Twisted lead dress for pentodes?

[jimboyogi]

Hi all,

In Merlin's pre-amp book, he recommends twisting the leads together for each triode gain stage. The benefits are for physical strength of the lead bundle, and some extra capacitance between the electrodes to reduce high frequency gain, and therefore reduce the chances of HF oscillation.


I have been following this advice with my builds, and so far I'm happy with the way it is working out.


My question though is does this advice still hold true for pentodes?

I'm designing an input stage using a pentode as V1.
The gride and cathode wires travel via shielded cable from the front of the amp to the pentode socket.
The plate and screen wires run from the board to the pentode socket.

Should I be twisting the plate and screen wires together, to increase the capacitance, and rolling of HF gain? Or will this cause other craziness that I won't like ?!

[Steve Conner]

The screen is usually bypassed to ground by a capacitor. So from an AC point of view it is ground, hence twisting it together with the plate wire won't do anything spectacular.

[jimboyogi]

Thanks Steve.

I have the screen bypassed to the cathode with a 0.1 uF cap. So if I understand correctly twisting the plate and screen wires together is akin to a low value plate to cathode capacitor. Which should help prevent HF/RF oscillation.

[soundguruman]

Hi all,

In Merlin's pre-amp book, he recommends twisting the leads together for each triode gain stage. The benefits are for physical strength of the lead bundle, and some extra capacitance between the electrodes to reduce high frequency gain, and therefore reduce the chances of HF oscillation.


I have been following this advice with my builds, and so far I'm happy with the way it is working out.


My question though is does this advice still hold true for pentodes?

I'm designing an input stage using a pentode as V1.
The gride and cathode wires travel via shielded cable from the front of the amp to the pentode socket.
The plate and screen wires run from the board to the pentode socket.

Should I be twisting the plate and screen wires together, to increase the capacitance, and rolling of HF gain? Or will this cause other craziness that I won't like ?!

Plate and grid...

That is quite correct. Twisting the plate and grid adds interelectrode capacitance that causes a small amount of negative feedback at higher frequencies. This can prevent the tube from ringing with microphonic oscillations.
I don't see the advantage of screen and plate with a pentode. You certainly don't want to get the filament wires anywhere near the grid, so those won't twist together.
Marshall used shielded wire. The center conductor is the grid. The shielding connects to the plate.
Fender uses a 7pf capacitor between the plate and grid.
No matter what method the voltage rating of the wire is important, or the voltage rating of the capacitor.

[Wombaticus]

Plate and grid...

That is quite correct. Twisting the plate and grid adds interelectrode capacitance that causes a small amount of negative feedback at higher frequencies. This can prevent the tube from ringing with microphonic oscillations.

Soundguruman, once again you're conflating the notion of microphonics with ringing or oscillation resulting from inadequate phase margin in a feedback network -- whether this network is intentional or not. Once again, these are very different and unrelated phenomena, and it's unhelpful to confuse the two in the minds of your readers.

While I'm here grousing, it's worth noting that the intrinsic interelectrode Cg-a capacitance of pentodes is so much lower than triodes, that even with a large grid stopper there is usually little high-frequency roll-off. This is exacerbated by the fact that gain for power pentodes is typically low, so the Miller effect won't multiply whatever capacitance there is by very much. So, if anything, twisting together grid with plate runs should be even more beneficial for pentodes than for triode gain stages - at least in a MI amp context where bandwidth is intentionally limited.

[soundguruman]

My post is absolutely correct. The only one confused is YOU.

[J M Fahey]

Twisting the plate and grid adds interelectrode capacitance that causes a small amount of negative feedback at higher frequencies.

You are talking ELECTRICAL feedback here.

This can prevent the tube from ringing with microphonic oscillations.

You are talking MECHANICAL feedback here.

The only one confused is YOU.

You are talking to the mirror, aren't you?

Oh well.

[tubegeek]

In my experience (which - in this area - has come mostly from working with the high-gm triode 6C45P) a tube that is prone to parasitic oscillations will "trigger" off of various things: microphonics and hum being two common culprits.

So I suppose that might be what soundguruman is talking about - a tube that will oscillate when triggered by mechanical microphonics might calm down if some electrical NFB was applied at HF.

Just a guess, because he doesn't explain himself, but I can see it as possible. I don't have a lot of experience using pentodes, though.

[diagrammatiks]

tried to work this out but my brain melted.

the grid pins and the plate pins of most tubes are right next to each other. if there was any feedback it would be unavoidable right at the pin. or inside the tube.

therefore it must be a magic ball or some sort.

[Wombaticus]

In my experience (which - in this area - has come mostly from working with the high-gm triode 6C45P) a tube that is prone to parasitic oscillations will "trigger" off of various things: microphonics and hum being two common culprits.

So I suppose that might be what soundguruman is talking about - a tube that will oscillate when triggered by mechanical microphonics might calm down if some electrical NFB was applied at HF.

Just a guess, because he doesn't explain himself, but I can see it as possible. I don't have a lot of experience using pentodes, though.

Soundguruman has previously demonstrated that he some funny ideas about microphonics, including denying that they are even a real phenomenon. Still, as with everything, there's at least a grain of truth in what he says. My understanding of the situation is more-or-less along the following lines:

1. Tubes (and other components) are most susceptible to mechanical-acoustical coupling at lower, mostly bass, frequencies. Sound waves at these frequencies are more energetic than high frequencies, and physical structures such as tubes are more likely to resonate at a lower frequency than at high frequencies; If mechanical coupling occurs, it's almost always at lower frequencies. Limiting bass frequencies *is* a recognized method of controlling microphonics that is attested to in the wisdom literature. This method was often employed in non-hi-fi applications such as portable radios. Incidentally, limiting bass is also often an effective means for controlling undesirable acoustic coupling in the context of reverb tanks.

2. A couple of things happen as the relative geometry of the internal elements shifts that also explain the noises one hears when you (for instance) tap a tube with a pencil. First, and most significant, there are instantaneous variations in the gain of the tube as things rattle around. In the case where the microphonic input come from the output of the amp, these changes result in audible variation in the volume at frequencies already present in the signal.

3. There *is* also a small change in the inter-electrode capacitances of the tube as things move about, and I suspect that this is what Soundguruman is on about. However, if you once again look at the old literature, this is never identified as a problem associated with audio frequency gain stages, and I have *never* seen adding additional grid-plate capacitance advocated as a method for controlling this. Notably the squeal that *does* sometimes result from a microphonic tube in a high gain position tends to be very much within the audio band. It would be pretty much impossible to add sufficient capacitance to control this without totally bonking your overall frequency response. Indeed, if you look at RDH4 and other sources, this was largely an issue relating to frequency stability of tube stages that had been intentionally set up as RF oscillators or were otherwise part of tuned circuits where even small variation in the frequency could be problematic. In any case, in this circumstance the inter-electrode capacitance most frequently mentioned as critical in this application is the heater-cathode capacitance, rather than the grid-plate capacitance. It seems to me exceedingly unlikely that the small variations in Cga that occur upon microphonic coupling could have much influence on the stability of an audio amplification stage, unless the stage was very very marginal in stability already. But there, admittedly, is the crack in the door for Soundguruman's theory.

[diagrammatiks]

Soundguruman has previously demonstrated that he some funny ideas about microphonics, including denying that they are even a real phenomenon. Still, as with everything, there's at least a grain of truth in what he says. My understanding of the situation is more-or-less along the following lines:

1. Tubes (and other components) are most susceptible to mechanical-acoustical coupling at lower, mostly bass, frequencies. Sound waves at these frequencies are more energetic than high frequencies, and physical structures such as tubes are more likely to resonate at a lower frequency than at high frequencies; If mechanical coupling occurs, it's almost always at lower frequencies. Limiting bass frequencies *is* a recognized method of controlling microphonics that is attested to in the wisdom literature. This method was often employed in non-hi-fi applications such as portable radios. Incidentally, limiting bass is also often an effective means for controlling undesirable acoustic coupling in the context of reverb tanks.

2. A couple of things happen as the relative geometry of the internal elements shifts that also explain the noises one hears when you (for instance) tap a tube with a pencil. First, and most significant, there are instantaneous variations in the gain of the tube as things rattle around. In the case where the microphonic input come from the output of the amp, these changes result in audible variation in the volume at frequencies already present in the signal.

3. There *is* also a small change in the inter-electrode capacitances of the tube as things move about, and I suspect that this is what Soundguruman is on about. However, if you once again look at the old literature, this is never identified as a problem associated with audio frequency gain stages, and I have *never* seen adding additional grid-plate capacitance advocated as a method for controlling this. Notably the squeal that *does* sometimes result from a microphonic tube in a high gain position tends to be very much within the audio band. It would be pretty much impossible to add sufficient capacitance to control this without totally bonking your overall frequency response. Indeed, if you look at RDH4 and other sources, this was largely an issue relating to frequency stability of tube stages that had been intentionally set up as RF oscillators or were otherwise part of tuned circuits where even small variation in the frequency could be problematic. In any case, in this circumstance the inter-electrode capacitance most frequently mentioned as critical in this application is the heater-cathode capacitance, rather than the grid-plate capacitance. It seems to me exceedingly unlikely that the small variations in Cga that occur upon microphonic coupling could have much influence on the stability of an audio amplification stage, unless the stage was very very marginal in stability already. But there, admittedly, is the crack in the door for Soundguruman's theory.

even from the other thread it's pretty clear that he is clearly confusing 2 very different things.

You can have a microphonic tube with perfect lead dress pop on tap with very little gain and you can make a perfectly tested and quiet tube squeal without it becoming microphonic.

In relation to the OP's question and how twisting the wires would somehow cause feedback in the circuit...

The plate is an elevated AC signal running on top of a significant DC component. The level of interaction it would have with the grid input signal is minimal even if they were twisted.

On a pentode, the grid is already literally balls deep inside the screen. Those wires can be twisted.