ground side for mallory 150's

[Enzo]

Maybe not here so much, but in discussions I often say "I'd rather be PROVEN wrong than to GUESS I am right." Meaning that I would then KNOW the truth instead of assuming it. So I am all for experimentation. When we are all rationalizing parts of theory to fit our notions, then thought experiments are fair game. Obviously empirical data trumps ideas.

AS a thought exercise, if there is RF on the plate, would it not flow freely through a coupling cap? If there is an ambient RF near a stage (from where?) how exactly does the outer foil shield the cap? WOUld not any RF picked up on the outer foil flow freely through the cap to the other side? In other words, at RF, what is the impedance of the cap? Pretty much zero if the cap passes audio, I would think.

In my mind, the only difference I can come up with is the different freqs on either side of the cap. If the cap was a 500pf coupler, the signal on the input side of it would be fuller range - more bottom - than the downstream side. So if any signal were picked up by the outer foil if it was connected to the grid, then some of that bottom could crostalk past the cap. Seems like a tiny amount. But turned around, then that greater area outer foil would radiate more low end. I guess.

I don't see innards of amps all that dusty.

Of course I could be wrong, and I am typing far more words relatively speaking than my concern over the issue really merits, but I have the impression this is like some other things like skin effect, that people learn of then try to make fit into guitar amp theory.

I totally agree that outside foil matters in actual RF circuits, but unless there is some grounding issue at the input jack, I don't see RF problems in guitar amps. The exception is when unstable amps make their own "RF" like 50kHz or 100kHz.

[dai h.]

maybe something to do with the way many film caps are constructed (rolled up)? You have the innermost part of the roll, then many layers sort of like caps in series. Capacitance in series equals less capacitance getting though and maybe that is the shielding effect?

I also remembered some old polystyrenes with a red band on one end. Maybe this signifies the outer foil. Polystyrenes are also one of the cap types working well at high freqs., so maybe there is some connection.

I don't see innards of amps all that dusty.

I left a 50W on for a while (weeks?) without tubes to keep voltage on the electros (which was kind of dumb). Anyway I opened up the chassis after some time and there was a quite noticeable layer of dust on the parts. The source I read was saying something about air cleaners working on the same principle (at work with the dust attraction).

[Tobacco Soup]

I just go by what some of the guru's have posted here in the past, long before the forum interface/format was changed to this one.

The RF guys on here will tell you point blank that it matters for RF. Since that is what we are trying to guard against, RF interference, then it matters somewhat for us.

If it weren't theoretically valid, why on earth would penny-pinching manufacturers bother to mark the outside foil? Find me an old-school EE that would purchase a million caps based upon hype, and well...

[Enzo]

Penny pinching cap makers mark it because it only costs once to set up what the label says, and adding a line at one end really does not add to the cost. Tha caps come from the machine all facing the same way, so the difference is printing the label with or without a line at one end.

They bother because they have no idea who will be using the cap after it is made, might just BE someone working in RF. An engineer isn't buying a cap on hype, he buys what he needs. If his circuit requires attention to outer foil, then there it is marked on the part. The people who set up the pick and place machines for the board line will set their gear to orient the part accordingly. On the other hand, the people who make doorbells will probably just ignore the matter altogether.

And here we are not working with RF, it is something we don't want. That is different.

Look, it won't hurt anything to follow your rule, so go ahead.

dai, I understand how they are made. The entire point is that when two strrips of foil are rolled, at the end, ONE of them will be outside the other. That is the shielding referred to - the one outer foil can "shield" the one that wound up inner. I think trying to consider each turn as somehow an independent cap in series with others is maybe try ing to outthink ourselves here. If the cap is rolled out, it will still be a cap. Rolled up the two foils still make a cap, but now both sides of the foils see the other foil nearby. I would think the capacitance woul dbe a lot larger rolled up than spread. Nothing is in series.

Yes, the band at one end or sometimes just a stripe on the front is the indicator of outside foil

I work on video monitors, and they ARE dust magnets. But mainly just the parts with high voltage on them - I mean thousands of volts. Stuff nearby starts to accumulate the dust too as it is attracted to the HV parts, the LV parts just get in the way and gather dust. Combine that dust with the wax of old paper caps, and working on old TV chassis covers you with this supreme gunk that is hard to wash off.

Air cleaners expose a high voltage anode to the air, and the dust is attracted to it.

[dai h.]

my understand is there is a shielding effect and not a shield.

on the other point of seeing little tiny portions in series--if that is incorrect, how does the shielding effect take place (only for RF, only for caps that are rolled and not say a cap with different construction like a mica--as I understand it). L(inductance) from being rolled up?

going back to "hearing a reduction of RF buzz when orienting correctly", would it be buzz (harmonics of the rectified power line frequency--100Hz or 120Hz) radiating from the rectifiers to first filter to ground (that loop)? Would correct orientation actually effect significantly less noise coming in? (My gut/intuition doubts this. I'm guessing if there was buzz radiating in, it'd be the same either way.) But I'm not sure.

[tbryanh]

I'm thinking that "shielding" might be the wrong concept.

The outer foil might be acting more like an "antenna" than a "shield."

Having the outer foil connected to the plate side (lower impedance to ground) means that more RF will be diverted to ground from the "antenna" than when the outer foil is connected to the grid side.

Capacitors can change characteristics at high frequencies. I have heard that electrolytic capacitors act more like resistors than capacitors at high frequencies, so some circuits have little bypass caps installed across electrolytic capacitors to deal with this. I don't know what frequency this becomes an issue, but it is probably way above the audio range.

But applying this concept to the coupling capacitor might shed some light. At some high frequency, the coupling capacitor probably starts acting more like a resistor than a capacitor, and this is probably when the orientation could matter. But these frequencies are probably a non-issue to the audio stages other than maybe the first preamp stage.

Ezno is well seasoned in the field, and he pointed out that RF is not an issue past the first preamp stage. That sounds like good news, and I tend to believe it.

I am pretty new to vacuum tube audio, and the only caps I have bought are the 716P orange drops. These are supposed to be some of the best caps, and there is no outer foil marking on them. As far as I know, these caps are made specifically for audio, so this tends to imply that it doesn't matter which way the caps are installed.

In spite of that, I am going to jump on the bandwagon and install the outer foils to the lowest impedances. Morgan Jones of Building Valve Amplifiers says to do it, and as mentioned Aiken amps and others say to do it too.

One reason I am going to do it is if I have any problem with RF, I will know it is not because of the orientation of the capacitors, so I can rule that out right off the top.

The other reason is that I know I am using the best methods possible when building amplifiers regardless if the methods are needed or not.

It only takes a little more time, and it does not cost any more money.

As a side note, I have heard that a guitar amp can be substituded for an O'scope. Just ground one end of the cap and hook the other to the tip of the guitar cord (while the other end of the cord is plugged into the amp). Sqeeze the cap and note the amount of hum. Reverse the cap and do the same. The side of the cap that causes the most hum is the side connected to the outer foil.

[Tobacco Soup]

The fact that it produces an audible effect when trying to determine which end is which, begs the question...

I don't necessarily promote the practice, as I haven't performed an A/B comparison. However, it's like chicken soup when you have the flu: it won't hurt, but it might help...

[dai h.]

interesting. Thanks.

[Enzo]

A quick look through my drawers found some Spragues banded for outer and some not.

I think it matters a whole lot more at RF, which indeed is not what we work with. If I had a pentode in an RF converter circuit or an IF circuit, and I needed to bypass the screen, I would make outer foil grounded.

As I said though, as far as RF is concerned, a 0.022 cap is just a wire, so whichever end of the cap has the outer foil doesn't matter. WHatever would come through the cap one way would be picked up by the cap the other way.

And I think that the plate circuit versus the grid circuit impedances might be something like 50k versus 220k, so NEITHER is ground or much of an impediment to any RF. My shop is very close to an AM radio station, so there is a fairly high ambient Rf field here. I can watch their modulation envelope on my scope without even trying. but the only time it gets into the guitar amp circuits is when I touch a sensitive point - like a preamp grid - with a metal object.

Elytics may become opaque to Rf at some point, but I also would say that was way above audio. Ecaps were used as speaker outputs in many single rail amps over the years, and they were expected to pass the entire audio range. And they were large caps.

how does the shielding effect take place

All the concept says is that since the outer foil is wrapped outside the inner, the outer shields the inner from picking up stuff from outside. Whether I accept that myself is a separate issue, but that is the reported phenomenon. Rollup two sheets of paper into a tube. One of them will completely surround the other when you are done, thus shielding it. Nothing more complex than that.

Scope, amplifier, even an AC voltmeter. Whichever end of the cap generates the most signal when you wrap your fingers around it, that is the outer foil.

This will be the rolled cylinder type caps. The square ones like micas are alternating flat plates, connected to one end or the other. No inner or outer there.

It certainly will not hurt the construction to pay attention to the foil orientation, and to me it makes a neater and more professional build if the parts all face the same way. I view this discussion as an interesting chat, not a pronouncement as to how anyone shoul;d install caps.

[tbryanh]

I'm thinking that "shielding" might be the wrong concept.

The outer foil might be acting more like an "antenna" than a "shield."

I think I should probably correct myself here a little bit.

Take the example Ezno made where the capacitor is connected from the screen grid of a pentode to ground.

When the outer foil is connected to ground and the inner foil is connected to the screen grid, the outer foil shields the inner foil to prevent the inner foil from acting like an antenna to create a feedback loop that might cause a parasitic oscillation in the pentode.

In the reverse connection, when the outer foil is connected to the screen grid and the inner foil is connected to ground, the inner foil does not shield the outer foil, so the outer foil is free to act like an antenna to create a feedback loop that might cause a parasitic oscillation in the pentode.

As mentioned earlier, the impedances the coupling capacitors are connected to between the stages of the audio circuits are high impedances, not low impedances like the ground connection, so it is questionable if the outer foil can provide shielding here. But, as was also mentioned earlier connecting, the outer foil to the lowest impedance (plate side, not grid side) will allow shielding to take place if shielding is possible here.

… the 716P orange drops. These are supposed to be some of the best caps, and there is no outer foil marking on them. As far as I know, these caps are made specifically for audio, so this tends to imply that it doesn't matter which way the caps are installed.

I think I should probably correct myself here a little bit too.

To say there is no outer foil marking on the 716Ps migh not be true.

It is correct that there is no band or stripe on the cap body near the lead that is connected to the outer foil, but I have heard (maybe in this thread) that the lead to the right of the printing on the body of the capacitor is always connected to the outer foil in the 716Ps.

(In other words, if you hold the capacitor in a position so the the part number and other printing on the cap is facing you and the printing is right-side-up so that you can read it, the lead to the right is the one connected to the outer foil.)

I have not tested any of the caps yet, but if this is true, then the outer foil is marked. This means that the conclusion (because the caps are not marked, it doesn’t matter which way the caps are installed) is not a good conclusion (not that it was such a great conclusion anyway).

If the caps are marked, just because they are used frequenly in audio circuits does not mean that it matters which way they are connected when they are used in audio circuits.

The caps are also apparently used in high-speed pulse circuits and possibly in RF circuits too. These circuits are probably where the orientation of the outer shield matters. So it might be that the marking on the capacitors are there for when the capacitors are used in these circuits.

I am going to jump on the bandwagon and install the outer foils to the lowest impedances

Yes, I still am going to do this.

[Bruce / Mission Amps]

...

It is correct that there is no band or stripe on the cap body near the lead that is connected to the outer foil, but I have heard (maybe in this thread) that the lead to the right of the printing on the body of the capacitor is always connected to the outer foil in the 716Ps.

(In other words, if you hold the capacitor in a position so the the part number and other printing on the cap is facing you and the printing is right-side-up so that you can read it, the lead to the right is the one connected to the outer foil.)

I remember years ago someone actually talking to Sprague about this "ink printed in a direction means something" and the guys from Sprague thought it was very funny that anyone could possible think they'd have time and extra money for labor to figure that out... in other words. it is TOTALLY random and with no merit.

[Enzo]

Wouldn't it depend on the machinery? I mean someone would have to figure out which end was which only once. The parts all come out of the machine the same way at some point. Now if they get tossed into a shaker bin before labelling then it is random, but if they are printed as part of the creation process then all would face the same way. If you buy taped resistors they all face the same way for example.

[tbryanh]

Well, I decided to test some of my 716Ps.

I have a quantity of 5 of each value, and I chose to test 2 values, one large value, and one small value.

It was random.

By the way. Using a guitar amp to test them works good.

[Bruce / Mission Amps]

Well, I decided to test some of my 716Ps.

I have a quantity of 5 of each value, and I chose to test 2 values, one large value, and one small value.

It was random.

By the way. Using a guitar amp to test them works good.

Oh, I am so glad to see you do that... I have too, many times
The same thing happens when I test all my O'drops, IC axials, Mallory axials or SK poly axials as a matter of fact.... almost ever cap I check this way comes out the same way ... totally random in which lead is connected to the outside... however, as silly as it is, I still always check and install the outside lead to the lowest impedance side of the circuit.
Does it matter?
I don't know... probably not, except in those very high gain cascading situations where every little bit of stability helps.