Stereo Tremelo Panner Idea

[zpyder]

Hey everyone.  I've got an idea for a circuit that would pan my signal between two amps cyclically.  I currently use the "Improved Hum Free A/B/Y" by R.G. (pics: http://aronnelson.com/gallery/zpyders-build-pics, layout: http://aronnelson.com/gallery/humfreeaby)to split my signal to two amps.  I have the idea for a circuit which is essentially two tremolos, one for each amp, that are always at states opposite each other.  This should create the effect of my signal "panning" from one amp to the other.  My thought is to make an LFO using a CD40106 or something that drives two opamps oppositely.  Here's my hand-drawn note/schematic:



The idea is an LFO drives two opamps, one via + input, the other at - input, so that the opamps' outputs are opposite eachother.  Then these opamps both drive an LED/LDR pair which are essentially the "gain" resistors of two other opamps which control the attenuation of the two signals.  I would really appreciate a more expert opinion on this circuit and whether or not the idea can work.  The drawing is missing coupling/decoupling caps for in & out as well as pulldown resistors.

thanks for taking a look!

zpyder

[caress]

you should check out the 4ms stereo panneur.  it's very similar to what you're interested in building and might give you some more ideas.  it's basically 2 tremulus lunes that can sync up or not and also pan or not...very cool

[zpyder]

got a link to a schem?

zpyder

[gez]

If you use a CMOS square wave LFO it would be better to use one of the unused devices to invert the output signal and wire up the lower op-amp buffer as a follower connected to said inverted output.

The LEDs need stop resistors to stop them frying.

I don't really follow what you've done with the audio part of the circuit.

[gez]

PS  LM358 or something similar would have to be used for the buffers...something that doesn't mind its inputs being pulled to ground.

Now that I think about it, it would just be easier to use some more of the unused 40106 gates as the buffers.

[$uperpuma]

http://commonsound.org/panneur/panneurscheme.gif Here's the link to Dann's schem... I really want to try this one...but the PCB layouts aren't completely ready for transfer... they have grey lines across traces that don't seem to be traces. Any genius that wants to do a layout for this one there would be a lot of people in your debt...

[nosamiam]

Is there really something wrong with that layout?  I'm not seeing the grey lines.

http://commonsound.org/panneur/panneurboard.pdf
http://www.commonsound.org

I built the Phaseur Fleur on PCB (great build, BTW) and I don't recall having to fix anything that was wrong.

[Transmogrifox]

Hey Zpyder- You can do this whole thing with one LDR, I will try to put together a schematic for this.

What you do have will work, only it needs a few things like a resistor to limit current in the LEDs, and a resistor to limit the gain to 1 on the inverting  op amp LED driver.  I would also recommend changing the order of the input resistor and LDR so the LDR is in series with the resistor shunted.  That way you can make the shunt resistor about 10k, while the LDR varies from about 10k to 1Meg, or so.  If you go the opposite direction, the series resistor needs to be about 500k or 1Meg, and the total input resistance to ground is higher.  this equates to a more noisey op amp stage.

[$uperpuma]

Hey Zpyder- You can do this whole thing with one LDR, I will try to put together a schematic for this.

What you do have will work, only it needs a few things like a resistor to limit current in the LEDs, and a resistor to limit the gain to 1 on the inverting  op amp LED driver.  I would also recommend changing the order of the input resistor and LDR so the LDR is in series with the resistor shunted.  That way you can make the shunt resistor about 10k, while the LDR varies from about 10k to 1Meg, or so.  If you go the opposite direction, the series resistor needs to be about 500k or 1Meg, and the total input resistance to ground is higher.  this equates to a more noisey op amp stage.

perfect...thanks!

[Transmogrifox]

Oh, one thing I didn't take a mental note of was that you had two inputs.  If you want the stereo channels to be isolated, you need to do what you're doing with 2 LDR's.  I'll post a schematic with one of each.

I still need to finish drawing it up and scanning it

[zpyder]

FYI-

I got the basis of information for this idea from reading these two articles at Geofex (well, these articles and having 2 amps running at the same time): http://www.geofex.com/Article_Folders/panner.pdf (Panning for Fun), and http://www.geofex.com/Article_Folders/LFOs/psuedorandom.htm (Digital Generation of LFO's for Modulating Effects - covers square, sawtooth, sine, and pseudorandom waves from hex triggers).  Both great articles!  Thanks R.G...

Transmogrifox- I'll be looking forward to see what you draw up.  Thanks for the comments.

cheers,
zpyder

[Transmogrifox]

Here are a couple different schematics:

http://www.geocities.com/transmogrifox/DualPan
http://www.geocities.com/transmogrifox/SinglePan

The "DualPan" is Stereo In, Steeo Out.  The circuit is essentially the same as the "SinglePan", except the single is Mono In, Stero Out.

The SinglePan Leaves the outputs out of phase with eachother, so there will be some cancelling  between the speakers as well, when they pan.  I suspect this may create a somewhat interesting effect, so I didn't invert on of the sides back to compensate.  You could put another inverting op amp after one of the outputs to switch the phase so both sides are in phase, if you like the adding effect between speakers.

I hope that gives you some ideas.

[zpyder]

Transmogrifox-

Taking a look at the dual pan:  First thing I notice is that you orient the two LED's in opposite bias configurations, so that they light up opposite eachother.  I like this, much simpler than my idea to use inverting & non-inverting opamp inputs.  On the top LED (the one oriented to light up when the LFO is negative), I can't read what you have in series with the 2.2K resistor - that's power supply right?

Other questions:
What about isolation of grounds & power supplies??  If we tie all of the grounds together, will we lose isolation?  Do we need to take care to isolate the LFO?  Does star grounding apply to any of this (if there is only 1 connection there is no loop)??
Why do I not see a resistor in the feedback loop of the first opamp?
Does this circuit leave the output in phase with the input?


Thanks again.  I'm still learning some of the basics, so please be patient as some of my questions may seem silly.

zpyder

[Transmogrifox]

First thing I notice is that you orient the two LED's in opposite bias configurations, so that they light up opposite eachother.  I like this, much simpler...

Credit goes to RG Keen for this one.  I saw something like this in one of his articles/threads/schematics at one point, but can't remember where now.

About the first op amp:  This is a unity gain configuration.  The "resistor" in the feedback loop is just a wire.  This creates a high impedance input for the guitar, and is able to drive the following stages regardless of their impedance, so that you don't have this periodically changing input impedance due to the LDR to load down your guitar pickups (although this may create a somewhat interesting effect on its own).

The "star grounding" would apply in principle to isolating the LFO from the audio op amp.  In this case, just take a separate connection from where the voltage negative lead (defined as ground in this circuit) connectst to the board.  It may or may not be important depending on what kind of LFO you use.  If you use a square wave LFO, or a "bang-bang" comparator+integrator type triangle wave LFO, then you'll want to try your best to keep them apart.  Do the same thing with the positive supply.  YOu're concerned about the entire current loop, not  just the ground.

That symbol you see with the top LED just says "+V".  It's the voltage supply positive, assumed to be 9Volts.

One other note:  Depending on your LED & LDR you may want to use different resistor values than 2.2k (larger or smaller depending on the brightness/efficiency of the LED).  I recommend that you have an LDR that is about 10k at full bright for this circuit.

That 100k resistor and 10k trim could be done away with to the consequence of no adverse effects (you would just connect the 2nd op amp's + input to the "Vb" virtual ground).  This feature is added so that the Tremolo can get to infinite depth.  To tune this properly, you would turn the LED to full bright (by tying the LFO connection point to Vcc or Ground depending on which side you were doing.  You then turn the trimpot one way then the other while plucking the guitar on the input until you find the sweet spot where almost no audible sound can be detected.

Finally, the output is inverted from the input.  This can be fixed with  a pair of 10k resistors as in the Single Pan.  You would actually want 2 single pan circuits and use only "Out1" on either.  What this does is it takes the inverted output, adds it to the noninverted input, so that it cancels with itself.  Thus, the overall output is in phase while the out of phase component just changes the amplitude by cancelling to the varying degree that its own amplitude is modified.  I hope that makes sense.

I personally can't think of why it matters.  My preference is having outputs inverted from inputs so that parasitic coupling effects from output to input in the box equate to negative feedback (which is good), so it's less likely to go unstable and generate a noisey squeal in the event you happen to do a sloppy job building the circuit.  Positive feedback more often results in instability.

Some people think you should just build your box better if it's going unstable from a poor assembly job.  That's generally a good idea and I agree in principle, but I have few qualms about a ratsnest circuit if it works just as well as a meticulously bent, routed, clipped and organized board (although that's an art form in itself).  I am less "neat" particularly when I'm trying something experimental like this.

[zpyder]

First thing I notice is that you orient the two LED's in opposite bias configurations, so that they light up opposite eachother.  I like this, much simpler...

Credit goes to RG Keen for this one.  I saw something like this in one of his articles/threads/schematics at one point, but can't remember where now.

I have a very basic 40106-based LFO running on my breadboard right now at ~.1Hz (full cycle once every ~8 seconds).  It uses one schmidt trigger to create the oscillation, and then one more trigger just as a buffer.  The circuit is this:

Only difference is that I placed a .1uF cap across the +Vdd and -Vss pins.  I'm using a 4.7uF cap and a 1M resistor.  The formula f=1/R*C gives us ~.212Hz (once every 4-5 seconds).  Firstly, for whatever reason my oscillator is running at nearly half the frequency predicted by formula for some reason.  This doesn't really bother me since it still works.  I have a 10k resistor in series with an LED going to ground at the output.  I can visually verify that the LED turns on and off every 4 or 5 seconds.

However:

When I place two LED's on one output, with one reverse-biased as you showed in your Dual Pan schem, the reverse-biased one does light up, but VERY VERY dimly.  Much much dimmer than the normally biased one..  What can I do to make them the same luminosity?  What kind of LFO would you use for your Dual Pan?  I've only done a Schmidt trigger LFO.  Perhaps if I ran a bipolar supply or biased the 40106 at half supply!?!?  If I could have the schmidt putting out either 9v or 0v, and have GND at 4.5v for the LEDs?  Just brainstorming... I don't really know what I'm talking about here.

thanks in advance,
zpyder

[Ben N]

Don't know what you're talking about?  Welll, that makes me feel right at home.   
Why not stick a single bipolar transistor configured as a phase splitter/driver ahead of the dual LDRs?
Ben

[puretube]

http://www.diystompboxes.com/smfforum/index.php?topic=29374

[zpyder]

Why not stick a single bipolar transistor configured as a phase splitter/driver ahead of the dual LDRs?
Ben

Could you elaborate on this?  Possibly provide a quick schematic ??

thanks,
zpyder

[Transmogrifox]

http://www.diystompboxes.com/smfforum/index.php?topic=29374

I think I miscredited this idea.  It was this post by puretube that inspired me, not RG.  I know this because this is the very post I was talking about

[Ben N]

Why not stick a single bipolar transistor configured as a phase splitter/driver ahead of the dual LDRs?
Ben

Could you elaborate on this?  Possibly provide a quick schematic ??

thanks,
zpyder

Look at most any discreet octaver, and look for a transistor with equal value resistors on the emitter and collector, driving the rectifier.  Like the second tranny in this:
http://www.fortunecity.com/tinpan/humperdinck/223/diffoct.gif
Ben