Read back. The Frozen One seems to have settled on PWL (which is the right way to go if it's good enough), but everyone felt the need to expand on my quadratic/processor suggestion.
Go figure.
-- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Fixed up for single +15V supply...
Newsgroups: alt.binaries.schematics.electronic Subject: Re: OpAmp Circuit (S.E.D) - PWL-Example-SingleSupply.pdf Message-ID:
...Jim Thompson
-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |
FrozenOne wrote: > I am looking to map voltages from one point to another, in a specific > maner. For example: >
This is almost certainly an exponential transform. Assuming that your numbers are exact- which I doubt very seriously- the transform Vout=2.8+0.9*exp(2.9*(Vin-11.7)) will fit with less than 1% error in the middle, and nearly exactly at endpoints, and this does make some sense for transforming between PIN diode configurations-loosely. Given your background, it would be very unwise to attempt this in any other way than a piecewise function generator, the above expression can/should be used as the formula for interpolating between your measured reference points at Vin=11.5, 12, and 12.5V. Obviously, the transform tends to grow significantly faster in the interval 12
I read in sci.electronics.design that John Larkin wrote (in ) about 'Op-amp circuit....', on Sun, 9 Jan 2005:
Have you allowed for the time-zone?(;-)
-- 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
Two fingers, type by sight (gotta see the keys), lots of mistakes.
John
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
In article , John Larkin wrote: [...]
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
Or for the time it took to type the message. John may be a touch typest.
-- -- kensmith@rahul.net forging knowledge
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 Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
HFA1130 and its clones, but not especially suitable for slow stuff... expensive and low swing. An 1130 will come cleanly off either rail in about 1 ns.
That's it. Internal nodes will saturate, current sources may collapse, and it will take a long time to recover, and may do weird spikey things when it comes off the rail. Some chopamps can take seconds to recover. And if you clip one stage of an LM324, the other amps on the same die will glitch, even if they're happily in-range. But a 324 is hardly a "modern opamp."
Most ideal-diode circuits have some sort of glitch problem, too.
John
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 Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
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
In article , John Larkin wrote: [....]
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
In article , Jim Thompson wrote: [...]
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 ! --- ! ! ! ---/\/\--- GNDThis 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
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.