Some non-electrolytic capacitors have a banded end, occasionally labeled "outside foil". These capacitors are typically made by taking a long narrow strip of insulating material and placing a strip of metal foil on both sides of it. The two pieces of foil become the plates of the capacitor, and the insulator is the dielectric. This long strip is then wound into a cylindrical shape, leads are attached to the two foils, and the entire assembly is then potted in some type of material designed to keep moisture out of the capacitor and to keep the capacitor mechanically stable. Since the capacitor is wound into a cylindrical shape, one of the foil sides is on the outside, and the other is on the inside. The outside foil terminal connection is then marked with a band to indicate the outer foil position.
Why is the outside foil marked?
Why do the capacitor manufacturers go to the trouble of marking the outside foil with a band? Aren't electrolytic capacitors the only ones where polarity matters? While it it true that polarity on a non-electrolytic capacitor doesn't matter, signal-wise, the outer foil is marked because it can be used as a shield against electric field coupling into the capacitor. In order to take advantage of the shielding properties of the outside foil, the capacitor must be connected in the circuit in a particular orientation.
Where to connect the outside foil?
The proper way to connect the outside foil is to the low impedance side of the circuit, which, in the case of coupling caps, will normally be the plate of the previous stage. If it is a bypass cap to ground, connect the outside foil to the grounded side. If it is a bypass cap from a signal to B+, connect the outside foil to B+. The outside foil will act as a shield against electric field coupling into the capacitor, so you want it to have the lowest impedance return path to ground.
For AC signals, the power supply rail is effectively at ground potential, just as the ground rail is. This is why it makes a good point to use as a shield ground. This concept is sometimes difficult to understand, but if you think about how a capacitor works, it will become clear. A capacitor has a capacitive reactance that calculated as follows:Xc = 1/(2*Pi*f*C)
where: Xc is the capacitive reactance
f = the frequency of the signal being passed through the capacitor
C = the capacitance of the capacitor.As you can see from the above equation, the frequency term is in the denominator, so as the frequency increases, the capacitive reactance decreases. Since reactance is effectively a measure of the "AC resistance" of the capacitor, the capacitor will exhibit a very low resistance at higher frequencies, while looking like an open circuit for DC and frequencies low enough to make the capacitive reactance significant. This means that the large electrolytic bypass capacitors in the power supply are effectively "short circuits" to AC signals above a certain very low frequency. For all practical shielding purposes, connecting the outer foil to the power supply rail is just as good as connecting it to ground. As a side note, electrolytic capacitors have an internal resistance that tends to rise with frequency, which can make the capacitor less than ideal as a bypass at higher frequencies. For this reason, it is sometimes a good idea to bypass electrolytic capacitors with a smaller value foil or other type capacitor.
I have seen where a well-known guitar amplifier "guru" said to connect the banded end to the grid of the next stage because it is at ground potential. This is completely wrong, because the grid circuit is a very high impedance point in the circuit. The grid of the tube itself is very high impedance, and it is usually shunted by a high resistance of 220K to 1Meg, and also usually has a large series resistance as an interstage attenuator as well. Because of this, it would make a very poor choice for electrostatic shielding. The plate, on the other hand, has an impedance equal to the internal plate resistance of the tube in parallel with the plate resistor (assuming the cathode is bypassed), which for a typical 12AX7 is around 38K total. If the cathode resistor is unbypassed, the output impedance is a bit higher, around 68K or so, depending on the value of the cathode resistor, but still well below the input impedance of the next stage. Tubes with lower internal plate resistances, such as the 12AT7, will have even lower output impedances.
What if the capacitor doesn't have a banded end?
This marking of the outside foil was very common in the "good ol' days" of electronics, but, sadly, most capacitor manufacturers nowadays do not bother to mark the outside foil, so we're left to fend for ourselves. If the capacitor has no banded end, the outside foil connection could be on either end, so there is no easy visual method to determine the best orientation of the capacitor. However, if you have access to an oscilloscope, you can do a simple test to determine which is the outside foil terminal. Set the scope up to the most sensitive vertical scale (20mV or less, preferably) and connect the scope probe across the capacitor (ground to one side of the cap, probe tip to the other). Grab the capacitor tightly with your fingers, and note the amplitude of the induced 60Hz AC signal (or 50Hz if you are on the other side of the pond). While still holding the capacitor tightly, reverse the scope leads and you should see a dramatic difference in the amplitude of the induced AC signal. 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. If you cannot see a large enough induced AC signal by holding the capacitor between your fingers, place the capacitor on top of an AC line cord (that is plugged into the mains wall socket, of course!) instead of holding it between your fingers and you will see a larger signal on the scope. If you are new at this, start with a 0.022uF cap or thereabouts, as it is easiest to see the difference between the two orientations. The induced signal is smaller at 60Hz with larger value capacitors, and is more difficult to see on the scope.
In the case of some types of capacitors, such as ceramic disks, multi-layer ceramics, or silver micas, there is no "outside foil", because the capacitor is made of a single-layer, or stacked layers of dielectric material and conductor. The orientation of these capacitors makes no difference. Also, some higher-voltage film caps (typically the 1000VDC/450VAC and higher values, such as the Orange Drop 716P high-voltage units) use a "series-wound" technique that has two separate sections, side by side, with a common "floating" connection layer, usually at the bottom of the layer stack. These caps will have no inherent shielding either.
Proper orientation of the capacitors will make the amplifier much less susceptible to outside noise, including hum, interference from fluorescent lighting, and tendency towards oscillations or frequency-response peaks and dips due to unwanted feedback from nearby signals within the amplifier, which can affect the tone of the amplifier (and is the reason why some people claim the amp sounds different if the caps are oriented in the opposite way - if there is no accidental coupling, there will be no tonal difference, but there will still be a noise benefit gained from orienting the caps the correct way).
Copyright © 1999, 2000, 2001, 2002,2003,2004,2005 Randall Aiken. May not be reproduced in any form without written approval from Aiken Amplification.