Output cap size?

[carboncomp]

How do you know what size to use?

For example, I'm making a Rangemaster with a few different size input capacitors, will I have to use a larger output cap too, how will I know what size to use, is there some math, or a measurement that can be made? 

[bool]

You can calculate, measure and simulate ... but in the end, you will do it BY EAR (and multiple amps and guitars).

[duck_arse]

You can calculate, measure and simulate ...

.... or just plain steal from someone else's design. work out the cut-off frequency of the standard rangemaster output cap. too low?

[carboncomp]

Can you tell me how to calculate, measure?

Saying "by ear" is the same as telling someone asking for a cake recipe to bake what taste nice!

[cloudscapes]

I usually go for between 0.1uf and 1uf. Keeps bypass pop away.

[Elijah-Baley]

In my Brian May treble booster and my TS808 I put a pin socket. So I tried some caps.

It is like you taste the cream of the cake while you make it. 

[antonis]

Can you tell me how to calculate, measure?

The cutoff frequency of a RC filter is calculated from: f = 1 /2π*R*C  where f in Hertz, R in Ohms, C in Farads & 2π = 6.28

[anotherjim]

You need to know 2 things before you calculate the value.

1 - The lowest frequency you want to pass without significant level loss.
For guitar work, this is often taken to be 100Hz.

2 - The Impedance of the circuits AFTER the output capacitor.
The output capacitor forms a High Pass Filter with this "Load" impedance. This could be a high 1Meg ohm impedance amp channel or a stomp-box with a much lower impedance such as an original Cry Baby Wah. To ensure against the variable nature of what comes next, it's often done to have a buffer stage with a very low source impedance at the output. and then an output capacitor that's so large, it won't limit the bass even if the load is as low as 10k ohm. The low frequency performance of the effect circuit can be controlled  in a predictable way by the capacitor feeding this buffer circuit which has a known input impedance. Many stomp-box designs make no attempt at having such a buffer and rely on the following load to be around 1M or haven't considered it much at all, which leads to another point...

Beware that some combinations of effects sound the way they do largely because of the way the different designs affect the overall frequency response. Installing a buffer or changing the coupling caps will affect this interaction.


Or, ignore the clever stuff and choose a capacitor size that has insignificant impedance whatever happens after it.

An on-line calculator like this one can easily show you the impedance of a capacitor at a given frequency...
http://www.qsl.net/pa2ohh/jslcimp.htm
Your computer needs Java for this to work.
If you select "Capacitor - Khz" and enter 0.1 for the frequency of 100Hz
Then enter a Capacitor value (it wants Nano Farads). I know it's going to be a fairly big Electrolytic to get a low impedance so try 100uF which is 100000nF.
Click calculate on impedance and you get about 15ohms. That will do.

Edit.. I forgot to add...
Somebody mentioned bypass popping. That's the DC on the circuit side charging the coupling capacitor when it "sees" the connected impedance when the bypass switch connects the effect. The bigger the capacitor, the bigger, and longer lasting is the pop.
Putting a resistor to ground from the output side of the capacitor and before the bypass switch will let the capacitor charge when it's powered up and stay that way - the anti-pop resistor.
This resistor is often 10k, and sometimes it's the volume pot, but it's often a plain resistor. 10k across the output means you now have a good idea of the load impedance. Plug the effect into 1Meg and it will have negligible change on the 10k value. Even a load as low as 50k won't have too much effect.

Now you know the load is 10k, you can calculate a precise capacitor value.

You could use another on-line calculator for working out an RC High pass filter such as this one...
http://sim.okawa-denshi.jp/en/CRhikeisan.htm

This one's easier to use since you can type in the range multiplier symbol. So enter 10k for resistor and try some capacitor values then click calculate.
After trying a few around 100n I find 220nF gives a low frequency cut off of 72Hz. Near enough gor guitar, and a smaller component than the brute force 100uF I picked earlier. But maybe I should play on the safe side and have some insurance against a really low impedance load being connected. So double it to 470nF. Now I have Bass down to 35Hz, but if it gets plugged into (worst case) a 10k load, I'll still have performance down to 70Hz.

[nocentelli]

Saying "by ear" is the same as telling someone asking for a cake recipe to bake what taste nice!

Not at all. No-one would suggest building an entire pedal before listening to the sound of the output cap. Buy a breadboard, build the circuit, hook it up to your amp to see how it sounds. Swapping caps will take 10 seconds if you're slow. If you build a simple looper to bypass the breadboard, you can also easily hear how it sounds running into your other pedals.

Since the range master is a treble booster which uses a smaller input cap that rolls off some bass at the input, I'm not sure the output cap value is that important unless you go too small and make it too thin.

[deafbutpicky]

like trying to boost a germanium FF into full-bore. Caps in series behave like resistors in parallel, adding low input
impedance can end up in fairy jingles. I've made the 10n on my naga viper switchable to 1µ e.g.

[merlinb]

Something that hasn't been mentioned is that if the circuit has a low output impdance (and any modern circuit should do, although most amateur designs don't), then the output cap should be big enough to maintain that output impedance down to low frequencies. The whole point of having a low output impedance is to give immunity from hum being picked up on the output cable. There's not point in having a nice low-impedance opamp output buffer and then sticking a puny 0.1uF cap on the output, since the cap has a reactance of 26k at power line frequencies! I would never us anything less than 10uF. Tone adjustment should be done before the output cap, not with it.

[R.G.]

I think I would put it another way.

In picking an output cap, you are providing a low impedance (* to be defined below)  from whatever source impedance the pedal puts out into the unknown impedance that will be connected to it as a load.

Yes, a modern design should have a low output impedance, but what matters most is the impedance of the cap compared to the unknown load. At high frequencies, the cap will always let the signal level tend to the voltage divider result of the load impedance and the output impedance of the circuit. That's not very moveable. What the cap does is change the frequency at which that starts being true.

*A quick quide to whether something is much smaller than something else, in the absence of any other way to judge, is if it's 1/10th as big. If the cap is so big that for all values of signal frequency its impedance is much smaller than either the output impedance or the source impedance, the output level is Rl/(Rl+Rs). It's value compared to Rs doesn't change the end game significantly.