Op-amp circuit....

[John Larkin]

On Mon, 10 Jan 2005 04:25:59 +0000 (UTC), kensmith@green.rahul.net
(Ken Smith) wrote:

In article <d513u0dsgphuk18dd9q7fe63vg90dj4a5s@4ax.com>,
John Larkin <john@spamless.usa> wrote:
[...]
Let the record show that I posted it 33 minutes before Ken did! So how
come it wasn't *my* nice idea?

Can you prove that it wasn't first thought of by Mr. Armstong or Mr.
Tesla?

If so, you have good ground for just such a claim to priority.

Oh, I'm sure somebody thought of it before we did, but it's unlikely
that it was Mr A or Mr T. They were more RF guys.

I'm just *so* terribly disappointed that Fred didn't value my
contribution.


But the glitch problem is still serious. Nearly all opamps wind up
like hell when they go open-loop.

John

 

[Ken Smith]

In article <d513u0dsgphuk18dd9q7fe63vg90dj4a5s@4ax.com>,
John Larkin <john@spamless.usa> wrote:
[...]

Let the record show that I posted it 33 minutes before Ken did! So how
come it wasn't *my* nice idea?

Can you prove that it wasn't first thought of by Mr. Armstong or Mr.
Tesla?

If so, you have good ground for just such a claim to priority.

--
--
kensmith@rahul.net forging knowledge

 

[Ken Smith]

In article <S6SxYnF$NZ4BFwP+@jmwa.demon.co.uk>,
John Woodgate <noone@yuk.yuk> wrote:

I read in sci.electronics.design that John Larkin <john@spamless.usa
wrote (in <d513u0dsgphuk18dd9q7fe63vg90dj4a5s@4ax.com&gt about 'Op-amp
circuit....', on Sun, 9 Jan 2005:
Let the record show that I posted it 33 minutes before Ken did! So how
come it wasn't *my* nice idea?

Have you allowed for the time-zone?(;-)

Or for the time it took to type the message. John may be a touch typest.

--
--
kensmith@rahul.net forging knowledge

 

[John Woodgate]

I read in sci.electronics.design that Tim Wescott
<tim@wescottnospamdesign.com> wrote (in <10u3hcs31fvthcb@corp.supernews.
com&gt about 'Op-amp circuit....', on Sun, 9 Jan 2005:

OK, why am't I looking at an AVR selection chart?

No, "amn't".
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[Ken Smith]

In article <4224u0p20pn5cdmv2p691jgk34t4ss5t1h@4ax.com>,
John Larkin <john@spamless.usa> wrote:
[....]

But the glitch problem is still serious. Nearly all opamps wind up
like hell when they go open-loop.

I don't know of any I.C. op-amps that don't have the problem to some
degree. On some externally compensated ones, you can add a circuit with a
schottky to remove the wind up but the swing is reduced in the process.

If you use a LM339 comparitor as an op-amp, you compensate it by placing a
capacitor on the output. That would yeld an op-amp with almost no wind up
but it can only pull down. The frequency responce and slew rate will
really suck too.



--
--
kensmith@rahul.net forging knowledge

 

[John Popelish]

Jim Thompson wrote:

I'm puzzled by the use of the term "windup"? Most modern OpAmps will
do nothing when railed except, perhaps, be a little slow coming off of
the rail... similar to a little storage time in a BJT.

I think that overload recovery time is the windup they are talking
about. When it involves the compensation cap charging to a voltage
quite far from normal operation, it can be quite a while.

The term is used frequently in reference to PID controllers, applied
ot the integral function. If the integral of error is not halted when
the output saturates, then, when the error reverses, you have to wait
for it to integrate its way back toward zero before the output will
respond to the input reversal. Since the integral time constant (time
it takes for the integral term to produce as big a contribution to the
output as the proportional tern did, immediately) can be measured in
hours for some processes) waiting for the integral to unwind can be a
very obvious (and expensive) wait.

I seen no technical reason that anti windup could not be designed into
opamps so that they could be used with much higher performance in
saturating applications, except that op amp designers do not see to be
aware of these applications.

(hint)

--
John Popelish

 

[Jim Thompson]

On Mon, 10 Jan 2005 14:50:12 +0000 (UTC), kensmith@green.rahul.net
(Ken Smith) wrote:

In article <4224u0p20pn5cdmv2p691jgk34t4ss5t1h@4ax.com>,
John Larkin <john@spamless.usa> wrote:
[....]
But the glitch problem is still serious. Nearly all opamps wind up
like hell when they go open-loop.

I don't know of any I.C. op-amps that don't have the problem to some
degree. On some externally compensated ones, you can add a circuit with a
schottky to remove the wind up but the swing is reduced in the process.

If you use a LM339 comparitor as an op-amp, you compensate it by placing a
capacitor on the output. That would yeld an op-amp with almost no wind up
but it can only pull down. The frequency responce and slew rate will
really suck too.

I'm puzzled by the use of the term "windup"? Most modern OpAmps will
do nothing when railed except, perhaps, be a little slow coming off of
the rail... similar to a little storage time in a BJT.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Jim Thompson]

On Mon, 10 Jan 2005 11:30:58 -0500, John Popelish <jpopelish@rica.net>
wrote:

Jim Thompson wrote:

I'm puzzled by the use of the term "windup"? Most modern OpAmps will
do nothing when railed except, perhaps, be a little slow coming off of
the rail... similar to a little storage time in a BJT.

I think that overload recovery time is the windup they are talking
about. When it involves the compensation cap charging to a voltage
quite far from normal operation, it can be quite a while.

The term is used frequently in reference to PID controllers, applied
ot the integral function. If the integral of error is not halted when
the output saturates, then, when the error reverses, you have to wait
for it to integrate its way back toward zero before the output will
respond to the input reversal. Since the integral time constant (time
it takes for the integral term to produce as big a contribution to the
output as the proportional tern did, immediately) can be measured in
hours for some processes) waiting for the integral to unwind can be a
very obvious (and expensive) wait.

I seen no technical reason that anti windup could not be designed into
opamps so that they could be used with much higher performance in
saturating applications, except that op amp designers do not see to be
aware of these applications.

(hint)

An old, slow OpAmp might take 1.5us to get off of the rail (30pF
compensation capacitor, 30uA drive current, 1.5V "dead-head").

My suspicion, just verified with a quick simulation, is that the
integrator summing node comes un-glued when the OpAmp rails.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Ken Smith]

In article <ov65u0lmv1emol9bda6p7lede3or9ggl4q@4ax.com>,
Jim Thompson <thegreatone@example.com> wrote:
[...]

I'm puzzled by the use of the term "windup"? Most modern OpAmps will
do nothing when railed except, perhaps, be a little slow coming off of
the rail... similar to a little storage time in a BJT.

They are more than a little slow coming off the rail. For a given speed
of op-amp, they take quite a long time. The "wind up" is from the term
"integrator wind up" common to control systems. A example of how you can
be bit by this is:


+15V
Vin ! LT1498
AC--/\/\/\--------!+\
in ! >----+-/\/\/----+----
--!-/ ! !
! ! ! ---
! GND ! ---
! ! !
---/\/\--- GND


This looks like a 1/2 wave rectifier circuit. The LT1498 can follow the
signal when it is positive and stops at ground when it is negitive.
Although the LT1498 is good for about 6MHz, this circuit doesn't work at
200KHz.


--
--
kensmith@rahul.net forging knowledge