[Dan Banquer]

"I've seen equipment of both types of construction that needed a refreshing of solder joints to restore proper operation. "
Stuart: Could you be more specific here? Was the issue a cold solder joint?, or was the joint cracked due to mechanical stress? or was there some kind of "gunk" of some type on the surface of the joint?
Thanks;
               d.b.

[Dan Banquer]

"2) I, again, checked out a simple circuit checking out the transmission effects of a 6 inch piece of wire "A" to another 6" piece of wire "B", at 10khz.
I just condensed this post to make it easier to understand.
Depending on the values of R connected between ground and each wire (between 2.2k and 12k in this example), coupling varied from approx 59 to 72 db below the reference. Even moving my hand away and to the wires caused a radical in the reading.

As you can see, Jerry, even from this simple experiment, audio frequencies can easily couple to other portions of the circuit, depending on the impedance, parts themselves become antennas of transmission. "

Steve: this sounds like capacitive cross coupling. Have you tried shielding? did you find that the coupled signal was 90 degrees out of phase with the main signal?
Yes; lowering resistor values (input impedance?)will reduce cross coupling. Shielding can be far more effective, give it a try when time allows.
d.b.

[Steve]

"Steve: this sounds like capacitive cross coupling. Have you tried shielding? did you find that the coupled signal was 90 degrees out of phase with the main signal?"

Not possible to test as I only have one channel on the sensitive differential plug in for the scope. I would suspect even a few tenths pf would be enough at this frequency to couple to another wire or component, more or less signal present depending on the impedances involved. Investigating higher frequencies and amplifiers, with much higher signal currents, would also be interesting.

The main point I am making is that if there isn't good shielding or other measures taken, coupling can occur at audio frequencies. Parts can couple to other parts, such as capacitors to wires, capacitors to tubes, etc if proper measures aren't employed.

Hope this answers your question.

[Roger A. Modjeski]

Steve,

Why do you care about 7pf of coupling to ground when the following stage will present 150-300 pf of Miller (input) capacitance also to ground? Ground planes are good things, they keep the big parts from picking up signals. The best thing you can do with a big coupling cap is to push it right down against the chassis not float it up on a PCB.

Your experiment with the 6 inch wires sounds like simple capacitive coupling to me, there are no "transmission" effects at 10 Khz in 6 inch wires.


Roger

[Steve]

We don't want to get into an argument, so I will only respond with this.

You stated 150-300pf miller capacitance???? Sorry, but not correct with my component designs.

I simply offered my opinion, which was asked. Enough said.

[Dan Banquer]

Steve: If you can get your hands on a two channel scope it's definetly worth your while. I have found it to be an invaluable development tool and a real education. For all of the DIY folks I recommend the same.
There is a fair amount of decent scopes on the used market these days, if you  have the budget for one I'd thoroughly recommend it.
                   d.b.

[Steve]

Well, I raised the signal voltage sufficiently to be able to use the dual channel plug in in my Tek scope. Interesting results.

Measuring C, a pico or less between the wires, and R1, 12k on wire "B", wire "A" 38 v P-P signal source with R2 2.2k to ground, wires approx 3/8 inches apart, the results are. (C R1 form a high pass circuit, similar to a coupling capacitor in a preamp or amp. When Xc = R, we should have a 45 degree phase shift.)

With this backround.

1) At 1khz, Large phase shift, near 90 degrees, at "B" wire, which is to be expected, since 1pf for C between the two wires is small, thus Xc is approx 160 megs reactance. Tek scope input Z is approx 1 meg with 50pf input Z. Reactance of 50pf at 1khz is approx 3.18 megs. But R1 is only 12k, so 12k is basically total R1. Xc is much greater than R1, so the high degree of phase shifting.

2) At 10khz, phase shift near 0, maybe 5 degrees as a guess, which again is not a surprise to me.

At 10khz, 1pf's Xc would be approx 16 megs, still extremely high. 50pf Xc = approx 318k. R1 is still only 12k, so 1/26th of Xc. So phase shift should still be very high, but instead, it is very very low.

If C had a reactance of only 12k or so, we should still be seeing a 45 degree phase shift (as Xc would = R1). But C (capacity between the wires) would have to be approx 1500pf, which is impossible.

It appears there are other factors besides simple capacitive coupling in this experiment.

[audiojerry]

Steve, how would the findings of your experiment have a bearing on a  circuit board layout plan?

[Roger A. Modjeski]

Steve,

Thanks for your kind reply. I'm also interested in a friendly exchange of thoughts. I do apologize if I was abrupt with you. For everybody's information, what is the Miller capacitance of your input stage, what tube do you use and what is the gain of that first stage? I admit some tubes are below 150 Pf but what is the practical Miller capacitance when you include the socket and traces around it?

Roger

[Steve]

Hi Audiojerry,

     I would be careful how the parts on a pc board (same with point to point wiring) are layed out in order to minimize signal coupling between different parts. Of course, nothing is perfect. I don't get too involved in explanations as this is an area, in itself, that requires study and experimentation (which I have over a decade of).

Sorry I couldn't be more help, but there are yearning eyes to learn and copy.

Steve, how would the findings of your experiment have a bearing on a  circuit board layout plan?

[Steve]

Steve,

Thanks for your kind reply. I'm also interested in a friendly exchange of thoughts. I do apologize if I was abrupt with you. For everybody's information, what is the Miller capacitance of your input stage, what tube do you use and what is the gain of that first stage? I admit some tubes are below 150 Pf but what is the practical Miller capacitance when you include the socket and traces around it?

Roger

I do use, exclusively, the JJE88cc tube, which is mentioned on my website, as its distortion is a factor of 10 Less, or more, than the NOS Amperex bugleboys etc that I have tested the JJs against with 2nd harmonic predominating and odds minimized.

All I will say is that the miller capacitance with all parts mounted on the board is much below what you mentioned. All other information, except specs on my website, is proprietary information.

[Dan Banquer]

Steve: if I understand you correctly you are saying that the capacitance changes with frequency? Did I understand that correctly? Or am I missing something?
             d.b.