Where to put the outer wrap of caps in a Tone control circuit?

[Mark]

I am replacing a couple of caps in the tone controls of my 6G7a Bandmaster. Typically, I'd look for the easily path to ground or the HT. The tone circuit more or less throws this out the window as there is a 100K slope resistor on one side and a 250K pot on the other side of the 0.1uF bass cap. The 0.022uF mid cap isn't an issue as these is a 6K8 resistor to ground, making that the obvious choice, though the bass cap is a little different.

Thanks for your time, I look forward to your reply.

[wpaulvogel]

I’d face the outside foil towards the previous stage.

[R.G.]

Sigh.

[Mark]

Sigh.

In the past I've seen caps with copper wrapped around them and earthed, not to mention all sorts of other Voo Doo. The idea of having the outer wrap towards earth doesn't seem to be that bad an idea, if you're trying to minimise noise. Granted shielded cable on the input and the volume controls are more effective, as would be DC heaters in certain circumstances.

But if you're putting a cap in and one way is quiet, even by a small margin, what is there to lose?

Thanks for your reply Paul, I appreciate you taking time to reply.

[wpaulvogel]

Once I found out about the outside foil possibly reducing noise, I began to use the configuration in my builds. I have no idea if it really works or if so how much. I do believe that an amp works hard enough just amplifying the guitar signal and allowing the amp to focus on that only will produce better harmonics and maybe reduce oscillation.
I measure every resistor, cap, pot, diode and transformer to make sure it’s to spec before I install it, why wouldn’t I face the caps, I put the resistors in with the stripes facing the same way.

[R.G.]

In the past I've seen caps with copper wrapped around them and earthed, not to mention all sorts of other Voo Doo. The idea of having the outer wrap towards earth doesn't seem to be that bad an idea, if you're trying to minimise noise. Granted shielded cable on the input and the volume controls are more effective, as would be DC heaters in certain circumstances.

But if you're putting a cap in and one way is quiet, even by a small margin, what is there to lose?

Your post caught me in bad mood. What I should have said was:

I suppose it's a question of objectives. You only lose the time you spend tinkering with capacitor orientation, I suppose. and if you enjoy the time spent tinkering, no harm. I sighed because it's not likely to do much good either way round.

Noise is a terribly broad category. It's probably best defined as any signal you don't want, much like "weed" is any plant you didn't want in your garden. I'm guessing that you are trying to minimize low frequency (hum) pickup rather than high frequency audio and RF, just from the wording of your reply. If I got that wrong, let me know.

The general theory of where to hook up a shield was the subject of several chapters in one textbook on noise. The chapters went on forever about shield effectiveness to various frequencies and what the shield impedance looked like at frequencies of interest, coupling mechanisms, impedance of the free space around the shield, that kind of thing. A lot of it got down to the general idea of how the noise gets coupled in and how it is then disposed of.

Even when the cap in question has one terminal going to ground, you have to ask yourself what the "ground" is carrying. You can cause RF oscillation by tying a cap to the wrong ground. Ground has an impedance too, and may be carrying a feedback current that can mess things up by turning a cap-to-ground into a very low source impedance input.

In the case of a cap from a tube plate to an RC network and another plate, there is some plate resistance on one side and some RC impedance on the other. A 12AX7 plate has about a 60K plate resistance. The RC network you mention has various impedances at different frequencies. If the cap picks up any noise, it's got to go through one of those paths to ... somewhere else. One way it's pushing through a 60K plate resistance and the other it's the RC network.

But the noise had to come in by the cap acting as though it's connected to another invisible capacitor, that being the capacitance to the noise source. The outer foil gets directly connected to the invisible cap and the inner foil is connected to the outer foil by the capacitance of the subject capacitor itself. Stray capacitances tend to run to the few pF range. The subject capacitor is generally from two to several orders of magnitude bigger than the stray coupling capacitor, so effectively, both foils are connected by a very much lower impedance than the stray coupling capacitance. Any difference the inner versus outer foils makes is limited by the ratio of the subject capacitor to the stray capacitance coupling noise in.

Even then, you have to work out what the loading impedance on both ends of the subject cap. That's then going to determine the signal voltage or current coupled into the circuit. Both impedances are effectively in parallel because of the difference in the stray coupling coupling capacitance and the subject capacitance value. That's not to say that low frequencies - 60 and 120 hz - won't have a difference from side to side of the subject capacitor, but this is heavily ameliorated by that invisible stray coupling cap's high impedance causing the coupled-in noise to be divided.

What the textbooks (and my experience, for whatever that's worth) say is that it's vastly more effective at reducing total noise to cut the noise down at the source of the noise. If it's hum you're trying to cut, reduce it by reducing hum SOURCES with good layout, good grounding, and good heater and AC power wiring as twisted or even shielded twisted pair wiring.

The other question that hits me is how do you know you're getting lower noise? Do you measure it somehow? Or is it purely subjective?

[martin manning]

The success is generally measured by the end result, with the ears. Cap orientation has the potential to reduce the introduction of noise from 60 and 120 Hz hum, and also, I would think, the reintroduction of signal from adjacent leads and components.

[R.G.]

I simul-posted, I see.

That is a great example of what I was talking about. You said:

Once I found out about the outside foil possibly reducing noise, I began to use the configuration in my builds. I have no idea if it really works or if so how much. I do believe that an amp works hard enough just amplifying the guitar signal and allowing the amp to focus on that only will produce better harmonics and maybe reduce oscillation.
I measure every resistor, cap, pot, diode and transformer to make sure it’s to spec before I install it, why wouldn’t I face the caps, I put the resistors in with the stripes facing the same way.

Quoting Stevie Wonder:

When you believe in things
That you don't understand,
Then you suffer,
Superstition aint the way

My cut on this in amp building is that there are two different realms of results in amp building (or really, anything-building), those being whether it just makes you happy to build it that way because it just makes you happier for whatever reason, and whether it actually produces better eventual outcomes to other people or measurably better outcomes to instruments. If you just like arranging the components a certain way, especially swapping Rs and Cs end for end (excepting of course, polarized caps) then go for it. If it makes you happy, and does not cause actual harm, and is worth the cost to you in time and fancy components, fine - do it. But if you are relying on better end results on a measurable basis, in the real world, you really need to know the science behind it and have some scheme of measuring "better" somehow. Otherwise, you're dealing with things like charms and amulets.

Do whatever makes you happy, but don't confuse real world results with that. Keep a clear and open mind and learn the technical stuff behind what you're doing unless charms and amulets are fine with you.

[martin manning]

Ha, I’ve often thought of that Stevie Wonder lyric in this context myself ;^).

[R.G.]

The success is generally measured by the end result, with the ears. Cap orientation has the potential to reduce the introduction of noise from 60 and 120 Hz hum, and also, I would think, the reintroduction of signal from adjacent leads and components.

Success is generally measured that way, certainly in amp building. However, the ears are remarkably unreliable instruments, subject to all kinds of internal and external deceptions.

Cap orientation may have some potential to reduce noise. The point of my post was that the question of how much reduction can really be achieved, either to AC line frequencies or to adjacent wire pickup. The question is HOW MUCH?

The answer is of course, that it can be calculated to some degree by analyzing the circuits involved. Once you do some analysis of "how big or small can THIS" be, and what does that do to the circuit performance? The biggies here are the magnitudes of the stray capacitances versus the other circuit parts, and what amount of attenuation or gain follows that. It is absolutely true that you cannot predict exactly what values of stray capacitances or inductances you have in a hand wired unit. it is also true that you can make reasonable estimates and see whether you're off by a millimeter or a light year.

There is a whole range of homework problems in freshman physics involving calculating the capactances between wires. In general, the capacitance goes down by the square of the separation. That is - layout of every single wire matters, and physical separation is potentially more effective than flipping ends of a lumped capacitor.

That whole business of reducing noise at the source if you can is important too. How effective would you say that twisted-pair AC wires, wiring to/from rectifiers, and heater wires is compared to swapping ends of film caps?

Also - If you're really concerned with stopping noise pickup in a non-electro cap, why NOT cover the thing with copper foil and ground that shield? I know that was presented as a mystical mojo effort by the OP, but it's a real-world shield, it's very close to the pickup point, and instead of draining away stray pickup into the circuit, it drains it away to a ground (if and only if you do your grounding well, unfortunately) instead.

How much reduction in noise can be had by flipping ends of the caps? No one seems to be interested in that issue, only that it's vaguely good somehow for presumably meritorious reasons, but I'm still back at how much good are you really doing? If it's not measurable, it's fine if it just makes you happy to arrange things that way.

[martin manning]

One can certainly measure the voltage produced across a cap in one orientation vs. the other on a scope, but of course the source (fingers pinching the body) is not the same as it sees in situ. At least it provides some proof of the concept.

[M Fowler]

I have never oriented my caps in over 200 builds, I'm sorry if this upsets the purist among us but I just put them in so I can read them.

The only noise I've ever corrected were a batch of silver mica 500pf and 250pf I replaced with those new big Sozo SM caps.

Mark

[sluckey]

I always orientate caps unless I'm doing a true point to point build. My choice for caps is the Xicon brown chicklets. I line them up with the writing always pointing in the same direction... Like little soldiers standing at attention, awaiting inspection.

[R.G.]

One can certainly measure the voltage produced across a cap in one orientation vs. the other on a scope, but of course the source (fingers pinching the body) is not the same as it sees in situ. At least it provides some proof of the concept.

I was thinking more about measuring the output voltage at the speaker terminals when they're properly loaded, and figuring out whether there was a difference of more than a few microvolts in noise at the end, where the human ear has some chance of hearing it. This is probably better done with a spectrum analyzer so you can interpret noise per frequency range to see if some noise is more influenced than others. It probably needs a long sample time per run to integrate out the average level of noise.

My guess is that you'll get different answers per amp, per circuit, and per wire position. The point of the exercise is to see if there is a difference with ONLY the cap's orientation. However, I suspect that the difference in ONLY cap orientation is small to unmeasurable. Not that it's not there, but I suspect it's lost under the other things that contribute noise.

The point of using instruments is that ears have different sensitivities to noise in different frequencies and at different times. I know this first hand.

This gets me back to if you like to think it's unmeasurably better and more aesthetically satisfying to arrange all the stripes on way or the capacitor leads one way, go for it. If you think it's better, it probably is, to you at least. But don't fool yourself without measuring.

Another random thought - how do you tell which lead is the outer foil? Manufacturers don't in general specify that. I don't know any non-destructive way to tell, and I suspect that any caps made on two or more different manufacturing lines may well have the outside foil be on differing leads relative to the labeling. You might well have to test each cap somehow for which is the outside foil. I can think of ways that might do this, but it's a difficult test. If you just believe every cap from the same maker is forever the same and that the manufacturer will never change their process, you're good. If you can't verify it, you're back to simply believing in amulets.

[martin manning]

Measuring the output would be interesting. The trouble is it would be a lot of work removing and replacing. Testing each one is required. The two most common ways are to connect the specimen across the input of a scope one way and then the other, or similarly across the input of an amp and listen for the 60 Hz hum. Not all cap constructions have an “outer foil.” Randal Aiken has a page on doing it. There are those among us that claim there is a huge difference. I just do it and drive on, which is no problem for a hobby builder. AFAIK this practice was common in tube radios, I have seen it myself.