# Two opamp differential amp.

Started by May 23, 2012
``` I reinvented the two opamp differential amp yesterday.
And then found it in AofE* (2nd Ed. figure 7.36).  Though I drew it
differently.

+-R2--+--R1--+
|     | |\U1 |    +--R2--+
gnd    +-|-\  |    | |\U2 |
|  >-+-R1-+-|-\  |
V(-)-----|+/         |  >-+--out
|/   V(+)---|+/
|/

It=92s a natural addition to the single opamp difference amp.
(boost up V(-) rather than throw away V(+). (G =3D 1 + R2/R1)

The output of the first opamp U1 bothers me a bit.  As it has to
handle the current from U2=92s gain.  I=92d like to run this up at high
frequency, and I=92m worried about the output impedance of U1 screwing
up the CMRR.  My current thought was to keep the gain low, and use
R2=3D1k and R1=3D 10k. That should help minimize output impedance effects.

Anyone used this configuration?

Thanks,
George H.

*I think I really did reinvent it... but then seeing it in AofE I
wonder if the back of my brain is just remembering it.
```
```On May 23, 9:52=A0am, George Herold <gher...@teachspin.com> wrote:
> =A0I reinvented the two opamp differential amp yesterday.
> And then found it in AofE* (2nd Ed. figure 7.36). =A0Though I drew it
> differently.
>
> =A0 +-R2--+--R1--+
> =A0 | =A0 =A0 | |\U1 | =A0 =A0+--R2--+
> =A0gnd =A0 =A0+-|-\ =A0| =A0 =A0| |\U2 |
> =A0 =A0 =A0 =A0 =A0 | =A0>-+-R1-+-|-\ =A0|
> =A0V(-)-----|+/ =A0 =A0 =A0 =A0 | =A0>-+--out
> =A0 =A0 =A0 =A0 =A0 |/ =A0 V(+)---|+/
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 |/
>
> It=92s a natural addition to the single opamp difference amp.
> (boost up V(-) rather than throw away V(+). (G =3D 1 + R2/R1)
>
> The output of the first opamp U1 bothers me a bit. =A0As it has to
> handle the current from U2=92s gain. =A0I=92d like to run this up at high
> frequency, and I=92m worried about the output impedance of U1 screwing
> up the CMRR. =A0My current thought was to keep the gain low, and use
> R2=3D1k and R1=3D 10k. That should help minimize output impedance effects=
.
>
> Anyone used this configuration?
>
> Thanks,
> George H.
>
> *I think I really did reinvent it... but then seeing it in AofE I
> wonder if the back of my brain is just remembering it.

This circuit stinks, for anything that wants to be fast.  The V(-)
input has to go through two opamps, with the output delayed about
100ns from the V(+).  (opamps was opa2134)

George H.
```
```George Herold a &#2013265929;crit :
> On May 23, 9:52 am, George Herold <gher...@teachspin.com> wrote:
>>  I reinvented the two opamp differential amp yesterday.
>> And then found it in AofE* (2nd Ed. figure 7.36).  Though I drew it
>> differently.
>>
>>   +-R2--+--R1--+
>>   |     | |\U1 |    +--R2--+
>>  gnd    +-|-\  |    | |\U2 |
>>           |  >-+-R1-+-|-\  |
>>  V(-)-----|+/         |  >-+--out
>>           |/   V(+)---|+/
>>                       |/
>>
>> It&#2013266066;s a natural addition to the single opamp difference amp.
>> (boost up V(-) rather than throw away V(+). (G = 1 + R2/R1)
>>
>> The output of the first opamp U1 bothers me a bit.  As it has to
>> handle the current from U2&#2013266066;s gain.  I&#2013266066;d like to run this up at high
>> frequency, and I&#2013266066;m worried about the output impedance of U1 screwing
>> up the CMRR.  My current thought was to keep the gain low, and use
>> R2=1k and R1= 10k. That should help minimize output impedance effects.
>>
>> Anyone used this configuration?
>>
>> Thanks,
>> George H.
>>
>> *I think I really did reinvent it... but then seeing it in AofE I
>> wonder if the back of my brain is just remembering it.
>
> This circuit stinks, for anything that wants to be fast.  The V(-)
> input has to go through two opamps, with the output delayed about
> 100ns from the V(+).  (opamps was opa2134)
>
> George H.

How much fast is fast?

See AD8129/8130 for a single opamp solution...

--
Thanks,
Fred.
```
```On May 24, 2:37=A0am, Fred Bartoli <" "> wrote:
> George Herold a =E9crit :
>
>
>
>
>
> > On May 23, 9:52 am, George Herold <gher...@teachspin.com> wrote:
> >> =A0I reinvented the two opamp differential amp yesterday.
> >> And then found it in AofE* (2nd Ed. figure 7.36). =A0Though I drew it
> >> differently.
>
> >> =A0 +-R2--+--R1--+
> >> =A0 | =A0 =A0 | |\U1 | =A0 =A0+--R2--+
> >> =A0gnd =A0 =A0+-|-\ =A0| =A0 =A0| |\U2 |
> >> =A0 =A0 =A0 =A0 =A0 | =A0>-+-R1-+-|-\ =A0|
> >> =A0V(-)-----|+/ =A0 =A0 =A0 =A0 | =A0>-+--out
> >> =A0 =A0 =A0 =A0 =A0 |/ =A0 V(+)---|+/
> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 |/
>
> >> It=92s a natural addition to the single opamp difference amp.
> >> (boost up V(-) rather than throw away V(+). (G =3D 1 + R2/R1)
>
> >> The output of the first opamp U1 bothers me a bit. =A0As it has to
> >> handle the current from U2=92s gain. =A0I=92d like to run this up at h=
igh
> >> frequency, and I=92m worried about the output impedance of U1 screwing
> >> up the CMRR. =A0My current thought was to keep the gain low, and use
> >> R2=3D1k and R1=3D 10k. That should help minimize output impedance effe=
cts.
>
> >> Anyone used this configuration?
> >> Or other thoughts/ comments?
>
> >> Thanks,
> >> George H.
>
> >> *I think I really did reinvent it... but then seeing it in AofE I
> >> wonder if the back of my brain is just remembering it.
>
> > This circuit stinks, for anything that wants to be fast. =A0The V(-)
> > input has to go through two opamps, with the output delayed about
> > 100ns from the V(+). =A0(opamps was opa2134)
>
> > George H.
>
> How much fast is fast?

Grin... that's always a good question, short answer is better than I
have now.  I'd like to do ~100ns pulses.  But making things even
faster might be interesting.
>
> See AD8129/8130 for a single opamp solution...

Thanks I'll give those a look see.  I was looking at the AD8429
instument amp yesterday.  12 MHz at unity gain, but all sorts of
ringing on the step response.

George H.
>
> --
> Thanks,
> Fred.- Hide quoted text -
>
> - Show quoted text -

```
```On Thu, 24 May 2012 06:06:52 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

>On May 24, 2:37&#2013266080;am, Fred Bartoli <" "> wrote:
>> George Herold a &#2013265929;crit :
>>
>>
>>
>>
>>
>> > On May 23, 9:52 am, George Herold <gher...@teachspin.com> wrote:
>> >> &#2013266080;I reinvented the two opamp differential amp yesterday.
>> >> And then found it in AofE* (2nd Ed. figure 7.36). &#2013266080;Though I drew it
>> >> differently.
>>
>> >> &#2013266080; +-R2--+--R1--+
>> >> &#2013266080; | &#2013266080; &#2013266080; | |\U1 | &#2013266080; &#2013266080;+--R2--+
>> >> &#2013266080;gnd &#2013266080; &#2013266080;+-|-\ &#2013266080;| &#2013266080; &#2013266080;| |\U2 |
>> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | &#2013266080;>-+-R1-+-|-\ &#2013266080;|
>> >> &#2013266080;V(-)-----|+/ &#2013266080; &#2013266080; &#2013266080; &#2013266080; | &#2013266080;>-+--out
>> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; |/ &#2013266080; V(+)---|+/
>> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; |/
>>
>> >> It&#2013266066;s a natural addition to the single opamp difference amp.
>> >> (boost up V(-) rather than throw away V(+). (G = 1 + R2/R1)
>>
>> >> The output of the first opamp U1 bothers me a bit. &#2013266080;As it has to
>> >> handle the current from U2&#2013266066;s gain. &#2013266080;I&#2013266066;d like to run this up at high
>> >> frequency, and I&#2013266066;m worried about the output impedance of U1 screwing
>> >> up the CMRR. &#2013266080;My current thought was to keep the gain low, and use
>> >> R2=1k and R1= 10k. That should help minimize output impedance effects.
>>
>> >> Anyone used this configuration?
>> >> Or other thoughts/ comments?
>>
>> >> Thanks,
>> >> George H.
>>
>> >> *I think I really did reinvent it... but then seeing it in AofE I
>> >> wonder if the back of my brain is just remembering it.
>>
>> > This circuit stinks, for anything that wants to be fast. &#2013266080;The V(-)
>> > input has to go through two opamps, with the output delayed about
>> > 100ns from the V(+). &#2013266080;(opamps was opa2134)
>>
>> > George H.
>>
>> How much fast is fast?
>
>Grin... that's always a good question, short answer is better than I
>have now.  I'd like to do ~100ns pulses.  But making things even
>faster might be interesting.
>>
>> See AD8129/8130 for a single opamp solution...
>
>Thanks I'll give those a look see.  I was looking at the AD8429
>instument amp yesterday.  12 MHz at unity gain, but all sorts of
>ringing on the step response.
>
>George H.
>>

AD8130 is a great part, 270 MHz and very clean. Just be careful not to
exceed the maximum 2.5 volts diff input. It *will* clip just a bit
past there.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
```
```On 23.5.12 4:52 , George Herold wrote:
>   I reinvented the two opamp differential amp yesterday.
> And then found it in AofE* (2nd Ed. figure 7.36).  Though I drew it
> differently.
>
>    +-R2--+--R1--+
>    |     | |\U1 |    +--R2--+
>   gnd    +-|-\  |    | |\U2 |
>            |>-+-R1-+-|-\  |
>   V(-)-----|+/         |>-+--out
>            |/   V(+)---|+/
>                        |/
>
> Thanks,
> George H.

classical instrumentation amp.

--

Tauno Voipio

```
```On May 24, 10:18=A0am, John Larkin
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Thu, 24 May 2012 06:06:52 -0700 (PDT), George Herold
>
>
>
>
>
> <gher...@teachspin.com> wrote:
> >On May 24, 2:37=A0am, Fred Bartoli <" "> wrote:
> >> George Herold a =E9crit :
>
> >> > On May 23, 9:52 am, George Herold <gher...@teachspin.com> wrote:
> >> >> =A0I reinvented the two opamp differential amp yesterday.
> >> >> And then found it in AofE* (2nd Ed. figure 7.36). =A0Though I drew =
it
> >> >> differently.
>
> >> >> =A0 +-R2--+--R1--+
> >> >> =A0 | =A0 =A0 | |\U1 | =A0 =A0+--R2--+
> >> >> =A0gnd =A0 =A0+-|-\ =A0| =A0 =A0| |\U2 |
> >> >> =A0 =A0 =A0 =A0 =A0 | =A0>-+-R1-+-|-\ =A0|
> >> >> =A0V(-)-----|+/ =A0 =A0 =A0 =A0 | =A0>-+--out
> >> >> =A0 =A0 =A0 =A0 =A0 |/ =A0 V(+)---|+/
> >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 |/
>
> >> >> It=92s a natural addition to the single opamp difference amp.
> >> >> (boost up V(-) rather than throw away V(+). (G =3D 1 + R2/R1)
>
> >> >> The output of the first opamp U1 bothers me a bit. =A0As it has to
> >> >> handle the current from U2=92s gain. =A0I=92d like to run this up a=
t high
> >> >> frequency, and I=92m worried about the output impedance of U1 screw=
ing
> >> >> up the CMRR. =A0My current thought was to keep the gain low, and us=
e
> >> >> R2=3D1k and R1=3D 10k. That should help minimize output impedance e=
ffects.
>
> >> >> Anyone used this configuration?
> >> >> Or other thoughts/ comments?
>
> >> >> Thanks,
> >> >> George H.
>
> >> >> *I think I really did reinvent it... but then seeing it in AofE I
> >> >> wonder if the back of my brain is just remembering it.
>
> >> > This circuit stinks, for anything that wants to be fast. =A0The V(-)
> >> > input has to go through two opamps, with the output delayed about
> >> > 100ns from the V(+). =A0(opamps was opa2134)
>
> >> > George H.
>
> >> How much fast is fast?
>
> >Grin... that's always a good question, short answer is better than I
> >have now. =A0I'd like to do ~100ns pulses. =A0But making things even
> >faster might be interesting.
>
> >> See AD8129/8130 for a single opamp solution...
>
> >Thanks I'll give those a look see. =A0I was looking at the AD8429
> >instument amp yesterday. =A012 MHz at unity gain, but all sorts of
> >ringing on the step response.
>
> >George H.
>
> AD8130 is a great part, 270 MHz and very clean. Just be careful not to
> exceed the maximum 2.5 volts diff input. It *will* clip just a bit
> past there.
>
> --
>
> John Larkin =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0Highland Technology Incwww=
.highlandtechnology.com=A0 jlarkin at highlandtechnology dot com
>
> Precision electronic instrumentation
> Picosecond-resolution Digital Delay and Pulse generators
> Custom timing and laser controllers
> Photonics and fiberoptic TTL data links
> VME =A0analog, thermocouple, LVDT, synchro, tachometer
> Multichannel arbitrary waveform generators- Hide quoted text -
>
> - Show quoted text -

Roger, no more than 2.5 Volts.
Differential receiver amps are on order.
Half the problem is knowing what to search for.

Hey, I got the active quench to work!

+Vbias
|
R=3D1k
|
+---+----+
|   |    |
10k  L   C ~10pF
|   E    |    +--R---+ R=92s =3D 10k
|/   D    |    | |\   |
+-|    |    +-R--+-|-\  |
| |\>  |           |  >-+---> to comparator
|   |  +------R--+-|+/
|  GND           | |/ opa134
|                R
+-<quench        |
pulse from    GND
comparator

There are some =91tweaker=92 C=92s (5.6pF) on the feedback R and R to groun=
d
of non-inverting input.  These gave the diff amp a nice pulse
response.

Here=92s a scope shot.

Chan. 1 is diff amp output
chan. 2 is the =91quench=92 pulse
chan 3 is the voltage change at the top of the LED. (AC coupled)

I got to use the four channel =91scope.
Things could certainly be faster.
But it works!

George H.
```
```On May 24, 11:47=A0am, Tauno Voipio <tauno.voi...@notused.fi.invalid>
wrote:
> On 23.5.12 4:52 , George Herold wrote:
>
> > =A0 I reinvented the two opamp differential amp yesterday.
> > And then found it in AofE* (2nd Ed. figure 7.36). =A0Though I drew it
> > differently.
>
> > =A0 =A0+-R2--+--R1--+
> > =A0 =A0| =A0 =A0 | |\U1 | =A0 =A0+--R2--+
> > =A0 gnd =A0 =A0+-|-\ =A0| =A0 =A0| |\U2 |
> > =A0 =A0 =A0 =A0 =A0 =A0|>-+-R1-+-|-\ =A0|
> > =A0 V(-)-----|+/ =A0 =A0 =A0 =A0 |>-+--out
> > =A0 =A0 =A0 =A0 =A0 =A0|/ =A0 V(+)---|+/
> > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|/
>
> > Thanks,
> > George H.
>
> classical instrumentation amp.

Grin,  Yeah, but then I'd have to stick another DIP on my copper
clad.  I'm mostly a one chip wonder.  Two IC's and I start to become
confused :^)

George H.

>
> --
>
> Tauno Voipio

```
```On Thu, 24 May 2012 09:52:03 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

>On May 24, 11:47&#2013266080;am, Tauno Voipio <tauno.voi...@notused.fi.invalid>
>wrote:
>> On 23.5.12 4:52 , George Herold wrote:
>>
>> > &#2013266080; I reinvented the two opamp differential amp yesterday.
>> > And then found it in AofE* (2nd Ed. figure 7.36). &#2013266080;Though I drew it
>> > differently.
>>
>> > &#2013266080; &#2013266080;+-R2--+--R1--+
>> > &#2013266080; &#2013266080;| &#2013266080; &#2013266080; | |\U1 | &#2013266080; &#2013266080;+--R2--+
>> > &#2013266080; gnd &#2013266080; &#2013266080;+-|-\ &#2013266080;| &#2013266080; &#2013266080;| |\U2 |
>> > &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080;|>-+-R1-+-|-\ &#2013266080;|
>> > &#2013266080; V(-)-----|+/ &#2013266080; &#2013266080; &#2013266080; &#2013266080; |>-+--out
>> > &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080;|/ &#2013266080; V(+)---|+/
>> > &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080;|/
>>
>> > Thanks,
>> > George H.
>>