Panning for fun, can this be done?

[Scruffie]

I've been staring at this circuit snippet for about an hour trying to work out a way of getting my intended response. I want to pan one signal to two different places as in the circuit in the top right of the first page here http://www.geofex.com/Article_Folders/panner.pdf

However, instead of a constant blend, I want the pot to be silent in the center position, as illustrated in this diagram.



Can it be done with a single gang pot or do I need to break out a dual gang?

[Mark Hammer]

Just so we're clear, you want to do this manually or via foot control, not via an LFO or other control signal?

[Scruffie]

Sorry, yes, one manually adjustable pot blending an input signal between 2 points with a center dead spot.

[thermionix]

I can think of possibilities with a center-tapped pot (ct grounded), but it depends on the rest of the circuit, as usual.

[armdnrdy]

I can think of possibilities with a center-tapped pot (ct grounded), but it depends on the rest of the circuit, as usual.

Yep...that was what came to mind as well.
Scruffie...take a look at some audio panning circuits. Punch in center tapped pot panning...see if it "pans" out. 

Take a look at the Multiflanger drawing. That circuit uses a dual center tapped pot at the output.
You might get some ideas.
You can find CT pots on Ebay and Radio shack actually sells a dual 100KA CT pot.
I believe they call it a loudness tap.
https://www.radioshack.com/products/radioshack-100k-dual-ganged-stereo-volume-control?variant=5717293381

[PRR]

As shown, this is 25:1 down in center, 50:1 down at extreme; (nearly) full-up at the other extreme.

Not "silent", but awful quiet.

The drop as you come off the stop will be very steep.

The 500K and 10K are somewhat arbitrary. A smaller ratio may be smoother off the stop but less-silent middle and dead-end. A larger ratio (1Meg+1K??) may get the center down so much the bleed is negligible, but will have trouble with end-resistance on typical pots.

It seems to me I "did" such a thing, here or elsewhere, using cancellation. But it isn't coming to me tonight.

[duck_arse]

I have a balance pot pulled from a hifi amp. it confuses me every time it comes to hand. it is a single-decked wafer, and measures 200k end to end. but from one end to wiper, it goes from 0-100k at 12:00, and is still 100k to full rotation. from the other end, it measures 0-100k at 12:00, and 100k at the other full rotation.

it has a centre dentent, which is nice.

[PRR]

500K against 10K may be too abrupt coming off the ends; try 250K+10K.

There is a more-clever but more complex way (and I am SURE I derived this before!). Take a bit of the input and subtract it from the output. This can give infinite loss at the end, but not infinite in center.

"Infinite" except the "bit subtracted" depends on potentiometer resistance, and pots are very loose tolerance. And in a real room, if one speaker is playing full (0dB), another speaker playing well below -20dB will not be noticed (even under Haas). So "silent" only has to be 20+dB down. And the "bit subtracted" divider does have to be adjusted for the specific pot in your build. (If you accept a bit of whisper, you can probably find a divider which "works well enuff" for +/-30% pot tolerance without a trim.)



Nearly 0dB at one end, -19dB at the 1/4 point, -27dB at center, and around -40dB at the end.

The "10K" should "match" better than 10%, but 10% parts from the same tape probably will; 5% parts are now common.

The "~7K" part trims the "silence at end" null. If you under-correct (short) there will be a bleed at -28dB. If you over-correct (>9K) the "quiet" half of rotation will null to zero and then come back up. This marginally improves the center loss, but of course makes the off-end drop more abrupt, and bleeds more at the silent end. Values of 10K-20K get you in this zone. 10K may indeed be a happy (and handy) compromise.

Input impedance varies from ~50K to ~10K. Don't drive this with a guitar, or a hi-Z pedal. A buffer would be a wise frill.

Note that this thing inverts. If it matters, add another inversion.

A high-spec version then wants four opamps. Buffer, low-Z (for hiss) inverter, and the nulling amps.

[thermionix]

I have a balance pot pulled from a hifi amp. it confuses me every time it comes to hand. it is a single-decked wafer, and measures 200k end to end. but from one end to wiper, it goes from 0-100k at 12:00, and is still 100k to full rotation. from the other end, it measures 0-100k at 12:00, and 100k at the other full rotation.

it has a centre dentent, which is nice.

Whaaaa???  That don't make no kinda sense!

[Fender3D]

If you're going to use this for phase/out-of-phase feedback in flanger or phaser, you can play with phase gain as Paul suggested.
Check the schematic on my post about TC Phaser on the other forum or PM me...

[Scruffie]

If you're going to use this for phase/out-of-phase feedback in flanger or phaser, you can play with phase gain as Paul suggested.
Check the schematic on my post about TC Phaser on the other forum or PM me...

Nope, actually I was thinking of being able to use a single pot to work as an input gain control for 2 separate circuits, i.e CW you'd have a tube screamer and CCW a whatever but I didn't want them to bleed in to each other too much.

500K against 10K may be too abrupt coming off the ends; try 250K+10K.

There is a more-clever but more complex way (and I am SURE I derived this before!). Take a bit of the input and subtract it from the output. This can give infinite loss at the end, but not infinite in center.

"Infinite" except the "bit subtracted" depends on potentiometer resistance, and pots are very loose tolerance. And in a real room, if one speaker is playing full (0dB), another speaker playing well below -20dB will not be noticed (even under Haas). So "silent" only has to be 20+dB down. And the "bit subtracted" divider does have to be adjusted for the specific pot in your build. (If you accept a bit of whisper, you can probably find a divider which "works well enuff" for +/-30% pot tolerance without a trim.)



Nearly 0dB at one end, -19dB at the 1/4 point, -27dB at center, and around -40dB at the end.

The "10K" should "match" better than 10%, but 10% parts from the same tape probably will; 5% parts are now common.

The "~7K" part trims the "silence at end" null. If you under-correct (short) there will be a bleed at -28dB. If you over-correct (>9K) the "quiet" half of rotation will null to zero and then come back up. This marginally improves the center loss, but of course makes the off-end drop more abrupt, and bleeds more at the silent end. Values of 10K-20K get you in this zone. 10K may indeed be a happy (and handy) compromise.

Input impedance varies from ~50K to ~10K. Don't drive this with a guitar, or a hi-Z pedal. A buffer would be a wise frill.

Note that this thing inverts. If it matters, add another inversion.

A high-spec version then wants four opamps. Buffer, low-Z (for hiss) inverter, and the nulling amps.

Thanks Paul, that looks perfect for my purposes, I was already going to have a buffer up front and then an inverting mixer from the two sources at the end so that should all work out nicely.