Op-Amp Stacking

[MetalUpYerEye]

I've heard people mention a little about op-amp "stacking." What exactly is it and what does it do?

[JHS]

Details about opamp-stacking can be found on the the Barber website, the advantage is, you get more current on the output
I've tried it and dropped the idea after a test. For me 2 identical preamp circuits wired parallel and an active splitter is the better way to go. The lower input impedance and the missing load ballance is the most disadvantage on stacked opamps often yielding in a muddy sound.

JHS

[rockgardenlove]

You do just that...stack them.  Put one right on top of the other.  Try it, it can change the tone for the better or worse.  In my experience it usually makes it sound somewhat smoother.

[MetalUpYerEye]

I heard it can be used to alternate between 2 or more op-amps in a circuit via a switch (ie; a tubescreamer with op-amp switching). Besides that does it just affect the tone of the circuit? Can higher gain be achieved by using 2 stacked op-amps as opposed to 1 of the same kind? For example; what would happen if I stacked 2 386's together in a Ruby or 2 4558's in a Tubescreamer?

[Mark Hammer]

Parallelling is something derived from the design of power stages, where one might have multiple power transistors or power tuibes in parallel so as to be able to deliver more current to the speakers (and more wattage as a result).  It also is regularly used for things like headphone amps where the outputs of two op-amps are paralleled to provide enough power/current to the headphones to be audible.

The thing is that when those sorts of elements are paralleled, interaction between their device inputs is not an issue because of how they are used.  Although one CAN simply piggyback op-amps (of the same type/category - you can't piggyback a dual on top of a single) by soldering their pins together, and I've done it with interesting results, one of the problems or byproducts is that the input impedances of the devices are now placed in parallel.  This didn't dawn on me at first when I paralleled a pair of NE5534s in my Distortion+ and noticed a certain "timidness" to the high end (and maybe that is what rockgardenlove means by "smoother").  Then, someone noted here the other day that the input impedance of the Signetics NE553x series op-amps was pretty low, and I realized that placing two such low-impedance inputs in parallel was unlikely to be a great idea for a circuit intended for a single op-amp to serve as input and output stage - often as the first thing the guitar sees.

Having said that, placing a high-current-delivery/low-input-impedance op-amp in parallel with a lower-current-delivery/high-input-impedance opamp should give the best of both worlds.

[343 Salty Beans]

Sorry to revive such an old thread, but I have a question about this. I don't suppose that stacking the opamps will increase headroom? I have a Paralooper that distorts either side of the blend dual-opamp (a 4558, actually) in a horrible buzzy way when the blend pot is turned too far to either side. Will opamp stacking work?

I don't think so. But in that case, I assume that the reason it's distorting is that my signals are too hot running into both sides of the opamp (while my pickups are passive, they're pretty high-output). If that's the case, would it make sense to stack another dual opamp on top of my TL072 to give it more boost on the way out, so I could bring it up to unity gain easier?

[bajaman]

I got some interesting results from stacking an OPA2134 with an LM1458. The crisp attack and dynamics of the OPA2134, with the smoooooth sustain of the LM1458. The current drain will of course be higher but well worth experimenting with disimilar types of opamps (as long as they are both single or both dual OR both quad opamps!!!!!!).
Cheers
Steve

[dirk]

opamp stacking also reduces the noise floor. So headroom is increased. Max amount is 6dB.
Not realy worth it imo.

[WGTP]

I soldered an 8-pin socket on top an op amp to experiment, and now I don't remember which op amp was on the bottom!

[R.G.]

I don't suppose that stacking the opamps will increase headroom?

Not in the way you're talking about it. It may increase signal to noise, but it does not increase headroom.

Increasing headroom takes more volts as you're describing it. Paralleling an opamp might increase the output voltage a fraction of a volt by loading the the opamps sharing the load. To get a noticeable (e.g. 6db) increase in headroom, you need double the output voltage, and thus a bit more than double the power supply voltage.

[MKB]

FWIW, my coworker who has been working with hi-fi audio construction and design for decades, told me that they used to stack op amps all the time in early circuitry, it was a very common practice.  But he couldn't recall exactly why.  He also sold all his old Amateur Audio magazines, which probably had articles on stacking.  Oh well...

[db]

We've made headphone amps using paralled op-amp outputs.  Works OK.  I can't imagine why anyone would do it for effect?
Would someone care to explain how the noise floor is reduced?  I haven't come across this before.  I would have thought that overall noise would be increased.

[R.G.]

Adding active devices and/or resistors will always increase noise. However, paralleling devices so they can drive loads at lower impedances can add less noise than would otherwise be there.

Noise always adds in a root-mean-square per bandwidth way. If one component has X noise in Y bandwidth, two of these will have 6db more noise, and four of them will have 12db more noise. But the input or output impedances may be better matched to the source/load impedances such that this is lower noise than you'd have with the same gain and single devices, because noise is impedance sensitive as well.

This is very far afield from anything needed in effects work. It's more in line with the kind of amplifiers you'd need to amplify moving coil phonograph cartridges. (yeah, I know. "phonograph?? What's that??")

Did you know that 22,476,832 angels can dance on the head of a pin? 8-)

[Tony Forestiere]

Nowadays only four angels can dance there. Formerly there was no limit, but OSHA passed the Angel Safety Law recently, which also requires that the pin must be inspected twice each year for structural defects.

[Joe Hart]

And just like that we have a 5-year-old thread reborn!  :-)
-Joe Hart