buffer steals my distortion!

[nepalnt21]

wanted to see what adding a buffer to the input of my breadboarded r.o.g. supreaux deux would do



so i used the voltage divider opamp buffer from here http://www.muzique.com/lab/buffers.htm

this one:


...

basically, i removed the 1M resistor to ground (im assuming i dont need it, since the impedance is set by the input of the buffer now), and placed the buffer right at the input of the signal, output goes to the first 100K resistor. i used two 2M resistors in the voltage divider, opamp is a NE 5532. 

when i bypass the opamp, the pedal distorts, but when i use the buffer, it seems to just suck ALL of the distortion away.

what the hecking heck?

[nepalnt21]

dag nabbit! i forgot about Vcc+/-

i also measured the two 2M resistors, and found two that were closer in value.

louder, but still, the distortion is much less.

i measured the drain voltage on each of the transistors, and they were severely out of whack. i had to turn the trimpots DRASTICALLY to bias them to 12v. weird.

[ElectricDruid]

You've got *no* DC bias.

If you remove the 1M resistor from the input of the Supreaux Deux, it no longer has any bias level at all at its input, since it is only connected by capacitors (one 10uF output cap from your buffer, and one 270pF cap to ground). Since caps pass no DC, they don't serve any purpose setting a DC bias level - you need resistors for that.

One solution would be to also remove the 10uF output cap from the buffer. That allows the bias level fo the buffer to feed the FET. That *might* work, but originally the FET is only biased by whatever it does internally and that 1M to ground, so I'd expect it to be a lot lower than the midpoint voltage that the opamp uses (9V on an 18V supply). I don't know much about FETs though, so perhaps someone else can step in here.

The simplest solution is just to put the 1M resistor back. It was there for a reason in this case, and it wasn't just to set the input impedance.

[nepalnt21]

gotcha. thanks for that.

i put it to the side for the night, cause im working on a v846 thing, but ill see how it sounds tomorrow.

[antonis]

The simplest solution is just to put the 1M resistor back. It was there for a reason in this case, and it wasn't just to set the input impedance.

It actually dominates input impedance but FET Gate needs a DC path to GND..
(or any other voltage bias..)

i also measured the two 2M resistors, and found two that were closer in value.

You don't have to be such a perfectionist..
(when dealing with relatively high supply voltage, matching of voltage divider resistors isn't a "must"..)

[nepalnt21]

thanks for the explanation.

i ended up removing the buffer, sounds much better. even with the buffer removed from the signal path, but with the positive voltage still applied to it, the distortion was sucked away.

i'll probably end up making a dedicated stompbox buffer.

[antonis]

even with the buffer removed from the signal path, but with the positive voltage still applied to it, the distortion was sucked away.

That makes absolutely no sense..!! 

Are you sure your buffer doesn't "load" power supply..??
(maybe due a defective chip or -by mistake- very low value bias resistors..)

[ElectricDruid]

even with the buffer removed from the signal path, but with the positive voltage still applied to it, the distortion was sucked away.

That makes absolutely no sense..!! 

+1 agree with Antonis. That indicates some other problem, nothing to do with the buffer itself, but maybe how it was wired up.

Still, if it works now without it, that's all good.