OpAmp Gain Values Revisited

(snip)

I haven't followed the story very well, up till now, so I don't know what NS app notes you refer to and what the middle amplifier device is.

Hi, Mark. For Mr. Popelish's reference, the earlier thread was

and the National Semiconductor appnote was AN-31, "Op Amp Circuit Collection". That's the AN consisting of basic circuit schematics for most of the basic op amp configurations, along with a relevant equation or two for each.

As shown in your ASCII art above, the first op amp you've drawn is set up as a diff amp with a gain of 1 * (V2 - V1). The second op amp is set up as a non-invering amplifier with gain of 1 + (R6 / R5), or 2.

R7 just seems to be a plain old 10K load resistor, which should not be anywhere near enough to bog down the op amp output. Even with the second op amp railed at V+, it will only draw about 1mA. But R7 doesn't change the basic gain equation of the op amp. I believe it's there because the NPN transistor internal to the LM833 output is a lot better at sourcing current than the PNP is at sinking current. See the data sheet. Keeping the output in a current sourcing mode will tend to improve slew rate, and reduce crossover distortion issues.

The third op amp is a voltage follower (a non-inverting amplifier with a gain of 1). The 50 ohm output resistor at the third op amp output may be part of a voltage divider, or it may just be made so that the amplifier output impedance is about 50 ohms (51 ohms is the closest 5% value). The thing is, the 51 ohm resistor may be acting with the input impedance of the signal destination to create a voltage divider, which may be slightly reducing your output. Either way, you're right -- if you want a gain of 1 for the whole block, you'll have to adjsut values on the diff amp resistors.

By the way, I would also check the exact values and particularly the matching on the resistors in the diff amp. Small changes here can make a proportionately larger change in the output (see the equations). Use

1% resistors, and use an ohmmeter to get a match in the input resistors and the feedback resistors. Diff amps don't work very well without adequate resistor matching.

Moral of the story -- even techs who just work with the circuits should take the time to look at the data sheets. They have good information for everybody, from design engineers to hobbyists.

Good luck Chris

I see they used a 2k pull down load resistor for the distortion and slew rate graphs. I wonder how much worse the result would be without it.

I still don't see this 3 stage amplifier in any reference. Why not leave the middle amplifier out, entirely? Can't get less distortion from it than that way.

Good morning, Mr. Popelish. From what the OP has said, I'm tending to think he's working on somebody else's homebrew stuff (poor guy -- no schematics, no help at all other than muscling out the circuit schematic from board inspection). I had mentioned in the prior post that someone who has no design experience, but just wanted to have a quick "cheat sheet" for elementary op amp circuits on the bench could do worse than printing out AN-31, which is just sketches of basic op amp circuits with gain equations added for clarity. You would look at the individual op amps, then figure out what each one does (diff amp, non-inverting amp, voltage follower). That might be some help for the newbie.

You're right of course, the middle amp of the three does look like an advanced case of "clean your plate" syndrome. It's actually a good indication the original circuit was done by a hobbyist. The LM833 _is_ a dual, after all. ;-)

Cheers Chris

Chris, you hit that one dead on :-) I just went in and created a schematic by reverse engineering the layout with a meter. Man, those SM resistors are hard to read :-)

That has come in really handy.

Well, I was going to ask about that, then I figured I would not ask for too much 'free' help, :-) but I wondered about the middle stage, and figured it was for loading purposes in order to attain four outputs from the single MC33708 input.

Can the one MC33708 drive four LM833's? I can do digital loading analysis, but never have gone into the concept using opamps, and am not sure what parameters I would have to deal with.

Thanks very much, again.

John

|\\| LM833 -|-\\ ----| >------- | -|+/ | |/| | LM833 | | |\\| | |\\| LM833 -|-\\ | -|-\\ - - | >----- | >------- | -|+/ -|+/ |\\| | |/| |/| -|-\\ | | >--| -|+/ | LM833 |/| | | |\\| |\\| LM833 | -|-\\ -|-\\ MC33708 |-- | >-- -- --| >------- -|+/ | -|+/ |/| | |/| | | | | |\\| LM833 | -|-\\ --- | >------ -|+/ |/| (created by AACircuit v1.28.4 beta 13/12/04

(created by AACircuit v1.28.4 beta 13/12/04

It is a 'one in' to 'four out' box, so I assumed it was a loading issue that required the middle stage of the circuit.

Can the MC33708 drive four inputs of the LM833?

An ideal op amp has infinite input impedance. Of course, none are perfect, but nearly all standard op amps have very light loading inputs. The datasheet for the LM833 says the input bias current is typically 1/2 microamp. I would guess you could actually drive dozens of just about any op amp inputs from a single op amp output, or hundreds if the signal was slow enough that input capacitance and lumped transmission impedances weren't a problem. In general, "unit loading" just isn't a problem.

Do yourself a favor -- take a good look at the manufacturer data sheets. They provide a lot of really good information that, once you decode the terms, is very useful in using the little beasties.

Don't worry too much about wearing out your welcome -- honest questions are always welcome here, and are frequently answered elegantly and well by several individuals. (Homework questions unaccompanied by offers of payment usually aren't. I believe the unofficial going rate is 20 zloty per clue. But even just giving them teh answers to their homework questions really doesn't help them. Darwin would have postulated that the potential techs and engineers who are capable of surviving their classes are the ones who will be able to do the job. The ones who find their way around Darwin usually aren't a good fit in electronics, and even if they don't electrocute themselves, would have been better off transferring majors to journalism. In that profession, actual facts and analytical thought are irrelevant.)

Also, you may wish to use google groups with their search engine to skim through questions in this group over the last several years relating to your question. Typically, if the question was answered well 10 years ago, the answer is still good today. You'll be able to learn a lot more that way.

Mr. Popelish in particular provides a lot of attention to detail, is almost always able to check the newsgroup daily, and his answers are always worth reading.

Good luck Chris