plate@cdsmn.mn.org (Doug Plate) (03/07/91)
*----------------------------------------------------------------* I'm looking for a simple circuit that will clamp the output of a summing amp to 5 volts. The circuit uses an OP amp in a non-inverting summing configuration (unity gain) with 2 inputs, either of which can go from 0 - 5 volts, however the output must come up to 5 volts and stop (plus or minus about .009v) even if the sum of the two input > 5 volts. I've tried several different clamping techniques, but keep running into the same two limitations. One condition I run into is that I can clamp the output so that if the sum of the outputs > 5v, the output = 5v, however the problem is that a sum of exactly five volts in gives me an output a few hundered millivolts short of the 5v output I need. The second condition occurs when I clamp so that a 5v input sum produces a 5v ouput. Unfortunately in this case, the clamp doesn't kick in until my output is about 5.22v. So basically what I'm looking for is a clamp that allows the output to follow the input all the way up to 5v then clamp to 5v for anything above that. Also, I'm trying to keep things as simple as possible. Doug Plate plate@cdsmn.mn.org
robf@mcs213j.cs.umr.edu (Rob Fugina) (03/09/91)
In article <1991Mar06.181426.10500@cdsmn.mn.org> plate@cdsmn.mn.org (Doug Plate) writes: >I'm looking for a simple circuit that will clamp the output of a summing >amp to 5 volts. The circuit uses an OP amp in a non-inverting I'm sure I can't tell you exactly how to do this, but it seems to me this would be a good place to use a zener diode... Rob robf@cs.umr.edu
raoul@eplunix.UUCP (Nico Garcia) (03/10/91)
In article <1991Mar06.181426.10500@cdsmn.mn.org>, plate@cdsmn.mn.org (Doug Plate) writes: > *----------------------------------------------------------------* > I'm looking for a simple circuit that will clamp the output of a summing > amp to 5 volts. The circuit uses an OP amp in a non-inverting > summing configuration (unity gain) with 2 inputs, either of which > can go from 0 - 5 volts, however the output must come up to 5 volts > and stop (plus or minus about .009v) even if the sum of the two > input > 5 volts. Ummm, like this? ----------------- | |\ | ---|- \ | | \ | V1--------R1----| | \______| Vout |------| / V2--------R2----| | / | / |/ Why not just use a good 5 Volt supply op-amp, with rail-to-rail capability? -- Nico Garcia Designs by Geniuses for use by Idiots eplunix!cirl!raoul@eddie.mit.edu
zimmer@calvin.stanford.edu (Andrew Zimmerman) (03/10/91)
In article <1041@eplunix.UUCP> raoul@eplunix.UUCP (Nico Garcia) writes: >In article <1991Mar06.181426.10500@cdsmn.mn.org>, plate@cdsmn.mn.org (Doug Plate) writes: >> *----------------------------------------------------------------* >> I'm looking for a simple circuit that will clamp the output of a summing >> amp to 5 volts. The circuit uses an OP amp in a non-inverting >> summing configuration (unity gain) with 2 inputs, either of which >> can go from 0 - 5 volts, however the output must come up to 5 volts >> and stop (plus or minus about .009v) even if the sum of the two >> input > 5 volts. > >Ummm, like this? > ----------------- > | |\ | > ---|- \ | > | \ | >V1--------R1----| | \______| Vout > |------| / >V2--------R2----| | / > | / > |/ > >Why not just use a good 5 Volt supply op-amp, with rail-to-rail capability? > >-- > Nico Garcia > Designs by Geniuses for use by Idiots > eplunix!cirl!raoul@eddie.mit.edu I don't see how the above circuit solves the original posters question. I believe that Vout = (V1+V2)/2. (No unity gain) I don't see how it clamps, unless it makes use of the upper rail of the op-amp supply. How about the following circuit instead? --------R1----------R1-------- | | |\ | a --- ---| \ | --|>|-- +5 - | \ | | | \_____|_____|_____ Vout V1------R1----| | / | |------| / | / z V2------R1----| | / |--|<|--- |/ / | --- - This actually is two different methods. (a) The first just uses a diode. (The +5 reference should be closer to 4.3 to take into account the voltage drop on the diode.) The second method (z) uses a 5 volt zener diode to clamp the circuit. (I haven't done op-amps in a while, but I believe the above is correct.) Andrew zimmer@calvin.stanford.edu
brianr@tekig5.PEN.TEK.COM (Brian E Rhodefer) (03/11/91)
In article <1991Mar10.093752.18289@neon.Stanford.EDU> zimmer@calvin.stanford.edu (Andrew Zimmerman) writes: >I don't see how the above circuit solves the original posters question. (referring to Nico Garcia's posted circuit, not shown here) >I believe that Vout = (V1+V2)/2. (No unity gain) >I don't see how it clamps, unless it makes use of the upper rail of the >op-amp supply. > >How about the following circuit instead? > > --------R1----------R1-------- > | | |\ | a > --- ---| \ | --|>|-- +5 > - | \ | | > | \_____|_____|_____ Vout >V1------R1----| | / | > |------| / | / z >V2------R1----| | / |--|<|--- > | |/ / | > "C" --- > - > Two problems here: 1) Trying to control the output of an opamp by shorting it to a voltage source (even through a diode) is a bit brutal. 2) The original poster wanted fairly high clamping accuracy: 9mV out of 5V. Variations in clamping current and operating temperature will prevent this circuit from meeting that goal. I suggest that the following circuit be used to clamp the voltage at the noninverting input of Andrew's circuit: D1 r------------|>|---------, | | To Point | D2 | "C" <---*---R2---*---------|<|---* | |\ | ---|- \ | | \ | | \_____| +5V --R2----, | / *------|+ / r--R2----' | / | |/ --- - R2 should be >> R1. To compensate for the small current through D2 (which keeps this second opamp balanced even while it isn't clamping), a resistor of 2*R2 should be connected from +5V to the inverting input of the summing opamp. The clamping opamp should be chosen or trimmed to have an input offset voltage less than the allowable clamping error of nine millivolts. I tried sending this through email, but it probably didn't work, and I hadn't gotten the unity gain topology right, anyway. Brian Rhodefer
zimmer@calvin.stanford.edu (Andrew Zimmerman) (03/11/91)
> > D1 > r------------|>|---------, > | | >To Point | D2 | > "C" <---*---R2---*---------|<|---* > | |\ | > ---|- \ | > | \ | > | \_____| > +5V --R2----, | / > *------|+ / > r--R2----' | / > | |/ > --- > - > > >Brian Rhodefer In an earlier messge, Brian mentioned what was wrong with a circuit I proposed. He of course was right that it would be hard on the op-amp, and therefore you should not use it. His other point was that the original posting wanted it accurate to 9 mV. I wasn't aware of this. (oops) I was also under the wrong assumption that the circuit was not to use an additional op-amp. If that was wrong, I agree with the change of Brian's which is putting a precsion diode where it should have been in my circuit, at the input, with the clamp voltage either bing 2.5 (1.8) or with a 2.5 zener. Sorry about the previous post. Brian: Thanks for correcting my circuit. Andrew zimmer@calvin.stanford.edu PS. In my original post, I did say it was a while since I had done op-amps. :-)
sgm@otter.hpl.hp.com (Steve Methley) (03/11/91)
In article <1991Mar06.181426.10500@cdsmn.mn.org> plate@cdsmn.mn.org (Doug Plate) writes: > I'm looking for a simple circuit that will clamp the output of a summing > amp to 5 volts......... Why not use an op-amp whose output is designed to be clamped? You'll find that the RCA3140 has this feature, along with a few others I can't recall just now. Details in RCA databooks....... (A zener from pin 8 to ground, clamped o/p level is zener voltage less two diode drops) Do bear in mind that if you clamp the output of a feedback circuit you lose the effect of the feedback! Incidentally that's why the approach I detail is good, because it stops the o/p drivers saturating, regardless of input. This is important if speed/recovery from overdrive is an issue. Cheers, Steve. (personal opinions)
lewis@tramp.Colorado.EDU (LEWIS WILLIAM M JR) (03/12/91)
If you are fortunate enough to have a copy of the Philbrick Applications Manual (published circa 1965), look on page 23 for a complete discussion of clamping or bounding circuits. For all but the most demanding applications, a pair of matched zener diodes back to back in parallel with the feedback resistor work just fine.
R_Tim_Coslet@cup.portal.com (03/13/91)
In Article: <1991Mar10.093752.18289@neon.Stanford.EDU> zimmer@calvin.stanford.edu (Andrew Zimmerman) writes: >How about the following circuit instead? > > --------R1----------R1-------- > | | |\ | a > --- ---| \ | --|>|-- +5 > - | \ | | > | \_____|_____|_____ Vout >V1------R1----| | / | > |------| / | / z >V2------R1----| | / |--|<|--- > |/ / | > --- > - > > >This actually is two different methods. (a) The first just uses a diode. >(The +5 reference should be closer to 4.3 to take into account the voltage >drop on the diode.) > >The second method (z) uses a 5 volt zener diode to clamp the circuit. > >(I haven't done op-amps in a while, but I believe the above is correct.) You obviously havn't. NEVER directly clamp the output of an opamp... the opamp and the clamp will fight each other, and generally damage the opamp output transistors. The clamping circuit should always be in the "feedback" network. It would take longer for me to design and draw the appropriate circuit than I have while reading news, but the place to put the clamps are either across the feedback resistor (in an inverting amplifier) or from the noninverting opamp input to ground (in a noninverting amplifier like the one above). R. Tim Coslet Usenet: R_Tim_Coslet@cup.portal.com BIX: r.tim_coslet California could REALLY use 40 days and 40 nights of rain right now!