Coupling cap values

[Randall]

With respect to coupling caps in an audio circuit, I just watched an old timer's video on the subject where he has a capacitor bank and clips in different value caps to demonstrate the filtering effects of lower value = less bass. (his little bread boarded amp was transistor). He states at the end that "the minimum value capacitor suitable for passing all audio frequencies in an audio amplifier is 0.047mF"

I don't know if this is true, but if it is why do we see .1mF in coupling stages all the time? Is it when we want blocking caps with no filtering?

I ask because I am trying to figure out why A) Fender changed the PI to output caps from the usual .1 to .022 on a AB165 Bassman. Wouldn't that chop off a lot of low end? B) Why someone who supposedly knows says to change the cap to the PI from .1 to .001, again doesn't that seem like a lot of low end will be cut?

[g1]

Sometimes you want to prevent bass from making the final product sound flabby. Also you may want to cut the fundamental and leave the harmonics.
Oh, and be very careful of mF versus uF, they are very different.

[Enzo]

Back in the 1950s and earlier, before something so insanely large as a whole millifarad existed on the shelf, mf was a common way to write uf. Likewise before picofarad became popular, the unit was micro-microfarad, or mmf, later uuf. There were transition years where we saw both mf and uf. Not on the same drawing of course.

I do see some old schematics - early Gibsons, I think - where they used M instead of K for thousand, so the 250M volume control, and a 470M grid resistor, etc. Off-putting at first, but I get it, M for thousand, like Roman numerals. But I have seen folks write in asking how an amp could work with 470 meg resistors.

[Justin Thomas]

Compared to earlier Bassmans (AA864 & AA165), the lows were trimmed from the circuit earlier in the channel. They WERE still trying to make an actual BASS amp at this point, and so we're constantly trying to find the best balance of "what to put where." The selection may also have to do with what stages are getting overdriven and how much, because bass content can dramatically effect that. The lower I play my own AB165, the higher I turn the Bass knob. As I crank the volume, the Bass goes down. Volume 10, Bass stays at 2 or less - otherwise it farts like ouch.

Another variable is, not every Bassman (from the 6G6 all the way to the Bassman 50) came with the same speaker cabinet. Most of them were 2x12s of various dimensions, with I believe some coming with a 2x12-loaded cabinet that was the same size as the 2x15 that my own 1967 AB165 came with. I have a 1968 2x15. They were also basically meant to be run with 2 cabs. All of those speaker changes probably had a say, too.

Later Bassmans experiment further still with cap values...

As to your second question, well, "too much bass" is a subjective thing, right? And the rest of the circuit may make up for those cuts. I think one thing to keep in mind is that a certain cap value may or may not cut ALL of the selected frequency - as in, a .01 may cut 80Hz 6dB, and a .0068 may cut it 9dB. I did just pull those numbers out of a hat, it the point being that a cap value doesn't cut ALL of the lows out, just a larger or smaller proportion of them. If that makes sense.

Justin

[Randall]

OK. Sorry about the misnomer, but you know what I meant, right? I understand there are times when you want to cut bass, but's that not really what I was asking.

[Enzo]

Oh, I was figuring the guy you quoted was using the old form, not you yourself. Not an issue, just history.

To amplify what Justin said: when you ask isn;t that cutting a lot of low end, it takes it out of context. It cuts more than the larger cap does, but within the overall context of the amp, there may already be more bass than needed in the signal path. THAT is what we have to look at.

[Dave H]

You shouldn't really talk about the value of the coupling cap in isolation as its low frequency roll off is determined by the cap value and the value of the resistance it's driving. For example a 0.1u cap driving a 100k resistor has the same LF response as a 0.01u cap driving a 1M resistor but the caps are a factor of 10 different in value.

The plot below shows the LF response of 5n, 2.5n, 1n, 0.5n and 0.25n caps driving a 500k volume pot. It's simulating the effect of the Orange FAC switch or the switch in the Matchless DC30.

[nickb]

With respect to coupling caps in an audio circuit, I just watched an old timer's video on the subject where he has a capacitor bank and clips in different value caps to demonstrate the filtering effects of lower value = less bass. (his little bread boarded amp was transistor). He states at the end that "the minimum value capacitor suitable for passing all audio frequencies in an audio amplifier is 0.047mF"

In general this statement is complete bull. Ignore it.

I don't know if this is true, but if it is why do we see .1mF in coupling stages all the time? Is it when we want blocking caps with no filtering?

I ask because I am trying to figure out why A) Fender changed the PI to output caps from the usual .1 to .022 on a AB165 Bassman. Wouldn't that chop off a lot of low end? B) Why someone who supposedly knows says to change the cap to the PI from .1 to .001, again doesn't that seem like a lot of low end will be cut?

There are several considerations that come to mind.

The most obvious is how much low frequency you want in then final output. Next, if you cut out the low frequencies you can use a smaller output transformer core and save weight/money or push more power through a given transformer, important considerations for volume production. There is only the effect on blocking distortion that happens when you drive the power tubes hard enough to cause grid current to flow. A bigger cap will lengthen the time constants which might cause problems when yo have a loud passage quickly followed by a quiet one for instance.

[cjenrick]

if you look at all the Bassman schematics you will see that they were always playing with the negative feedback circuits,

so less NFB at the low end means you can use a smaller cap and maybe save a few cents,

also, they went from 6SL7 to 7025 to 12AT7 PI tubes in various circuits so that will affect gain and thus the choice of caps to get their desired -3 db rolloff freq, how much AC current the tubes can pump out also makes a difference, look at the Ampeg SVT with those 47K grid resistors, three in parallel for top and bottom, better be able to pump some juice, better use a honkin 12BH7a and better forget about .1 caps and go direct coupled,

coupling caps in the phase inverter will form a voltage divider with the grid resistors which are usually 220K,

you can calculate the reactance of various caps for the lowest frequency of the typical bass guitar (41 Hz - low E)

then you can calculate the percentage of voltage seen by the grids of the power tubes, which should have very much more input resistance than the 220K resistors, thus the parallel resistance factor of the tubes can be ignored,

so here are the numbers for .1 .05 .02 and .001 caps

0.1 uf = 39K at 41 Hz = 85% into grids if using 220K grid resistors
.05 uf = 78 K = 74% if using 220K grid resistors
.02 = 194 K = 53% if using 220K grid resistors
.001 = 3.9 Meg = 5.3% if using 220K grid resistors


notice that going from .02 to .1 only adds about 30% more voltage to the 6L6 grids,

i would stay away from .001, maybe the dude was talking about the PI input cap which is sometimes .001 uf.


so you need to consider the Z that the cap works into before giving a blanket statement like ".047 uf is best"

look at the coupling caps for transistor circuits, 1 uf, 10 uf, etc, because the have to work into the lower input Z of the emitter base junction,

[Dave H]

so here are the numbers for .1 .05 .02 and .001 caps

0.1 uf = 39K at 41 Hz = 85% into grids if using 220K grid resistors
.05 uf = 78 K = 74% if using 220K grid resistors
.02 = 194 K = 53% if using 220K grid resistors
.001 = 3.9 Meg = 5.3% if using 220K grid resistors

The phase shift will affect the top three percentages. For example when Xc = R (as is nearly the case for the .02 cap) it will be at -3dB (70%) not -6dB (50%)

[catalin gramada]

Just my thought. Coupling meant how the wave preserve the energy during its cycle.When work with complex wave forms is matter. Just feed the input with a square wave at 50 hz to say so.and scope the output to see how looks like. Of course a musical signal is not a pure sine so can do a better comparison just listen the diferences between, quick switching the input with a sine and square for same frequency then playing with different coupling caps to hear how it affect.When work with fundamentals we missed whole harmonic contens. Full riched harmonics signal have a different transfer function for the same impedances than a pure sine .How much the current source count in freq.response for the same impedance bridge?

[Mick Bailey]

The response of the amp needs to be considered alongside the speakers and guitar. An old Silvertone guitar sounds a lot thinner than something fitted with a 22k+ humbucker. The same thing with speakers.

[SoulFetish]

You shouldn't really talk about the value of the coupling cap in isolation as its low frequency roll off is determined by the cap value and the value of the resistance it's driving. For example a 0.1u cap driving a 100k resistor has the same LF response as a 0.01u cap driving a 1M resistor but the caps are a factor of 10 different in value.

The plot below shows the LF response of 5n, 2.5n, 1n, 0.5n and 0.25n caps driving a 500k volume pot. It's simulating the effect of the Orange FAC switch or the switch in the Matchless DC30.

[ATTACH=CONFIG]44276[/ATTACH]

This is exactly right!

[SoulFetish]

Actually, you guys all kind of nailed it on this one.
It's like showing up to a gun fight after the dust has settled.