op amp buffer feedback loop frequency response

[JasonG]

I have not read anything about adjusting the  frequency response  for a non inverting op amp buffer anywhere.  My guess is that you would put a super small cap in the loop to trim down the higher frequencies but I have never seen anything dealing with that.
You could always put a RC network at the tail end of the buffer but is that the only option we have? 

[John Lyons]

An Ideal buffer is neutral.
The reason you use them is to preserve the frequency response of whatever comes before or after them.
Just curious why you would want to change the frequency response.

john

[George Giblet]

The common method of putting a cap across the feedback resistor doesn't work with the buffer (sometimes you will see this done but it is done for other reasons). The buffer works with 100% feedback so the RC network has no effect.   You have to put the filtering before or after the buffer.

As a side note putting a cap across the feedback resistor of any non-inverting opamp stage produces a shelving type filter instead of a low-pass filter.   The more gain you have in the non-inverting stage the more it looks like an ideal filter.   At 1/(2 pi RC) the response starts to drop off and the shelf point is at f = gain * 1/(2 pi * RC).    So if you have a gain of 2 then the roll-off will always shelve at 2 * the roll-off point.  Most overdrive pedals have significant gain and that's why the feedback cap more or less creates a low-pass filter.

The inverting stages do produce a low-pass filter but unfortunately the input impedance is low unless you want to put up with the additional noise produced by the large feedback resistors.

[JasonG]

An Ideal buffer is neutral.
The reason you use them is to preserve the frequency response of whatever comes before or after them.
Just curious why you would want to change the frequency response.

john

True the Ideal buffer is neutral but the real buffer is not. Mark Hammer pointed out in a thread a thread a while ago that buffers  do add some treble.  I was also thinking of the RC boost a lot of people seem to agree that it is a very transparent boost. Why do so many people agree? Because it has a subtle eq in it that lets you dial in the quality that makes you feel like its not coloring the tone.
I know its an odd question but  I thought that ether I have not been reading the right books or its imposable for some reason I had overlooked. My understanding of filters at this point is basic. I have not spent the time to work through the different styles , there advantages and disadvantages.
Thanks George

[alanlan]

You also have to factor in what transparent means in a practical situation.  If your amplifer has an "ideal" 1Meg, 20pF input (for arguments sake) and so does your op-amp cct with unity gain, then I would argue the combination of either cct and your guitar will yield the same result.

However, if your amplifier has a 100K input impedance and then you switch in your 1Meg 20pF buffer, it will more than likely yield a different result and you may be forgiven for thinking that it is the buffer's fault for not being very "transparent".

Things are never quite as simple as you think.

[Steben]

A buffer makes all what comes after it see a low impedance. This means that the loss of either lows or highs because of "loading" (difference in impedance of two circuits) is cancelled. Usually a guitar (pickups are coils, inductors) get's loaded by low impedances which gets the treble down. With a buffer between the two, the guitar is no longer "seen" as an inductor. With a decent guitar amp (read: high impedance) you hardly get any loading, meaning a buffer is very subtle to non-colouring at all. A buffer in front of a fuzz face (low impedance) for example, gives a very treblish fuzz sound.

[George Giblet]

I'll say it in a different way.  As John said a buffer is completely neutral, ie. it does not modify the signal or frequency response.  So how can it affect the frequency response?   What it does is prevents the change in frequency response (and level) when two things are connected together.

For the sake of a simple argument imagine your guitar being a voltage source in series with a 10k resistor.  A real guitar is more complicated than a 10k resistor but the details will cloud the point.  The frequency response of a 10k resistor is flat.  However your guitar connects to a cable and amplifier.  For simplifications sake say the cable will has 500pF capacitance.  When the guitar connects to the cable the two elements interact and form a low pass filter with cut-off frequency 1/(2*pi*R*C) = 1/(2*pi*10k*500pF) = 32kHz. 

An ideal buffer has a high input impedance and a low output impedance.  The low output impedance and it is this low output impedance which will drives the cable capacitance.  If the buffer had  say 100ohm impedance then the cutoff frequency would be 1/(2(pi*100*500pF) which is 100 times higher than without the buffer.

So when people say a buffer adds treble it is not the buffer that is adding the treble.  The buffer removes the drop in treble that is already present when a guitar connects to a cable (and amp).

Buffers don't have to affect frequency response they can also affect level, even though an ideal buffer has a gain of 1.  If the idealized guitar connected to a 10k load the level would drop to 1/2 because the interaction forms a voltage divider.  If the buffer is placed in between the guitar and the 10k load the level would be twice as high.   The buffer isn't providing gain it is removing the attenuation.

[John Lyons]

Thanks for the detail George!

john

[JasonG]

Thanks I guess I can lay that one to rest.