Opinions.....outside foil

[Enzo]

At the risk of sounding cynical, outside foil is on my list of discovered properties. Guys can work on amps for years with just a meter and a scope, then they bu8y an ESR m3eter and "discover" ESR. And all of a sudden now ESr has become important. I think outside foil is the same. One can work on these amps for years, then become aware of outside foil, and overnight it becomes important.

It should not surprise that a test designed to identify the outer foil will demonstrate a difference between the inner and outer foil connections. The real question though is what difference does it make in the CIRCUIT, not in the tester.

[trobbins]

I would expect that if the cap was a typical coupling cap between say the first and second input stages of an amp, and the cap was physically close to a higher signal level cable passing by from a latter stage, then if I had the choice I would connect the outer foil to the first stage anode, rather than the higher impedance node on the other side of the coupling cap. dV/dt capacitive coupling could well be a problem if space and cable routing couldn't simply bypass the concern.

Another appropriate situation could be the non-driven grid of a long-tailed PI, which would have a physically large cap, and be in a region of the amp with high signal levels (like the PI anodes).

Another situation could be a capacitor coupled input stage grid, or the large screen decoupling cap of an input stage pentode.

[R.G.]

Let me go non-tongue-in-cheek for a few minutes. I could not agree more with both Enzo and Trobbins.

A good working definition for "magic" is "any technology you do not understand". Humans tend to group anything they don't understand into a mystical grouping, and they'll happily blame it on the weather gods, the entrails of a chicken, a shaman, or some other mentally (only!) connected thing that is observable in some way. It's a well-studied effect in the human psychology biz that if a human makes some change to something, as long as the change is not obviously, screamingly bad, the human will think it's "better" in some way.

If a cap even has an outer foil, then any effect that causes simply has to be a result of the workings of the laws of physics. And as Enzo notes, the differences are what it makes in the circuit. And as Trobbins notes, what capacitive effects it has from its position in the circuit; more importantly, what it's connected to and how sensitive that is.

A capacitor's physical construction matters in terms of the parasitics, the funny little imperfections it has. Capacitors were first planes (plates of glass, or the inside/outside of a glass jar) then for compact size they were wound strips of metal foil and insulators, first paper and later paper/plastic or plastic. The drive for small size made everything smaller; the insulators thinner, the metal "foils" as thin as a few atoms of metallization on a plastic strip. Where the external lead connects to a stack of capacitor plates or where it connects to a roll of conductors makes a big difference in whether the current going in/out has to travel long distances through thin metal layers and around wound layers, increasing inductance, or whether the maker has been clever and made the taps for the outside layers be on the sides of rolls or ends of stacks. Both of these tricks dramatically drop the ESR and ESL.

Frequency matters, too. At higher frequencies, a cap is such a small impedance that the two leads are connected by a very low impedance shunt through the capacitance itself. At lower frequencies, the impedance of the cap rises and you can see some difference in voltage between cap terminals. So it's at the low end where which cap terminal is connected to which circuit node will make a difference if the circuit lets it.

External proximity is usually from stray capacitance, especially where we're worried about "outside foils". An outside foil (which might be an atoms-thick vaporization, or not even exist in a stacked construction) would shield the inner foil from external strays if the impedances connected to each cap lead let it. Stray capacitance is usually in the few-pf range, so the impedance from an external source to the outer foil is higher than the impedance of the cap itself by the inverse of the ratios of the capacitances. For nF sized caps, that puts the ratio in the 1000s easily, around -60db as a capacitive divider best case. And that's to get the signal into the cap in question. The circuit impedances then determine what happens with that coupled signal.

This idea of an outside foil effect would actually make a good lab question for sophomore physics labs. There is an effect. WHAT effect and how big it is depends much more on the circuit and the physical layout than any issue of which way you turn the foil.

[Malcolm Irving]

External proximity is usually from stray capacitance, especially where we're worried about "outside foils". An outside foil (which might be an atoms-thick vaporization, or not even exist in a stacked construction) would shield the inner foil from external strays if the impedances connected to each cap lead let it. Stray capacitance is usually in the few-pf range, so the impedance from an external source to the outer foil is higher than the impedance of the cap itself by the inverse of the ratios of the capacitances. For nF sized caps, that puts the ratio in the 1000s easily, around -60db as a capacitive divider best case. And that's to get the signal into the cap in question. The circuit impedances then determine what happens with that coupled signal.

Agreed. But as well as the impedance ratio we also have to think about the voltage ratio. As a worst case example, the signal we want to pass through the capacitor could be a few millivolts (e.g. if there is a series cap before the first amplification stage) while the noise signal we want to screen out could be from mains wiring at 230V or 120V.

[R.G.]

The impedance ratio tells you directly what amount the coupled-in interfering signal will be reduced.

And it's always been good advice to amp builders not to run mains voltage right beside the input stage, right?

Air (in the form of distance) is always cheaper than any shielding of any kind. And no amount of outside foil identification will prevent you from shooting yourself in the foot.

[trem]

The impedance ratio tells you directly what amount the coupled-in interfering signal will be reduced.

And it's always been good advice to amp builders not to run mains voltage right beside the input stage, right?

Air (in the form of distance) is always cheaper than any shielding of any kind. And no amount of outside foil identification will prevent you from shooting yourself in the foot.

I have said it before (I think).....but yes. That is my (FWIW) understanding about this whole outside foil thing. With a "properly" built guitar amp, the outside foil is a moot point.
Once the chassis slides into its cab, the RF Shielding is in force, and the layout of the amp prohibits any induced noise from nearby Mains, B+, Signal wires, etc etc from having "any" detrimental effect on the circuit (in this case caps) as they meander their way from start to finish.
It very well may be (as Enzo touched on) more of a Marketing/Selling point for the amp builder to use.
And, as others have said already, for hobby builders doing this one amp at a time. it is no big deal.....we have to install the caps anyway, so observing the outside foil direction is easy to do.
There, no doubt, has been Thousands and Thousands (hundreds of thousands) of dead-quiet amps built over the years with no regard for outside foil.
The amps probably do not give a shit, and wonder why we care.....they're doing all the work.

[GainFreak]

The question is which way do you turn the cap?

According to Aiken's article mentioned above:

The orientation with the lowest induced signal is the one you want, and the ground lead of the scope is connected to the outside foil in that position. Mark it, and connect that side of the cap to the lowest impedance point in the circuit, typically the driving source plate when used as a coupling cap, or the grounded end if used in a shunt position.

[R.G.]

Actually, my question was rhetorical, designed to stimulate thinking.

Yep, (1) if there is an outside foil and (2) interfering signals haven't been kept away from a sensitive, high impedance node, then the outside foil should be connected to the lower impedance side of the circuit.

That begs the question - which side is the low impedance side? As noted, for a plate to grid circuit, the plate circuit is generally lower impedance than the grid circuit, so if there is an outside foil, that would be a natural place to put it.

And then there is the question of the impedance of the cap itself. The reason one uses a coupling cap is that the cap makes a DC-impervious, AC short circuit between the two things being coupled. At what frequency does the cap itself "short circuit" the interfering signal so that both legs of the cap have essentially equal voltages induced by the interfering signal through the stray capacitance to the source of the interference?

The "plates" of the coupling cap are quite large and very close together. Capacitance goes up linearly with the area of the plates (*for parallel plates; other geometries get complicated enough to use as sophomore physics homework questions) and goes down as the square of the distance between plates. The stray coupling "plate" is in general a wire; worse would be another cap carrying a large signal with the signal on its outside foil; this produces more surface area for the stray capacitance.

Connecting the outside foil to a low(er) impedance point does cut down on the interfering signal, but how much good this does depends on the specifics of the impedances involved and the ratios of the capacitance to the stray capacitance. Two worst-case coupling caps, like a PI output right next to and touching an input cap side by side would provide a severe test. But consider that if they are nearly touching, that simply bending them apart causes many times more separation between them, and reduces the stray coupling by many-times squared. Proper layout (that is, putting the PI over away from the low signal stuff) would be even more effective - vastly. If you absolutely had to put a PI cap next to an input cap for some reason, outside foils practice would only do you so much good. In that case, introducing a grounded foil shield between the two, or around the bellies of both, would be dramatically better than relying on "outside foils" to cut stray capacitance coupling.

I'm synopsizing heavily because I'm too lazy to draw up and annotate the drawings. I hope this is coming through clearly.

[Malcolm Irving]

Once the chassis slides into its cab, the RF Shielding is in force, and the layout of the amp prohibits any induced noise from nearby Mains, B+, Signal wires, etc etc from having "any" detrimental effect on the circuit (in this case caps) as they meander their way from start to finish.

Yes, the chassis shields the internals from external electric fields, and if wires inside the chassis are placed close to the chassis there is some shielding effect from that, but there is stray capacitance and mutual inductance among the wires and components inside the chassis. Good layout can certainly minimise the unwanted stray effects, but I don't think it prohibits them.
I'm not saying that bothering about where to connect 'outside foils' has an audible effect - I'm just saying that as far as I'm aware we haven't yet proved that it doesn't.
Two (otherwise) identical high gain amps, one with all the caps the 'right' way round and the other with all the caps the 'wrong' way round would be interesting!

[R.G.]

I'm just saying that as far as I'm aware we haven't yet proved that it doesn't.

I'm pretty sure we (that is, the whole human race) has not yet proved that there isn't a rain god, but that question doesn't get a lot of play. We have not proved that the color of the tolex doesn't make a difference to sound quality either.

Then there's the whole issue of it being logically impossible to prove a negative - no one is every going to prove that it doesn't.

There is a book that I enjoyed reading a lot. It's by Michael Shermer: "Why People Believe Weird Things". I'm not sure that I agree with all of it, but there is sure a bunch of thought provoking reading there. I highly recommend it.

[Enzo]

Good choice. On my shelf alongside the Demon Haunted World and other classics about thinking.

[trobbins]

Another aspect to consider is historical.

Grid-leak bias was a lot more common in vintage days - hence the sensitivity of the 'receiver' side node to capacitive coupling from a nearby capacitor's outer foil would have been substantially higher.

Also, capacitor parts were physically a lot larger, especially the old paper-foil encapsulated in wax style axial leaded parts, and in a lot of commercial equipment (radio's and amp), the packing density of wires and parts was sometimes atrocious, or the physical location of those old wax caps would blanket the top of an octal base.

[Tom Phillips]

To add some food for thought, check out the attached stock 1964 Fender Showman gut shot. The outside foil of the blue molded caps, based on the banded marking, is very consistently placed in the pictured amp and in all other Fender models of the period. Someone decided to do that at the time for some reason. An interesting bit of history, lore or a procedure that was blindly copied or???

Another interesting fact is that I once did the outside foil scope test on a batch of the vintage blue molded tubular caps and I think about 15% of the batch were incorrectly marked by the manufacturer with respect to the outside foil band. Therefore, even after someone back in the day carefully installed the caps per the assembly instructions, some of them were oriented "incorrectly."

As I said. Just food for thought.

[trem]

.....Now that IS interesting.
So next time a guy (with a snooty attitude) asks me..."Why do you think Leo bothered to orient all those Blue Molded Caps with the line a certain way.?".........I can answer..."Well, have YOU ever tested that line.?"
I think I am getting all worked up over nothing...

[J M Fahey]

That is how I have "always" done it also. I chose a method, and made sure all the parts looked the same.
I have never worried about the outside foil, it probably Is Not Audible in a guitar amp. I was just wondering if there is a legitimate way to identify it, if the cap is not marked.

Well, there's always the "Finger-O-Meter" way



1) you get any amp you have on the bench (you won't *build* something for this test) , you disconnect the wire to the first grid coming from input jacks, wire a 100k to 1M resistor from grid to ground (if there's not one there) and connect the mystery cap in parallel with it.
2) you apply your NASA calibrated Finger-O-Meter[tm] to the middle of the cap body.
It may hum a lot or very little (try it both ways), the highest hum position has the outside foil connected to grid, while the less humming one has it grounded.
Mark the foil end with some Sharpie and test all others.