How a "Sweet Switch" really works

[László]

Yeah, it definitely gives the pickup more of a formant vowel sound.

Now I want to mess around with the sweet switch again on my guitar.

 

[Sage]

Now, anyone without a background in electronics or physics is probably wondering why the designer of this circuit chose a 135 nanosecond delay line as the filter for the sweet switch circuit. Well, the answer is quite simple after one performs a few calculations. An electromagnetic wave travels at the speed of light through a vacuum, which is 300,000,000 meters per second; therefore, an electromagnetic wave can travel 300,000,000 / 1,000,000,000 x 135 = 40.5 meters in 135 nanoseconds through a vacuum. However, an electrical wave does not travel at the speed of light through a coaxial cable because all coaxial cables have what is known as a velocity factor. A velocity factor is a fraction of the speed of light, which means that a coaxial cable has different electrical and physical lengths. The velocity factor of a guitar cable is roughly around 78%, which means that an electrical signal can travel 300,000,000 / 1,000,000,000 x 135 x 0.78 = 31.59 meters in 135 nanoseconds. That number is important because it roughly translates to 31.59 x 39.37 / 12 ~= 104 feet of guitar cable. In effect, the delay line simulates the frequency attenuation and the phase shift imposed on a guitar signal by a 100-foot-long guitar cable.

Resurrecting this old thread to add a link to this month's Mod Garage article in Premier Guitar, in which author Dirk Wacker explains how to make your own version of a sweet switch.

According to him, the delay line is meant for high frequency signals and doesn't actually work as expected in a low frequency system like an electric guitar; instead, it just acts like a very small tone cap. If that's true, there's no phase shift, just a low pass filter, like a fixed version of a tone knob, and it can be replicated via a capacitor between hot and ground.

I don't quite understand his argument for this, though. A delay line is a series of inductors in-line with the signal, and a capacitor to ground after each inductor. It makes sense that those capacitors in parallel would simply add up to a single larger capacitor, assuming the inductors aren't having any effect on the signal, but... why wouldn't they? I'm afraid the electronics theory involved here is a bit beyond my self-taught capabilities.

That's not to say Dirk's assertion doesn't make sense... Paul notes in the demonstration video posted earlier in this thread that the sweet switch is meant to simulate "35 feet of extra jack cord," and not 100 feet of cable as posited by em7 in the OP. That always struck me as odd, but in his article, Dirk does state that 1 nF of capacitance is roughly equivalent to adding 10 feet of cable. I asked him what capacitance the old 1513-135Y device would amount to, and he said roughly 3000pF. So about 30 feet or so, which is more in line with Paul's assertion.

I'm really curious to know what Em7 or anyone else here with a more formal knowledge of electronics thinks about this.

 

[watelessness]

washer or no washer. all that matters.

 

[Sage]

Okay, I took some measurements and did some math.

First off, Dirk's assertion that 1 ft of cable roughly equals 100pF of capacitance is way off. That's actually twice as high as almost any standard guitar cable on the market, and several times higher than modern low-capacitance cables. If you look at the manufacturer's specs, you'll see that standard instrument cable like Mogami 2524, Canare GS-6, etc. all have capacitance of 40pF to 50pF per foot (and interestingly, that capacitance is higher than that of lesser-known brands). And sure enough, my own homemade 10-foot cables made from Redco TGS-HD measure around 460pF, or 46pF per foot with connectors.

Secondly, I have a few 1513-135Y units (the exact device used in the PRS sweet switch) that I ordered directly from the manufacturer a while back, and I took them out and measured the capacitance between the input and ground. Every one of them measured about 1740pF, not the 3000pF posited by Dirk.

Paradoxically, while this demonstrates that Dirk's numbers are wrong, it actually bolsters his premise. If you divide 1740pF by 50pF, you get almost exactly 35 feet, which corresponds precisely to what PRSh said in the demonstration video: "about 35 feet of extra jack cord." To equal 104 feet, as initially calculated by @Em7, the cable would have to have a capacitance of less than 17pF per foot. That's rare even for today's ultra-low-capacitance cables; I highly doubt it was common or even possible in the late 80s.

Of course, that assumes that capacitance is the only factor. The 1513-135Y delay line isn't just composed of capacitors, but of inductors as well, and I don't know what effect the inductors have on the signal. But if capacitance is the real issue here, you could just use a common 1800pF capacitor in place of the 1513-135Y. I don't have one on hand to compare, but perhaps in the future I'll stick one in a pedal and try a/b'ing it against the 1513-135Y.

 

[Blues Lyne]

Okay, I took some measurements and did some math.

First off, Dirk's assertion that 1 ft of cable roughly equals 100pF of capacitance is way off.

Dirk said,

"Modern guitar cables have an average capacitance of approximately 100 pF per meter, which is very low and allows long cable runs without audible degenerations."

Easy to miss, I did a double-take when I read it as well.

100pf per meter is around 30pf per foot.

I don't have any experience with the PRS sweet switch but did a similar thing years ago on one of my guitars. The guitar has a thinner mahogany body, a glued mahogany neck, a maple cap, and two humbuckers. It had a harshness in the upper mids compared to my other guitars (S & T style) that I was having trouble dialing out.

Partly because of this issue and partly out of curiosity about the effects of cable capacitance, I rigged the guitar so that I could swap out different capacitor values across the output of the guitar. As I tried different values, most did what I expected and rolled off highs to different degrees. With some, you could also make out the shift in the resonant peak. However, when I connected a certain value, the guitar seemed to come alive. It seemed to move the resonant peak to a frequency that really works with the guitar and pickups. It still cuts, and almost has a mid-boost effect, but that upper mid-frequency no longer pokes out.

I also found that if you move the capacitor before the volume pot, the resonant frequency seems to be a little higher than with it after. Notes really jump out. It's a nice option that can sound especially good on a neck pickup that needs more definition and focus. This also allows you to have different capacitance values for the neck and bridge pickups. Of course, this doesn't simulate cable capacitance.

I installed an on/off/on switch so I can have stock in the middle and both capacitor options. As Dirk mentioned, you could wire the switch to switch between two different capacitor values instead.

 

[Sage]

Dirk said,
"Modern guitar cables have an average capacitance of approximately 100 pF per meter, which is very low and allows long cable runs without audible degenerations."

Easy to miss, I did a double-take when I read it as well.

100pf per meter is around 30pf per foot.

Ah yes, you're right, I missed that. I was using the math from his chart here:

As a little guideline to calculate the best additional capacitance, you can use this chart:

10 ft. cable (approx. 3 meters): 1 nF

15 ft. cable (approx. 4.5 meters): 1.5 nF

20 ft. cable (approx. 6 meters): 2.2 nF

30 ft. cable (approx. 9 meters): 3.3 nF

1 nF = 1000 pF and 1000 divided by 10 is 100, so he's suggesting 100 pF per foot here. Why that doesn't match his earlier statement I don't know, but long story short, 50 pF per foot is a more accurate measurement. I've got some 1500 pF film caps on the way that I'm building into a pedal to see if they achieve the same effect as the delay line. Will let you guys know the results.