A couple of questions about op amp distortion

[chromesphere]

Hey guys,

Occasionally I get the urge to like...you know...do some learnin'.  At the moment I'm looking at some basic op-amp circuits and I have 2 things I'm not totally sure about.  Lets use the Dist+ as an example:
(http://www.freeinfosociety.com/electronics/schematics/audio/pictures/mxrdistp.gif)

1) For biasing the positive input of the distortion plus there are the 1meg resistors coming from +9v.  I would have assumed that we'd want the positive input of the opamp to sit at +4.5v?  As far as my knowledge of voltage dividers go, +4.5v should be a the middle junction of those 2 (3) 1 meg resistors.  Wouldn't the point where the positive input is actually connected to be closer to ground?

2) The low pass filter on the feedback loop of the dist+ (the distortion pot) is connected to ground.  I have seen some circuits (the love pedal for example) connect this to the +4.5v point of the voltage divider as I described in the last point.  Is there a reason for this?  Is it effectively achieving the same thing? (your guitar signal sitting at a +4.5v bias point).

Hope this isn't too complicated a question(s) to be asking, thanks for any help!
Paul

[Seljer]

1) Probably. In a datasheet  I looked up for the 741 the input impedance of is listed as between 1 and 6Mohms. With high value resistors there for the divider in most likely pulls it down a measurable amount.

2) The capacitor in series with the distortion pot actually makes a high pass filter, not a low pass filter (the 10pF between the output and - input lowers gain at high frequencies). With the capacitor in series there where it is, the gain of the opamp for DC voltages is effectively set to 1 (pretend the cap is an open circuit at DC and you're left with what looks like a non-inverting buffer circuit), so the DC voltage at the output matches whatever  DC is present at the + input.

[chromesphere]

2) The capacitor in series with the distortion pot actually makes a high pass filter, not a low pass filter

Whoops, yeah high pass not low pass
Ok thanks for the response that makes pretty good sense.  I forgot about the cap in series!
Paul

[PRR]

> Wouldn't the point where the positive input is actually connected to be closer to ground?

The number you want is "input bias current". For '741, *typically* a little under 100nA or 0.1uA.

The 1Meg+1Meg from 9V makes 500K or 0.5Meg from 4.5V.

0.1uA times 0.5Meg is 0.050V.

So the 4.5V is loaded by 0.050V. In the case of '741, the input current is simple NPN transistors, so it will go low, 4.45V.

4.45V == 4.5V for all practical audio purpose.

[dwmorrin]

2) The low pass filter on the feedback loop of the dist+ (the distortion pot) is connected to ground.  I have seen some circuits (the love pedal for example) connect this to the +4.5v point of the voltage divider as I described in the last point.  Is there a reason for this?  Is it effectively achieving the same thing? (your guitar signal sitting at a +4.5v bias point).

+1 on the question here.
I've been wondering this lately, but haven't put much thought into it yet.
I have to admit, I was a bit disappointed PRR didn't touch on this question in his reply.
Are there pros or cons to choosing 0V or 4.5V for the ac path of the negative feedback?
I want to say that the theory is the same, so it doesn't matter.  But maybe there's practical considerations?  Or maybe I'm missing some element of the theory here?

[chromesphere]

Thank you for your reply PRR, I will try digest your explanation, cheers!

Yes I agree David, is there any difference?  I thought perhaps the question was answered with Seljer's response, because there is a cap blocking the DC, their is effectively no difference?

Paul

[Johan]

As far as the signal is concerned, the is no difference between ground and any well filtered voltage reference point. But if you are using an electrolythic cap to filter or block dc, that cap will be happier in the long run if there is a small voltage across it.... it is not something i would worry about in any diy application thou. ...

[Keppy]

As far as the signal is concerned, the is no difference between ground and any well filtered voltage reference point.

This is correct. It shouldn't matter whether that point is connected to ground or Vref as long as the connection is made through a blocking cap and Vref is adequately filtered.

[teemuk]

I would have assumed that we'd want the positive input of the opamp to sit at +4.5v?

This is correct.

As far as my knowledge of voltage dividers go, +4.5v should be a the middle junction of those 2 (3) 1 meg resistors.

This is correct too.

Wouldn't the point where the positive input is actually connected to be closer to ground?

Why? The input resistor connects to aforementioned 4.5VDC voltage reference so I don't see a reason why the opamp input should or would be "closer to ground".

The low pass filter on the feedback loop ... is connected to ground.  I have seen some circuits ... connect this to the +4.5v point of the voltage divider as I described in the last point. Is there a reason for this?

Yes. If feedback loop's "shunt" path isn't capacitively coupled the other opamp input needs to be referenced to 1/2 Vcc as well. Alternatively, if the other opamp input lacks the DC reference it can be provided via the shunt path of feedback loop.

Is it [low pass filter on the feedback loop ... connected to ground] effectively achieving the same thing?

It depends what you mean with same thing? If you mean a low impedance termination point for AC signals then yes: the AC signals don't care that they see a few largish capacitances on their path, they still nicely terminate to ground. If you mean a 1/2Vcc DC reference for the opamp input then no: The capacitors block DC. However, since DC reference is already provided at the non-inverting input the inverting one will automatically bias to same voltage potential as long as there is no path for DC signals that could skew the bias. ...And the capacitors block DC so there isn't.

Are there pros or cons to choosing 0V or 4.5V for the ac path of the negative feedback?

If the opamp lacks another reference then you have no choice but to select 4.5VDC and direct coupling. If reference is provided and the singal path capacitively coupled you can choose either 4.5VDC or 0VDC as reference but in case of 4.5VDC you need to use non-polar coupling caps for the shunt part of the feedback loop (no significant DC voltage difference across the cap) and the AC signals will see the DC reference's filtering capacitor in series with the others of concern. Depending on design the 4.5VDC point may introduce noises coming from power supply.

Non-polar caps are a bit more expensive and supposedly a tad more non-linear than polar ones. If capacitance values are low to begin with you don't even have to worry about that because you are not probably choosing an electrolytic coupling cap to begin with. The effect of the filtering cap to overall capacitance in signal path is usually taken into consideration already in the design stage. Adequate filtering usually ensures a steady-enough and noise-free reference point. One can easily turn even a true ground reference noisy with poor layout so the comparison of their noise-freeness as is and without further details is somewhat moot. So neither pros or cons are astounding and can be easily compensated.

I think people generally find more pros or cons from which reference happens to be easier to use layout-wise.

[chromesphere]

Im glad I asked this question I feel somewhat enlightened Thank you all for taking the time to respond, I appreciate it!