Why does the small clone lfo modulate the cd4047?

[Chico]

OK folks, I am confused on this one.

In the small clone schematic, the MN3007 is clocked by a CD4047.  The LFO modulates the CD4047 through a 10uf cap and 39k resistor.  There is a 150pf cap between pins 1 and 3  of the 4047, and a diode between pins 2 and 3.

When I look at the data sheet of the CD4047, pin 1 is labeled C, pin 2 is labeled R and pin 3 is labeled RC common.  In the example in the data sheet, a capacitor sits between pins 1 and 3 - same as the small clone GOOD.

But the data sheet shows a resistor between pins 2 and 3.  Here is where I am confused.  I assume that the diode simply keeps the lfo voltage from entering the chip.  Since the LFO does not generate any negative voltages with respect to 0 ground, all the lfo output will be between 0 and +9.  So the diode should block all of the lfo voltage (less the diode drop)

Is this correct?

OK, the signal fed into the 10uf cap and 39k resistor is a modulated voltage output from the lfo.  How is this turned into a modulated resistance seen by the 4047?

Any explanations would be greatly appreciated.

I am sort of trying to disect this circuit to make sure I understand it top to bottom, and this is my first of many hurdles that I am currently stuck on.

Best regards

[R.G.]

The diode is being used as a variable resistor to vary the speed of the 4047 outputs.

The forward resistance of a diode varies from infinity (the diode doesn't conduct at all) to a few ohms as you increase the current through it. It's a heavily nonlinear resistance, so you can only use it for very small audio signals to minimize distortion; this is not the case where it changes a high frequency clock though.

The LFO drive changes the current through the diode, and thereby its resistance. The change in resistance changes the speed the 4047 oscillates at, and therefore changes the delay timeof the chorus.

Why? Chorus done with a static clock sounds fairly lifeless and artificial. Sweeping the delay time by  changing the clock rate a little makes it sound more natural.

[Chico]

RG:

Thanks for the reply.  That makes a lot of sense.  And it triggers some other interesting questions/observations.

From what I have read, it is typically desirable to modulate a chorus with a sine wave.  I have also read that a flanger should be modulated with a triangle wave (and there are some that highly recommend hypertriangular wave) modulation.

OK.  So I am in the process of modding a small clone into a combo chorus flanger.  More for the mental exercise and learning experience than to create a great flanger.

So, by replacing my 39k resistor with a 100k pot, I can vary my clock to generate nice long delays for chorus friendly sounds, and shorter delays for flanger approved sounds.  I also added a switch and second cap to double the frequency of the lfo to slow the lfo down for even more flanger approved vibe.  These suggestions stemmed from comments by Mark Hammer, and they appear to be working really well.  (I still have a few bugs to work out with the feedback control, but that is another story)

Now, if the diode controls the resistance, and the resistance (along with the 150pf cap) controls the modulation, and the resistance of the diode varies nonlinearly with current, I expect that the modulation actually produced by the CD4047 is not an accurate representation of the waveform I observe on my scope looking at the lfo output.  (When I look at the clock signal on the scope, I see it modulating, but cannot tell what shape of modulation it is)

So, is there any tricks that can be played to take advantage of the nonlinearity of the diode to get more interesting modulation effects?

Thanks for your help.

and I am putting in a triangle wave into the dio

[Boofhead]

Is this correct?

I don't think so.  Don't agree with RG on this one either.

The underlying VCO is a two inverter oscillator, the VCO version is fairly widely known - the internal ckt of the 4047 is a little more complex but achieves thr same goal.  The diode on pin 2 rapidly charges the timing cap.  The LFO voltage provide the discharge path, hence only the LFO voltage determines the oscillation speed.  In the 4047's "normal" connection the charge and discharge is done by the external timing resistor R.  The rapid charge (or in some case discharge) is common on simple VCOs as it provides timing which is more proportional to the LFO voltage.

[Chico]

Boofhead:

Thanks for the insight.  Please help me to understand how this works.  As I understand your comments,in a "typical" application per the data sheet, a fixed R and fixed C derive a "fixed time constant", thus the 4047 clocks at a frequecy that is a function of the RC values.  However, by replacing the R with a diode,  I can think of pins 1-3 of the 4047 to behave like a VCO.  The diode acts to quickly charge up the capacitor and the lfo serves as a discharge path for the cap, thus the modulation of the of the clock frequency will "follow" generally, the modulation of the lfo.  



If that understanding is correct, I have (hoefully) one last question on this topic.  

If this is the case, what is the "base resistance" seen by the 4047.  What I mean is, there must be some R that established the "nominal" frequency about which the 4047 clock operates with respect to the C.  Is that nominal resistance set by the 39k?

That is, does the 39k serves as a discharge path for the C and thus determine the "nominal" clock fequency with respect to C?

Thanks everyone for your help.

[Boofhead]

If that understanding is correct, I have (hoefully) one last question on this topic.

You're right on!

Is that nominal resistance set by the 39k?

Yes.  The easiest way to look at this is all the action around the VCO is at relatively high frequencies.  The 10uF is pretty much a dead short at these frequencies so everything before it has all but no effect.

That is, does the 39k serves as a discharge path for the C and thus determine the "nominal" clock fequency with respect to C?

Yes, I think you have got a handle on this now.   One extra piece of info, the LFO input to the VCO only works with voltages below about half the  supply rail.  Voltage above this cannot disharge the cap below the CMOS switching threshold and the oscillator stops oscillating.

PS: thanks for the re-iterations and direct questions.

[Chico]

Boofhead:

Thanks for your patience with me.

That is a great tip.  I actually confirmed this last night.  I have been trying to get a more dramatic effect out of the modulation.  So, I breadboarded a dc coupled amplifier to boost the gain of the lfo, twiddled with the bias and applied it to the 4047.

Exactly like you said, above about 4-5 volts, the circuit burped and the modulation effects cut out until the lfo swept below 4-5 volts.


So, my next step was to disconnect the lfo from the 4047 and used it to drive an optoisolator.  I replaced the 39k between the lfo and 4047 with a 100k and put the resistor leads of the optoisoloator across the 100k.   After twiddling with the amount of current delivered to the LED portion of the optoisolator, I could adjust the depth of the modulation to a much greater degree (everything from subtle to extreme).

Funny, it was a great excercise, and I learned a lot (Thanks for pointing me in the right direction)

However, at the end of the day, with the range of control that I now have, I found that the settings I liked the best were within the range of the setup already provided by the original circuit.  

Thanks again.