: A noninverting amplifier has the following feedback resistors and internal characteristics:
RF = 9kΩ, RI = 1kΩ, AOL = 2500, Rin = 100kΩ and Rout = 10Ω. Then the nominal gain is
RF 9k
ACL = 1 + =1+ = 10 [V/V] (38)
RI 1k
The gain including all non-ideal characteristics is:
RF Rout
1+ RI
+ AOL Rin
ACL =   (39)
1 RF RF Rout Rout
1+ AOL
1+ RI
+ Rin
+ RI
+ Rin

1 + 19 + 2500·100k
10
= 1
(40)
9k
1 + 91 + 100k 10 10


1 + 2500 + 1k + 100k
1 + 9 + 40E −9
= 1 (41)
1 + 2500 (1 + 9 + 90E −3 + 10E −3 + 100E −6 )
= 9.9597622 [V/V] (42)
The reduction in gain with respect to to the nominal 10 [V/V] is about 0.4% (please check my math).
A similar procedure can be followed to obtain both the input impedance, Zin , and output impedance,
Zout for the whole circuit. Note that these impedances will be a function of the opamp’s Rin and Rout ,
but not equal to Rin and Rout .

2.2.1 Inverting Amplifier

Now let’s do the whole thing over for the inverting amplifier. Just kidding. Once you have the gain
error, we can simply extrapolate and it is possible to prove that the gain is
vout − RRFI + Rout
AOL Rin
=   (43)
vin 1+ 1
1+ RF
+ RF
+ Rout
+ Rout
AOL RI Rin RI Rin