pulldown resistors.....

[boog]

what exactly do they do?  if i recall correctly i read somewhere they help with "popping" when you engage/disengage the circuit but if you ground the bypass switch are they necessary?  and if they are necessary, how do you determine the value? i've seen them from 1M up to 5M

[sean k]

Correct me if I'm wrong but yes they do in one instance allow any charge of the input cap to bleed off to ground while also seeting the impedance that the previous stage sees. So if your previous stage is another pedal with a 100k out then that and the 1M are in parallel which means an impedance of just over a hundred k ohms. So high meg resistors are kinda an insurance that whatever the signal impedance is coming in that it won't be held down too much by the previous stage which may be the case if you were coming out of a tube pedal and had a 1M out. Usually about 500k though so even with a 1M input those in parallel are somewhere around 400k so the loading is not much under what was designed into the previous stage.

[R.G.]

This is one of the problems with signal switching, and especially with mechanical switches, and especially with true bypass switches.

When a mechanical switch opens, it is very, very open - hundreds if not millions of megohms through the air between the switch contacts. If you have a true bypass pedal which is run from a single battery, then the inside of the effect is biased somewhere other than at the zero volts of ground. So there is a blocking capacitor to block that DC bias from getting out to your guitar and your amp. That works fine as long as the capacitor is connected to something on both sides. The outside of the cap (away from the inside of the pedal) is at the 0V of the outside world, the inside pin is at the few volts of the internal bias.

When that mechanical switch opens, it breaks the contact between the capacitor's outside pin and the rest of the world. The cap has charge stored, and if it were perfect, when it was next connected, it would still be at the correct voltage across it, so no pop would be heard.

But no capacitor is perfect. So while one end is opened up by the switch, the capacitor leaks a little internally, letting its (for instance) 4.5V DC value relax to 4.45, 4.40, 4.35... and so on until some time much later, it's leaked down a lot. When you connect it back up by operating the mechanical switch, the outer end is pulled to 0V by the switch, but the inside is now too low (or high) for the insides of the circuit, and the difference is made up by the circuit letting a big chunk of current flow into the capacitor to bring it back to the right voltage.

That chunk of current flowing is the pop.

A pulldown resistor works by keeping the outside of the capacitor pulled to ground all the time, swamping any internal leakage. No pop.

[earthtonesaudio]

...and you determine the value of the pull-down resistor (for a given capacitor) based on a compromise between the impedance drop caused by said resistor, and the volume of the pop you can tolerate.

If the cap is very leaky, you need a lower-valued resistor to bleed the charge to ground fast enough to keep up with the leak.  But if you make the resistor too low (a short circuit for instance) it loads the signal down more.  If you make the resistor too high in value, the signal is less affected but the pop becomes louder.

The "effect input grounded when bypassed" method is good at reducing pops, because it holds the cap at zero volts until you engage the effect.  However, when you step on that switch, the cap is then allowed to "float" (open circuit) for a small, but finite amount of time, during which a DC voltage can (will) develop on the outside lead of the cap.  If the cap is very leaky you will still hear a pop.

So, it's good practice to take all of these 3 separate measures toward reducing this problem when using mechanical true bypass:
1. Use lower leakage signal coupling caps (avoid aluminum electrolytics for example).
2. Ground the effect's input when bypassed.
3. Use moderately high valued pull-downs*.

*It's also worthwhile to take into account the input and output impedance of the effect in question.  If you want to keep input impedance high, you have to use a high valued resistor, but on the other hand, if the effect has low output impedance, you can use a lower-valued output pull-down without the same undesirable side effects.

[boog]

i think this may be the route to me understanding a little bit of this stuff.....thanks a bunch you guys!  am i correct in thinking this is the cap that is usually tied to input voltage (not sure if that's the correct term); for example, most schems seem to follow 9V+ w/ a 100uf cap (or some such high value) before entering the rest of the circuit?  i'm unsure since earthtonesaudio mentioned avoiding electrolytics, which are about all i have for any value over .47uf.

[earthtonesaudio]

I meant avoid electrolytic caps for signal coupling.  Power supply bypassing is a fine place for them.  And I certainly didn't mean you should avoid them altogether; just there are better alternatives in terms of reducing pops.

[boog]

got it; kindof.  obviously still a bit over my head, but thanks none the less!

[earthtonesaudio]

Cool.  If you get hooked on this hobby, you'll learn more and more everyday... and asking questions like you did is in my opinion one of the fastest and best ways to learn.  Good luck!

[Paul Marossy]

On rare occasions, I have had to put a pulldown resistor on both the input and the output to stop popping upon switching.

[boog]

surprisingly i've managed to put quite a few of these together without getting too much smarter; must be just good at following directions    i've got a bit of a grasp on high/low filters, but that's about all.  really don't get the power stuff yet.  and i just tinkered with one the other night that had a pulldown at the front and back of the circuit, that's what triggered the question;

[amptramp]

Pulldown resistors may go a long way to reducing the popping when switching between stages, but if the effect is still not satisfactory, consider the following:

If you have a true bypass that switches both the input and the output of the effect, there is some relative delay between the switching action of each pole.  If you rewire the switch to interchange input and output, there may be a difference in the popping noise and it may go from unsatisfactory to satisfactory.  With no input, some effects amplify noise or DC offset.  If you switch the input first, the output may momentarily be driven to the effect output value corresponding to an open circuit input when switching off.  If you switch the output first, you may avoid some of the problems.  It would be interesting to have a switch with a mechanism something like a film camera shutter where one pole switches on first and the same pole switches off first.  In most switches, the pole that is first on is last off.

My own preference would be to avoid true bypass and switch the output only, possibly using a unity-gain buffer for the input stage so the guitar does not see any change in load impedance.  Then you switch between the buffer output and the effect output.  But there will still be a voltage difference between the effect output and the guitar (or buffer) output because the effect unit generally changes the voltage from the original to something different, so it may not be avoidable.

The only way to completely avoid popping is to use electronic switching with MOSFETs where there can be a smooth transfer over time from one output to the other, almost like a pan.  An ordinary analog switch like a CD4066 may be too fast - discrete implementations would allow the timing to be designed in.

[boog]

given my infantile knowledge of this stuff, i'm just going to stick w/ true bypass for now; i have been looking at the articles on geofex and the tone god about doing the cmos swithching stuff.  plan to try that over the winter.  been getting overly ambitious recently by trying to cram too many effects into one enclosure which increases the possibility of popping.  so this may be a pleasant alternative.

[R.G.]

It would be interesting to have a switch with a mechanism something like a film camera shutter where one pole switches on first and the same pole switches off first.  In most switches, the pole that is first on is last off.

Something similar does exist, and was once used extensively for audio circuits. They're called make before break switches.

The only way to completely avoid popping is to use electronic switching with MOSFETs where there can be a smooth transfer over time from one output to the other, almost like a pan.  An ordinary analog switch like a CD4066 may be too fast - discrete implementations would allow the timing to be designed in.

Discrete MOSFETs suffer from not having the substrate node brought out. It's internally tied to the source. This causes the reverse-biased junction used to isolate the channel from the substrate to look like a reverse biased diode from source to drain. So discrete MOSFETs are essentially unusable for this, unless you have a source for four-terminal ones with the substrate available. You can do something like this with the CD4007 MOSFET array, though. The N and P substrates are internally tied to V- and V+ and not to the sources, so you get to use the width of the power supply with this one device.

JFETs are about the only easy option for fast, externally timed switching.

Note that even the 4066 and the 405x series become almost pop free if you bias them so that the DC level of the signal is in the middle of the apparent power supply to the chip. The switching elements still have some bleedthrough of the control signal, but this is in opposite directions and if the switching is done in the middle of the power supply, of equal magnitude. The bleedthroughs cancel to a large degree.

There are also specially designed audio switch chips which have essentially zero switch pop from the control signal.

[captntasty]

I'm going to go a little off topic here but since the subject of switch pop has been brought up.... what do you do with an effect that no matter what you do the bypass produces a pop?  Would a lower (than 1M) resistor make any difference?

[earthtonesaudio]

If you've already tried using lower-leakage caps and grounding the effect's input when bypassed, lower value resistors might be a good next step.

[captntasty]

hmmm... how would I measure a caps leakage?

[R.G.]

I'm going to go a little off topic here but since the subject of switch pop has been brought up.... what do you do with an effect that no matter what you do the bypass produces a pop?  Would a lower (than 1M) resistor make any difference?

I personally would do more thinking. If the standard bag of tricks is not working, chances are that you're somehow switching between two different DC levels, and that implies that either the effect you're working on, the *amp* input, or the *guitar* may be putting out DC. Get out your meter.

But it's always good to question your assumptions. Electrolytic caps work - kind of - if they're not reverse biased too badly, but they leak so much that pulldown resistors are ineffective. Same for old, or heat abused electrolytic caps.

hmmm... how would I measure a caps leakage?

It's likely that you can't, at least without advanced equipment. Instead, you change to a known low leakage type of capacitor. Electrolytic caps are worst, then there's all the plastic film types with tiny variations and finally polystyrene, teflon, and vacuum caps as best.

[dukie]

well how do you put a pulldown resistor in GGG Hogs foot schematic because until now im still not succeeded to eleminate the pops. ive tried to put a 1M pulldown resistor before the input caps and after the output caps but its not working. is it because there is a volume pot in series with the effect input?

by the way i mod the hogs foot using 3P4T rotary switch to select values of the input and output caps, and pops occured when the rotary switch is engaged.
is there any way to prevent the rotary switch pops using pulldown resistor?


cheers!

[R.G.]

well how do you put a pulldown resistor in GGG Hogs foot schematic because until now im still not succeeded to eleminate the pops. ive tried to put a 1M pulldown resistor before the input caps and after the output caps but its not working. is it because there is a volume pot in series with the effect input?

Can you provide a URL for the schematic you used? I did some searching, found the GGG Hog's Foot schematic, but it does not have a volume pot in series with the input.

The Hog's Foot as presented at GGG should be fine with 1M pulldowns before the input cap and after the output cap **if the caps are oriented the correct way round** that being with the + end toward the circuit, and the caps are not somehow damaged.

Also, I did not see any volume pot in series with the effect at GGG, so I don't know what schematic you're actually using.

by the way i mod the hogs foot using 3P4T rotary switch to select values of the input and output caps, and pops occured when the rotary switch is engaged. is there any way to prevent the rotary switch pops using pulldown resistor?

There is. You have to use high value resistors to ensure that the unused end of any capacitor is always pulled to a constant voltage whenever it's not being used. Otherwise, you will inevitably have popping because the capacitor will leak a little while one end is open circuited, and the amount it "relaxes" will be heard as a pop when it's switched back in. One way to do this is to put a 1M (for instance) in series with each cap and use your selector switch to short out the resistor when that cap is in use. Another is to gang up all the caps with one end at the circuit's input, the other end of every single one pulled to ground. Then you can hook the rotary switch to route the input to the desired cap without popping because all of them are pulled down to the same voltage.

Finally, some switches just pop all by themselves.

The details matter.

[captntasty]

Sure enough I had a 1uf elec for an input cap... I put in a non-polarized film in there and pop is gone.