Varying Vref current: Sonically useful?

[Mark Hammer]

After posting something about deriving Vref for op-amps on another thread, and after reading the Vex interview in the Dave Hunter book, where he discusses the Stab control and "dying battery" mimickry on the Fuzz Factory, I started wondering about the way that people swap dual op-amps in and out, and the resistor values used to derive Vref currents in different circuits.  

In plenty of cases, you'll see a pair of 10k resistors and medium value cap for smoothing.  In the Distortion+, those resistors are ONE BLOODY MEG.  Surely to goodness the bias current has something to do with the workings of that particular circuit.  Perhaps because it is optimal for THAT op-amp, or perhaps because it is intended to bring something out in that op-amp that might require different values to bring out in another.

Either way, I offer it up as an idea for exploring to use a dual-ganged pot for systematically varying the Vref current to op-amps.  For example, start with a dual-ganged 500k pot, wired up as variable resistor, with each section in series with a 10k "default" resistor.

The other thing I wonder about is how much of people's claim that such and such an op-amp really livened up an effect while another one sounded awful, has more to do with the match between needed bias current and the  op-amp, rather than just the op-amp itself.  It is hard to imagine any single bias current would be ideal for any and every op-amp regardless of category.  This was illustrated in no uncertain terms by what Jack Orman had to do to make non-1458 opamps behave in the original Dr. Q circuit.

[Dragonfly]

obviously im a "beginner" at opamps, but in my early experiments ive tried hooking up various resistors in series with the 9v input to the opamp in order to "starve" the input voltage a bit, and noticed that it GREATLY affects the sound. if you start with something simple, like a 18v feed, and hook up a pot as a variable resistor, you can get a huge range of sounds just from varying the voltage. keep in mind that these experiments were conducted on LM386 opamps, though i'd guess that the effect is similar with most opamps...


now as far as the "bias" of the opamp, i would think that it would react similar to the way a transistor reacts to different biasing voltages...though im still new at the opamp thing...

interesting post...i'll definitely be watching this thread

RAWK ON !

[Dragonfly]

The other thing I wonder about is how much of people's claim that such and such an op-amp really livened up an effect while another one sounded awful, has more to do with the match between needed bias current and the  op-amp, rather than just the op-amp itself.  It is hard to imagine any single bias current would be ideal for any and every op-amp regardless of category.  This was illustrated in no uncertain terms by what Jack Orman had to do to make non-1458 opamps behave in the original Dr. Q circuit.

interesting thought....id bet that your theory is at least partially correct...

though opamps should have different "sounds", just as transistors do. even similarly performing, similarly constructed transistors can sound different.....but you already know that

[gez]

In the Distortion+, those resistors are ONE BLOODY MEG.  Surely to goodness the bias current has something to do with the workings of that particular circuit.  Perhaps because it is optimal for THAT op-amp, or perhaps because it is intended to bring something out in that op-amp that might require different values to bring out in another

With the dist + the values are high to provide high z in.  1M resistors are about the max you can get away with before starving the inputs of the amp used (if I recall) and the 1M in the feedback loop helps correct any offset.  You'd expect high value resistors for an input stage.

Either way, I offer it up as an idea for exploring to use a dual-ganged pot for systematically varying the Vref current to op-amps.  For example, start with a dual-ganged 500k pot, wired up as variable resistor, with each section in series with a 10k "default" resistor.

So many people use FET input amps for quietness (and this does make a difference) that I don't think the pot would do that much unless you set out to use an amp with bipolar inputs, in which case the pot would probably create an offset (though many moderns bipolar input amps can have pretty miniscule input bias currents so it may not be that much of an offset).  In a circuit with diodes in the feedback loop it's probably not going to do much sonically unless you really starve the inputs.  Back-to-backs at the output though, who knows?  The offset might cause some asymettry, or even correct the asymettry of a particular amp's output swing.

The other thing I wonder about is how much of people's claim that such and such an op-amp really livened up an effect while another one sounded awful, has more to do with the match between needed bias current and the  op-amp, rather than just the op-amp itself.  It is hard to imagine any single bias current would be ideal for any and every op-amp regardless of category.  This was illustrated in no uncertain terms by what Jack Orman had to do to make non-1458 opamps behave in the original Dr. Q circuit

If the inputs are being pulled to ground most bipolar input devices are going to cut off, which is the problem with the DrQ as far as I'm aware.  So input current varying from one type of amp to another isn't really the issue in this case, more the design of the input stages.

The orininal 741s were noisy so any modern amp is an improvement.  Some amps sound different in circuits due to things like gate capacitance if large value resistors are used.