op amp substitution question

[plugger]

Can someone explain why substituting one op amp for another in a preamp stage (Musicman 130 210 combo, chassis 2275) would result in more gain through the preamp stage? I can't really square this with my understanding of how op amps work.

Specifically, I've been substituting LM307H op amps (which are the original spec for the circuit) with LF356N, which sound better, but also result in noticeably higher gain. I understand that any real op amp will vary from an ideal op amp in terms of frequency response characteristics, slew rate etc., but gain is pretty much gain... unless you were pushing the op amps to their gain limits, I would expect them to result in functionally very similar gain for a guitar audio signal.

So, for example, the -ve feedback resistors of r1 = 10K and r2 = 220K for op amp IC3 (channel 2) in that circuit should show result in a gain on the non-inverted input of 1 + (r2/r1) = 23. But substituting in the LF356N for the LM307H for this IC results in noticeably more gain.

This is just an example. Gain increases are also noticed when the LM307H ICs are subbed with LF356N in other stages of the preamp, e.g. IC4 and IC7.

Why would this be?

The schematic for this amp as a pdf can be found at http://www.musicmanamps.com/schematics/2475-130_and_2275-130.pdf in anyone is interested in the specifics of the circuit. The preamp section looks pretty straightforward to me, but there is always the possibility there's something I'm missing here.

BTW, the reason I've been experimenting with substitutions is because I quite like the sound of the of the LF356N -- a bit warmer and organic sounding to my ear than the stock LM307H. A little less noisy, too. But the extra gain is causing problems with the trem circuit (which is on channel 2), so I've gone back to the LM307H on channel 2 for the time being. But in any case, I'd like to understand the practicalities of op amps a bit more fully.

[Enzo]

Read the data sheets for the two parts:

http://www.ti.com/lit/ds/symlink/lm307-n.pdf

http://www.ti.com/lit/ds/symlink/lf357.pdf

Note in particular the much higher large signal voltage gain spec.


It is possible also that the JFET inputs on the 356 interact with the circuits differently from the 307's bipolars, but I don't really know about that.

[R.G.]

Can someone explain why substituting one op amp for another in a preamp stage [...] would result in more gain through the preamp stage? I can't really square this with my understanding of how op amps work. [...]

Why would this be?

It probably isn't. At least it isn't until you look more closely at the word "gain".

If by "gain" what you really mean is "distortion", like many people casually and incorrectly use the word, you probably do get more distortion. This can easily happen if the opamps happen to have different degrees to which the outputs can approach the supply rails. An opamp that can only go to within 3V of its power supply rails will distort sooner than one that can approach the rails to within a volt, and be more distorted with the same signal voltage an power supply voltages. And it will sound different to one that can go to within 0.3V of the negative supply but only 3V of the positive supply, all of which exist as standard opamps with compatible pinouts.

[plugger]

Thanks guys.

RG, I did mean "gain" in the technical sense, rather than the informal sense. In the example above, the resistor values in the feedback would yield a gain of 23.

I have done some more reading, and I think I know the answer. Enzo is on the right track in looking at the open loop gain differences. Although, an open loop DC gain of 200,000 vs 160,000 isn't directly the issue when you are looking at a gain of 23. Rather, the critical difference in performance between the parts is looking at the AC characteristics -- what happens when you transition from DC gain to AC signal gain.

It turns out that for DC, both parts would behave fairly comparably for modest gain levels << open loop gain levels (e.g., 23 vs 160,000). However, if you look further down the data sheet, under "AC characteristics", and look at the attenuation in the open loop gain as frequency increases in the "open loop frequency response" plots, you see that even for modest audio range frequencies (e.g., 1000Hz) the open loop gain has dropped by something in the order of 30dB compared to DC. And the critical thing is that the plot for the LM301 part is even more attenuated, by about 10dB across all the audio frequencies. So the "effective" gain of the audio signal is about doubled for the LF356 compared to the LM301!

So it's not the open loop gain DC characteristics that make the difference in this application, it's the AC characteristics.

Anyway, some tweaking of those feedback resistor values would seem to be in order if I want to go over to the LF356 part as a sub for the LM307. And at least now I have at least the beginning of understanding why that needs to be done...

BTW, just a note on RG's informal sense of "gain": On a Musicman 210 HD130, there is no "gain" in that sense. The HD130 is the amp Leo Fender built because he just thought the Twin Reverb just wasn't loud enough. This amp does not break up at any (legal) volume level. (I once tried and got the UN WMD inspectors around knocking on my door with a security council resolution telling me to shut up or I'd be in violation of several international treaties. It was still pristine "blackface" clean at those levels -- as far as I could tell with the bleeding eardrums and all. ;-)
 

[Enzo]

Some earlier Marshall SS amps using 1458 op amps are unstable if you make the then common 4558 sub for the 1458.