Potentially Stupid Question About Capacitors and Filtering

[abc1234]

My question applies to all audio circuit designs, but is best illustrated with a specific example:

I'm building a Soda Meiser and want to make it more bass-friendly. One of the ways I attempted to do this was by doubling the values of the capacitors in the signal path (0.1uf to 0.2uf) in order to pass more low end. However I abandoned doing this for all but one of the caps when I realised doing so also increased the volume and therefore made the circuit have too much saturation and too little gating.

The obvious exception is the output cap; since it's not between stages but at the end, the only effects of a 0.2uf appear to be an increase in volume and low end passing through.

...Which leads me to my question: are there any disadvantages to using a higher value output cap?

Specifically, since all the prior caps will have attenuated some low end, won't the additional lower frequencies passed by a higher value output cap not be from the bass guitar at input but instead generated by the final transistor stage?

i.e. Will it just be passing mud that would remain rolled-off by keeping the stock value?

Here's the Soda Meiser schematic for reference:



And here's the MXR Distortion Plus as an example of a circuit using an output cap 100 times larger than the input cap:


Source: https://www.premierguitar.com/articles/mxr-distortion-plus-mods-1

[GibsonGM]

The output cap will work with the vol pot in an R-C filter arrangement..so you'll have some effect on the tone depending on the vol. pot setting.  What follows the circuit will also have an effect, but we can't really control that.

Note the 100k vol pot in the Effector, vs. 10k in the Dist+, and 10x difference in cap size.  Same cutoff frequency.

Whether upping that cap value will make your sound muddy, I doubt that very much - I suspect it won't do much at all.....I CAN say - TRY IT!!    Experimenting is the best way to see what effect these changes will have.   Since the orig. config. seems to pass everything over about 16Hz, a bigger cap will just pass "lower than that", and you may not notice a thing since we don't hear below 20Hz and there's nothing worth 'emphasizing' down there - but there could be other factors in play, so actually doing it might reveal something!   .1 vs .2u is tiny here....I'm talking, .1 vs. 1u!   It may make it seem 'fatter' or something, and it will act differently at different vol pot settings... 
 
Or it may not change the sound at all       

[antonis]

You might follow signal path (conventionally circuitry from left to right..) to find what happens..

You can't recover an already cut-off signal at, 100Hz say, with a passive RC filter of 200Hz cut-off frequency placed after the first one - to make a peadal more bass-friendly you have to establish bass-friendship from the begining..

As Sir Mike told you, first order passive RC filters "act" at a cut-off frequency (-3dB point, which corresponds in 0.707 voltage attenuation, further exhibiting -6dB per octave / -20dB per decade attenuation)  which is determined by the inverse product of their values and 2*π [1/(6.28*R*C)]

BEWARE of specific RC arrangement 'cause reverse RC setting result in totally different task..!!
(series Cap w Shunt Resistor form HIGH pass filter where series Resistor w shunt Cap form LOW pass filter..)

MXR Dist+ Input filter should be considered simple due to some general acceptances, like very high pin3 input impedance and very low R1 value in relation with R5 value, so we have to deal with a High pass filter of 15.9Hz cut-off frequency.
R3+R2 with C3 also form a High pass filter which should be added on Input filter, forming a second pole (12db per octave) on pin6, for frequencies equal to or lower than those of Input filter..
The above comes to add on C4/R6/OUT pot to form a third pole (18db per octave).

Soda Meiser is trickier due to "complicated" equivalent shunt resistance with 100nF series cap but if you consider as sole shunt resistance the parallel combination of BJT's intrinsic emitter resistane times h_FE and feedback resistor divided by stage gain you'll get close to reliable results.. 

P.S.
No question should be considered (potenially or not) "stupid" till it turns to a major face-palm..

[R.G.]

[...] I attempted to do this was by doubling the values of the capacitors in the signal path (0.1uf to 0.2uf) in order to pass more low end. However I abandoned doing this for all but one of the caps when I realised doing so also increased the volume and therefore made the circuit have too much saturation and too little gating.

Distortion pedals are tricky. If you give a read to the Technology of the Tube Screamer at geofex.com, you'll get some of this. It is common in distortion pedals to choke off some bass in the input stage, as bigger low frequency content into a clipper tends to obscure any highs. So it's common to put a long, slow bass rolloff on the input.

It helps to get a mental handle on what filters do. Series capacitors do act like high pass filters, but they are soft, slow rolloffs. A single capacitor in series with a a resistive input causes a reduction in frequencies below the nominal "cutoff frequency" of only half-power per octave. So it's better to think of a series cap as not so much a brick wall where everything above some frequency goes through and everything below it is stopped, as a ramping-down tone control. If you stack a bunch of these in series, it does add many filter sections and get more brick-walled, but - well, it gets back to there being no substitute for knowing the details.

Perhaps the best way to make a pedal "bass freiendly" is to do a clean mix of the dry bass signal with the pedal. This keeps the tone of the pedal the same on the parts of the audio spectrum it acts on, and mixes in the lower octave that the pedal may be filtering out.

...Which leads me to my question: are there any disadvantages to using a higher value output cap?

Specifically, since all the prior caps will have attenuated some low end, won't the additional lower frequencies passed by a higher value output cap not be from the bass guitar at input but instead generated by the final transistor stage?
i.e. Will it just be passing mud that would remain rolled-off by keeping the stock value?

How much extra bass a big output cap passes depends heavily on what the earlier capacitive rolloffs have done. If there's no bass there, it cann't let more through. More of zero is still zero. But especially in what I'd call ideosyncratic pedals, pedals designed by tinkering instead of actual design processes, you would have to either try it and see if you like the result, or do a design analysis to find out what the various break points are doing to you.

And as noted, the final transistor stage won't be generating low frequencies. Well, it will generate 1/f noise, but it's very unlikely that this will be noticeable. Don't sweat this one. Pretty much if it's not in the signal going into the last transistor, it won't be coming out.

[amptramp]

Is this being driven by the guitar itself or is there another effect or a preamp between them?  If you have to couple the guitar to this circuit, be aware that the inductance of a bass pickup may be around 5 Henries and there may be a volume control at the guitar output as well that changes the impedance drastically.  The input impedance to the Soda Meiser is quite low and will kill off a lot of the signal.  The MXR Distortion + is a lot better in this regard with a nominal 1 megohm input impedance.

[vigilante397]

Experimenting is the best way to see what effect these changes will have.

Experimenting is the best way to see what effect these changes will have.

Experimenting is the best way to see what effect these changes will have.

Copied three times for emphasis. This is SUCH A BIG DEAL. When I first started designing pedals from scratch I did math and analyses and simulations, and somehow nothing I made sounded like I wanted it to. Now 99% of my development is done on the breadboard. The end goal is to plug it into an amplifier, not an oscilloscope. Not that there isn't a time and place for math and simulations and such, but if the goal is to make something that sounds good, trust your ears!

[antonis]

Not that there isn't a time and place for math and simulations and such, but if the goal is to make something that sounds good, trust your ears!

True and Correct but, IMHO, an elementary knowledge of cut-off frequencies RC values combination, poles & slopes, passive interactions, etc, saves you A LOT of wasted experiment time.. 

(I would love to watch you, Nathan, trying to form a 16Hz high-pass filter starting with a 10pF Cap & a 10R resistor, implementing incremental raising value steps according to E12 rating..)

[ElectricDruid]

Not that there isn't a time and place for math and simulations and such, but if the goal is to make something that sounds good, trust your ears!

True and Correct but, IMHO, an elementary knowledge of cut-off frequencies RC values combination, poles & slopes, passive interactions, etc, saves you A LOT of wasted experiment time.. 

+1 agree. The math gets you into the right ballpark, and then you can adjust by ear. Anyway, for RC filters, it's easy to look up - you don't have to actually *do* any sums! (heaven forbid!)

[abc1234]

Thanks for everyone for their contribution.

I guess the answer to my question, "are there any disadvantages to using a higher value output cap?", is no.

To those who said I should experiment and listen, as stated in my original post I had already done so and concluded that it simply increased the output volume.

Pretty much if it's not in the signal going into the last transistor, it won't be coming out.

That's what I thought, so why are there designs like the MXR Distortion Plus that have an output cap that's so much larger than the input cap?

[Groovenut]

That's what I thought, so why are there designs like the MXR Distortion Plus that have an output cap that's so much larger than the input cap?

It's an impedance thing. Because of the input impedance the .01u at the input of the MXR has a knee point of about 16Hz. with the output impedance the 1u yields about the same knee. Input filter is basically .01u and 1.01M, output filter is basically 1u and 10k (give or take).
Whatever plugs into the output of the pedal will be in parallel with the 10k rough output impedance. Usually this is any where from 250k to 1M. The 10k in the pedal still dominates the external impedance resulting in about the same filtering.

[amz-fx]

so why are there designs like the MXR Distortion Plus that have an output cap that's so much larger than the input cap?

Because you do not know what the circuit will be plugged into, and the impedance of the next stage (or pedal) may be high, and of no consequence, or it can be low and will load the output of your pedal, which can cause frequency loss.

Most of the time, you will design for the worst case driving the input, and the worst case loading the output.

regards, Jack

[Jubz]

Hi

Ive experimented with bass distortions on breadboards and simulations tools lastly, and I ve learned that all the changes you will do to filters before the signals will change the character of the distortion (or fuzz or overdrive) a lot. And as RG said, and the simulations tools cant really show you that, adding more bass to the distortion circuit doesnt just "add" something to the sound, it gives me the impression to take the highs and a lot of the bite away.

And it can get very tricky because the output of bass guitars vary dramatically between passive and active instruments. A distortion that I like with my passive jazz bass will sound dull and boring with my g&l1505 because it has a very hot output and lot of bass going into the distortion circuit.

The clean blend has worked very well for me. Another trick I ve liked is to add a "tight" control. A high pass filtering at the beginning of the circuit which let you adapt the circuit to different basses. If you look at a darkglass b3k schem you will see a very interesting notch filter centered around 320hz at the beginning of the circuit and a treble boost (yeah in a bass distortion pedal).

But at the same time look at the schematic of the idiot box blowerbox (higly praised rat circuit for bass). All the caps values seem to have been raised with the "let pass the maximum low frequencies everywhere" logic in mind. And the pedal sounds killer. So I guess there are guys who manage to keep the bite of the sound and not cut bass at the beginning of the circuit. How they do is currently a mistery to me.

Or you could try to separate frequencies band and send them to different circuits, which let you control the volume and distortion of bass mid and treble. Others recommended that to me when I was on my quest for bass distortion. Look at the ZORG basstard and small bastard. It does it with a very cool fiter arrangement (state variable filter) at the beginning.

[vigilante397]

(I would love to watch you, Nathan, trying to form a 16Hz high-pass filter starting with a 10pF Cap & a 10R resistor, implementing incremental raising value steps according to E12 rating..)

I'll be sure to live stream it next time

[PRR]

> The math gets you into the right ballpark, and then you can adjust by ear. Anyway, for RC filters, it's easy to look up - you don't have to actually *do* any sums! (heaven forbid!)

Bah. Use a Reactance Chart. The answer is right there, just follow the lines.

[antonis]

What Paul said..



IMHO, on a full log chart without extremely detailed axis values isn't so easy to read ordinate or abscissa values..

I prefer to have in mind a "standard" R/C values product (like 100nF/100k -> 16Hz) and make by heart any other elementary calculation..
(e.g. a 10 times bigger cap or resistor lowers 10 times the cut-off frequency and vice versa or raising one value with simultaneously lowering the other value keep result unaltered..)

[abc1234]

Thanks Groovenut and amz-fx.

[ElectricDruid]

What Paul said..

Lol! Each to their own:

https://electricdruid.net/rc-filter-calc/

[GibsonGM]

I just made a visual basic program that does the same as Tom's page, like 10 yrs ago.  Easy as h3ll, gets you what you want...