Some grounding questions

[chas.wahl]

Here's Merlin Blencowe's advice for circuiting stages and getting the grounding in the right place:

The highlighting is mine and I presume that what he means in the text is what's shown by the highlight I added between one stage's coupling cap and the path to the next stage's grid and cathode ground.

Now here's part of a circuit I'm working on:

In the case of V2b (the last preamp stage before the PI) I'm wondering what the function of the 470k resistor highlighted is, and whether (unlike the original amp's schematic, which shows all the grounds under V2b gathered as shown) the ground side of that resistor shouldn't be grounded instead to the same node ("C") as the PI is, since that's the "next stage" receiving the signal output from V2b.

Also (and I feel more confident about this one) at V3a, whose signal output is going through the reverb tank and to the grid of V3b, shouldn't the ground side of the Reverb switch be grounded at the cathode ground of V3b? The original schematic shows all the grounds at nodes D and below V3b ("to F") gathered together, like one big happy reverb family, but V3a and V3b are each on their own node of the power supply (D and F), so it seems to me that they should be treated as two stages that come from different filter cap sources.

And a final one, that I noticed while reviewing Blencowe's "Grounding" article:

Unfortunately, this tantalizing suggestion from the ValveWizard doesn't say how to adjust resistor values to make such a change. Should one halve the dropping resistor value in the B+ and put a resistor of equal value (half) and wattage in the ground leg that's highlighted? I'd think about a choke, but this amp has current such that it would probably be a fairly large one.

 

[sds1]

In the case of V2b (the last preamp stage before the PI) I'm wondering what the function of the 470k resistor highlighted is, and whether (unlike the original amp's schematic, which shows all the grounds under V2b gathered as shown) the ground side of that resistor shouldn't be grounded instead to the same node ("C") as the PI is, since that's the "next stage" receiving the signal output from V2b.

470k is the AC load for V2B (it also participates in high pass filter with C18), I agree that according to Merlin that ground connection should terminate at star C.

Also (and I feel more confident about this one) at V3a, whose signal output is going through the reverb tank and to the grid of V3b, shouldn't the ground side of the Reverb switch be grounded at the cathode ground of V3b?

I'm not sure this one matters so much, since the ground there is intended to shunt signal before the reverb tank, as opposed to being a return path in an active part of the circuit.

 

[chas.wahl]

Thanks, @sds1. So when the Reverb switch is open, that's when the reverb is on?

 

[sds1]

Thanks, @sds1. So when the Reverb switch is open, that's when the reverb is on?

Yes that's correct!

 

[Lowerleftcoast]

Unfortunately, this tantalizing suggestion from the ValveWizard doesn't say how to adjust resistor values to make such a change. Should one halve the dropping resistor value in the B+ and put a resistor of equal value (half) and wattage in the ground leg that's highlighted?

Correct.

 

[chas.wahl]

Thanks you guys!

 

[Lowerleftcoast]

In the case of V2b (the last preamp stage before the PI) I'm wondering what the function of the 470k resistor highlighted is, and whether (unlike the original amp's schematic, which shows all the grounds under V2b gathered as shown) the ground side of that resistor shouldn't be grounded instead to the same node ("C") as the PI is, since that's the "next stage" receiving the signal output from V2b.

When pondering the DC ground, it is not helpful to consider "the "next stage" receiving the signal output". Just look at the coupling cap as a open circuit to DC. DC can't pass (or shouldn't pass) through the cap. The DC current has to find the way back to the source.

Looking at your example... any DC flowing through the 470k can not be from the previous stage. There is an open circuit* between the previous stage and the 470k due to the coupling cap. Any DC flowing through the resistor must be from the PI. Since the resistor has DC current from the control grid, it is a grid leak resistor and the DC current must be from the B+ servicing the PI.

* Looking at a capacitor as an open circuit to DC. AC signal can still pass.

 

[chas.wahl]

Looking at your example... any DC flowing through the 470k can not be from the previous stage. There is an open circuit* between the previous stage and the 470k due to the coupling cap. Any DC flowing through the resistor must be from the PI. Since the resistor has DC current from the control grid, it is a grid leak resistor and the DC current must be from the B+ servicing the PI.

* Looking at a capacitor as an open circuit to DC. AC signal can still pass.

Now I may be confused again. On the basis of the above, where should the 470k resistor be grounded? Is the above just explication ("think about the DC current, dude") of why it's better to ground it at node "C", or a correction to that advice by @sds1, suggesting leaving it at "E"?

Based on further inspection, I think it's the former: to me, although the 470k resistor is shown at V2b, it seems it's really a grid leak resistor for the PI V4a grid, in the same way that the 47k resistor to the right of the 470k resistor in question is a grid leak for V3a's grid. It's a bit confusing, since it could be moved over to the left of the PI grid, and two long wire runs from V2b to V4a shown in my markup above could be reduced to 1 (the wire from the coupling cap between them), with the 470k resistor hung off that right at V4a grid to the ground node "C". Did I get that right?

Comme ça?

 

[Lowerleftcoast]

Is the above just explication ("think about the DC current, dude") of why it's better to ground it at node "C", or a correction to that advice by @sds1, suggesting leaving it at "E"?

I hope you are leaning towards grounding it at node *C*.

it seems it's really a grid leak resistor for the PI V4a grid, in the same way that the 47k resistor to the right of the 470k resistor in question is a grid leak for V3a's grid. It's a bit confusing, since it could be moved over to the left of the PI grid, and two long wire runs from V2b to V4a shown in my markup above could be reduced to 1 (the wire from the coupling cap between them), with the 470k resistor hung off that right at V4a grid to the ground node "C". Did I get that right?

Yes, you have it right.

 

[sds1]

although the 470k resistor is shown at V2b, it seems it's really a grid leak resistor for the PI V4a grid

Good call, this is a much more relevant answer than I provided.

 

[chas.wahl]

Great! Thanks.

 

[andrewRneumann]

Cool schematics! I wish I had software that made such pleasing schematics.

R25 should also join nearby to the relocated R17?

I haven’t seen a paraphase exactly like this before. Did you get that from another design?

 

[King Fan]

I haven’t seen a paraphase exactly like this before. Did you get that from another design?

I have a hunch this may be from his current devotion to this amp (schematic lifted from his well-researched thread about it)….

 

[chas.wahl]

Cool schematics! I wish I had software that made such pleasing schematics.

R25 should also join nearby to the relocated R17?

I haven’t seen a paraphase exactly like this before. Did you get that from another design?

It's an Ampeg design from the early '60s, the first production amp with onboard reverb & tremolo, I believe: Reverberocket. I see what you mean about R25, since it's the grid leak for the "cathode-side" PI tube. But whether they join at the cathode resistor or at node "C" doesn't make any difference, does it?

I'll keep it in mind when doing the layout, which I've only just begun.

My schematic interpretation of the Ampeg one is drawn using AutoCAD; I'm in architecture when not doing this, so it's just a tool I know. I use it for layouts too.

 

[Lowerleftcoast]

I see what you mean about R25, since it's the grid leak for the "cathode-side" PI tube. But whether they join at the cathode resistor or at node "C" doesn't make any difference, does it?

Blencowe writes in Chapter 15.7 of his grounding article, "All power and ground connections are now made much closer to their relevant smoothing capacitors, thereby minimising the interaction of audio and power currents... truncated ...and all grid leaks are now connected to their respective cathode resistors, rather than directly to the bus.". All of his diagrams show separate paths for each cathode. So, he has some concern for the AC current, keeping each separate.

 

[chas.wahl]

I think I get it. Does this also mean that I should tie the tail of the 5.6k grid leak at the input (V1a) to V1a's cathode, rather than to the "chassis ground point" where I have it shown now? And that on V4b (the "cathode" side triode of the PI), I should connect not only the 470k grid leak, but also the 500 uF bypass cap paralleling it, to the tail of the cathode resistor shared by the two PI triodes?

 

[chas.wahl]

And another one, for confirmation:

View attachment 999500
Unfortunately, this tantalizing suggestion from the ValveWizard doesn't say how to adjust resistor values to make such a change. Should one halve the dropping resistor value in the B+ and put a resistor of equal value (half) and wattage in the ground leg that's highlighted? I'd think about a choke, but this amp has current such that it would probably be a fairly large one.

Correct.

Do the dropping resistors with halved value still need to have the same wattage rating as the single dropping resistor -- 10 W in this case -- or may the wattage be halved too? That's actually a series circuit, right?, not a parallel one?

 

[2L man]

And another one, for confirmation:



Do the dropping resistors with halved value still need to have the same wattage rating as the single dropping resistor -- 10 W in this case -- or may the wattage be halved too? That's actually a series circuit, right?, not a parallel one?

Half the power rating is correct when half resistance series resistors are doubled.

Parallelling half power as well but there resistance is doubled.

Pondering grounding (return current wires) nodes looking coupling capacitors as a divider is an easy way to approach this issue. Often also transformers divide the stages.

 

[chas.wahl]

To use a graphic example:

The above is correct, in terms of operation and power-handling capability?

My apologies if I'm being dense.

 

[King Fan]

If you're being dense, I'm being depleted uranium. Lead on, I'm learning.