switching between three different cap values

[TimS]

I'm thinking about using three-position switches for the treble and mid caps to allow for three different values of each (something along the lines of 250pf, 330pf and 500pf for the treble cap and 47n, 22n, and 10n for the mid cap). What's the best method for doing so? And is it worth having three possible values for each, or does optimizing the amp for one or two settings make the others useless?

[dogears]

Hi Tim,

I'd say two positions would be 270pf or 330pf and .01uf and .047uf on the mids....

If you add three, then I'd add in the .022uf mid and something like 470pf on the treb cap.

I would not use 250pf on the treb cap.

Many ways to do this.... You could do series or parallel caps. Lots of ways to skin this cat.

The one absolute certainty is to use a 250ka mid pot. The .01 midcap needs the bigger pot. It has no effect on the .047 fwiw. That is why it is a certainty.

[TimS]

Many ways to do this.... You could do series or parallel caps. Lots of ways to skin this cat.

Yeah, that's where I get confused. I know how the switches work and how to calculate the capacitance of two caps in series or parallel, but I can't decide which of the standard values to use, where to mount the components (on the board or on the switch), and the best way to wire them.

How are you doing it?

[kleinm]

Many ways to do this.... You could do series or parallel caps. Lots of ways to skin this cat.

Yeah, that's where I get confused. I know how the switches work and how to calculate the capacitance of two caps in series or parallel, but I can't decide which of the standard values to use, where to mount the components (on the board or on the switch), and the best way to wire them.

How are you doing it?

I'd use a DPDT on-off-on switch. Connect each pole to one side of a single board-mounted midrange cap. In this case, I'd use .01uF on the board. Your DPDT's poles will connect to each side of this cap.

On one side's throws, connect a .01uF cap across it. This will get you .02uF by paralleling the .01uF on the board and the .01uF on the switch.

On the other side's throws, connect a .04uF cap across it. This will get you .05uF by paralleling the .01uF on the board and the .04uF on the switch.

When the switch is in the middle, the poles will be broken and your signal will go only through the .01uF cap on the board. This is my only uncertainty. In a DPDT on-off-on, I think the poles are not connected in "off". Not sure, though.

[TimS]

I'd use a DPDT on-off-on switch. Connect each pole to one side of a single board-mounted midrange cap. In this case, I'd use .01uF on the board. Your DPDT's poles will connect to each side of this cap.

On one side's throws, connect a .01uF cap across it. This will get you .02uF by paralleling the .01uF on the board and the .01uF on the switch.

On the other side's throws, connect a .04uF cap across it. This will get you .05uF by paralleling the .01uF on the board and the .04uF on the switch.

When the switch is in the middle, the poles will be broken and your signal will go only through the .01uF cap on the board. This is my only uncertainty. In a DPDT on-off-on, I think the poles are not connected in "off". Not sure, though.

Thanks, that sounds a lot better than what I had been thinking of doing.

[heisthl]

One way to do it: Use an on-off-on SP switch put the feedpoint and the smallest value on the center terminal and from the other end of the smallest value connect the destination and the 2 larger values which also connect to their respective switch poles. this lets the switch parallel caps when toggled. The only drawback to this method is the switch panel label is not progressive.

[TimS]

Another dumb question:

Wouldn't switching tone stack caps in parallel mean that there would be quite a large amount of DC voltage on the switch? If so, isn't that potentially hazardous?

[mlp-mx6]

Not if you're switching the OUTPUT side of the capacitors. No DC at all.

[novosibir]

Don't switch the caps on the side, facing to the tube's plate! Switch it on 'the other side', facing to the pots. There's no DC already.

But if you've done it wrong by accident, you anyway are switching on the toggle - and aren't touching the solder lugs inside - and on the toggle there isn't any voltage at all, and moreover it's grounded through the housing to the amp's chassis.

So not categorial necessary, to conclude an additional life policy

Larry

[TimS]

Not if you're switching the OUTPUT side of the capacitors. No DC at all.

How do you switch in parallel on the output side only? Doesn't switching in parallel require leads from both the input and output sides?

[mlp-mx6]

No. To add the cap in parallel they can be permanently connected to the "hot" side (where the DC is) and the 2nd one is connected on the "cold" side to the other by the switch.

Make sense?

[TimS]

I see what you mean - instead of having one cap on the board and leads coming from both sides of that cap to the switch where the other two caps are, have all three caps on the board, connected together at the hot side, and have the leads coming from the other side of each cap to the switch.

Thanks for clearing that up for me.

[mlp-mx6]

Correct.

Another way, depending on how many combinations you want, is to have the "bigger" cap on the board, and have it run through a smaller cap in series. The switch would short out the smaller cap, leaving just the larger one in effect.

[TimS]

I'VE GOT IT!

I finally figured out how to switch between (for example) 10n, 20n, and 50n mid cap values AND have the values increment from one switch position to the next:

Using a DPDT on-on-on switch (the kind that in the middle position connects one center terminal to a bottom terminal and the other center terminal to a top terminal), wire the two bottom terminals together. One bottom terminal (the one that connects to the center terminal in the middle switch position) is connected to the lead coming from the output side of a 10n cap (all three caps are on the board and connected on the hot side). The other bottom terminal goes to the mid pot. The center terminal that connects to the bottom in the middle switch position is connected to another 10n cap, while the other center terminal is connected to a 30n cap.

When the switch is in one extreme, both center terminals are connected to their respective upper terminals. This leaves only the 10n cap on the bottom connected to the output, so the mid cap value is 10n. When the switch is in the middle, the other 10n is connected to the output giving 10n//10n = 20n. And in the other extreme, both center terminals are connected to the output, so 10n//10n//30n = 50n.

Does that sound right?

[dogears]

Why not just connect the .01uf to the midpot.

Then wire up a .01uf and .04uf to a center off dpdt switch. The center terminal also goes to the pot where the original .01uf is. If the switch is in the center, it is just the original .01uf. Up and down yield .02 and .05.

All caps are connected together at the hot, the output sides of the additional .01uf and .04uf go to the switch and woudl be in parallel with the original .01uf.