Input impedance and pulldown resistors

[bipedal]

I'm still working to grasp the concept and application of input impedance.  I'm familiar with the great AMZ article on the topic and have read through some of the impedance-related threads here.  In general, it seems that high input impedance is a good thing to aim for in pedal designs -- easy enough.

From my understanding, the pop-eliminating pulldown resistors that are common to many of our builds have a major impact (though are not the sole factor) on a pedal's input impedance.   Following this, is it a good practice to be using larger pulldown resistor values?

For example, the Tube Reamer specifies a 1 meg pulldown, but the actual input impedance of the circuit is also affected to some degree (I believe) by the impedance of the op-amp (I've been using RC4558s -- got a bunch in a grab bag...).

Is there likely to be any practical difference -- increase in input impedance -- by using an even larger resistor value for the pulldown, say 2.2 meg or more?

Comments, clarifications, and/or suggestions for other articles to read are appreciated.

- Jay

[PerroGrande]

Hi Jay,

The anti-pop pulldown resistor, as you noted, *will* reduce the effective input impedance of a pedal. It adds another parallel path to ground for the incoming signal.  The result is a lowering of the overall impedance presented by the pedal.

Most circuits have three areas that contribute to the overall input impedance: 

1) Incoming anti-pop resistor
2) Biasing network
3) Impedance of the first amplification/buffering stage

These three components end up in parallel with one another and obey Ohm's law just like they were resistors.

Increasing the value of the anti-pop resistor will decrease its contribution to the overall input impedance.  However, understanding which of the three components above that contribute to input impedance is dominating the overall impedance is better than merely cranking up the resistance.  If your signal looks into a biasing network that presents an impedance of 50K for example, the contribution of the 1M pulldown is very small compared to the biasing network.  In this case, raising it to 2M (or even higher) doesn't really buy much.  The 50K dominates in either case. 

What the larger resistor will buy you, however, is a greater potential for noise.  This is a natural and unavoidable consequence of resistors.  You can change to a different composition of resistive material to lessen the effect of such noise, but it will still increase as resistance increases.

What I usually try to do is select a compromise value that will minimize both noise and the impact on the impedance of the circuit. Naturally, high-gain circuits are going to be the most susceptible to noise and require the most caution when selecting a value. 

[R.G.]

I'm still working to grasp the concept and application of input impedance.  ... In general, it seems that high input impedance is a good thing to aim for in pedal designs -- easy enough.

Input impedance is easy - it's just "how much current does the input eat per volt of input signal". The only complexity there is that the current eaten by every single component that touches the input signal must be accounted for. As previously stated, that is pulldown resistors, biasing network, and the active device that the signal finally gets to drive. Every new thing that's connected there is part of the input impedance, as it may let signal current leak to ground before it can effectively drive the input.

In general high impedance is bad; it allows hum and noise to creep in. However we are forced to use high impedance on the inputs of pedals because of the output impedance of a guitar's pickup. A pickup is among other things an inductor of between one and five henries. At high audio frequencies, the pickup itself gets up to hundreds of K of impedance, so the input of the pedals must be higher yet to not eat the current of those high frequencies preferentially and give the treble loss most people call "tone sucking". So high impedance is bad - unless there is nothing else you can do, which is the case for the standard guitar/cable setup.

From my understanding, the pop-eliminating pulldown resistors that are common to many of our builds have a major impact (though are not the sole factor) on a pedal's input impedance. 

There is a rule of thumb in engineering that things that only make 10% or less difference can be ignored, at least for a first approximation. So a pulldown resistor which makes less than a 10% increase in the current eaten in the input can be ignored as having any impact at all. The impedance of most of the old tube amps was 1M, as set by the DC grid leak resistor of the input to the first tube. That's what our ears expect to hear. So if the combination of all the stuff on a pedal input is around 1M, including pulldown resistors, then increasing the pulldowns to 10M, 22M, etc. will not help. Likewise, if the effect itself has an input impedance of 100K without a pulldown resistor, then no pulldown resistor of greater than 1M will have a huge effect. You're well into diminishing returns. It is not true that the pulldown resistors have a major impact as a blanket statement. It may or may not. There is no substitute for knowing the details.

For example, the Tube Reamer specifies a 1 meg pulldown, but the actual input impedance of the circuit is also affected to some degree (I believe) by the impedance of the op-amp (I've been using RC4558s -- got a bunch in a grab bag...).

That is correct. FET input opamps may have an input impedance of many megohms, and so the biasing resistors and pulldowns are all of it - the device can be ignored. An NE5532 has such a LOW impedance that it has a huge effect. You have to read the datasheets for the opamps.

Is there likely to be any practical difference -- increase in input impedance -- by using an even larger resistor value for the pulldown, say 2.2 meg or more?

The details matter. Look up the input resistance of your opamps. Better yet, set up your guitar into your amp and start with a 2.2M resistor loading the signal at the amp. Then a 1M. Then a 1/2M, 100K, 47K and so on. Let your ears hear what happens. Then you'll really know what happens, as opposed to only having heard it on the internet. 

[drk]

so, if i understand correctly, the input impedance its kinda of a voltage divider, which is use so that the signal doesnt get smaller. so what eats up the signal, that makes this tecnice necessary? is it because of the pickup, been a inductor, which varies its impedance with the frequency, and with this arrangement that doesnt happen?

[R.G.]

Yes. What makes this necessary is the pickup inductance, which is almost nothing - a few K ohms - at DC, but the inductance runs it up with increasing audio. A single coil pickup may be only 6K ohms at DC, the resistance of the wire. But it may have 2 henries of inductance, so the impedance at increasing frequency is 6K + 2*pi*F*L, or 2*pi*10,000*2 = 125k ohms (about; yes, I know it's not exact) at 10kHz. That impedance looks like it's in series with the output of the guitar.

It forms a voltage divider with the input impedance (and the cable capacitance, which we haven't touched on yet. Using my "rule of 10X", the voltage lost won't matter if the input impedance on the end of the cable away from the guitar is ten times the output impedance of the guitar. With 1M, we're not quite making it, as that would require 1.25M. But then guitar pickups don't put out much over 7k anyway, so maybe it's really closer to
2*pi*7000*2 = 88K, and a 1M is more than 10X bigger than that, so we would arguably not lose much audible treble.

So you're right - it's the voltage divider effect caused by the OUTPUT impedance of the signal source (the guitar) and the INPUT impedance of the thing being driven, the effect. If it weren't for the frequency-selective issues with the pickup inductance, we could ignore this and just glom in more gain to the amplifier to make up for the equal losses at all frequencies. But it's frequency selective, so we make the input impedance bigger.

[drk]

thanks RG, im starting to understand why everyone is so worry about the input/output impedance

and what's the downsides?
the voltage is reduced of course but what else? PerroGrande said that the noise increased, why is that?

[Paul Marossy]

The details matter. Look up the input resistance of your opamps. Better yet, set up your guitar into your amp and start with a 2.2M resistor loading the signal at the amp. Then a 1M. Then a 1/2M, 100K, 47K and so on. Let your ears hear what happens. Then you'll really know what happens, as opposed to only having heard it on the internet.

Good suggestion. Let your own ears decide! 

[R.G.]

and what's the downsides?
the voltage is reduced of course but what else? PerroGrande said that the noise increased, why is that?

Thermal noise. Every resistive or active electrical part has electrons moving around simply by thermal agitation. At absolute zero, all electron spin/motion ceases and there is no noise generated. Other wise, at any higher temperature, electrons move around randomly proprotionately to the absolute temperature. The noise voltage this random movement causes is also proportional to resistance. So the higher the resistor, the higher the thermal noise it generates. In some situations, this is a serious threat. In pedals, it's between no issue and an annoyance. But it is there, and that's what causes it.

[bipedal]

Thanks to John and R.G. for jumping in and tackling my rambling question -- lots of good info to chew on, and a'chewin' I shall go.

I've always appreciated those veteran builders who have been willing to add their 2 cents, even to those questions that must seem pretty basic.  I've learned a ton from posting my own q's and even more from reading other threads here in the forum.  It has been noted before, but it's worth reheating it again: on behalf of those of us builders who 'know-enough-to-be-dangerous-but-don't-really-know-why-it's-dangerous' (so to speak), a big thanks!

Warm fuzzy complete.  And now, to plan my next build.

Cheers,

- Jay

[drk]

didnt know that
thanks for explaining things once again