Diode Strings as Voltage Controlled Resistance (VCF)

[ve3wwg]

I've added some wiki information that might be of interest to those interested in how the diode strings work, in voltage controlled circuits like the VCF.

http://www.experimentalistsanonymous.com/ve3wwg/doku.php?id=diode_ladders

I've been analyzing the Steiner VCF circuit, in preparation for a new synth module build. I took some time to play with the diode string idea in LTspice to sharpen my own knowledge on the subject. The link above explores this concept and plots some low pass filter responses.

[expanoncolin]

This is cool.  I hadn't seen it explained before how the diodes are acting like VCRs in these circuits.  Do you know how Ge vs. Si vs. LED's curves are different?  It seems like it would be nice if maybe Ge diodes had a sharper exponential conduction curve, so that you could reach higher cutoff frequencies while applying less voltage... allowing a larger dynamic range for your signal.

[ve3wwg]

I don't believe there is very much difference between Ge and Si diodes, apart from the voltage where they turn on.  They both operate on the same exponential conduction law. The first part of the following wiki link describes this fairly well:

  http://en.wikipedia.org/wiki/Diode_modelling

I've attached an image of the essential part here.

The formula is the same for both. The difference is that the parameter Is varies considerably between them. All this difference does is shift the point at which conduction occurs (0.2 volts for germanium vs 0.7 for silicon as you know). Notice that the value for germanium's Is is 3 orders of magnitude higher than silicon's!

I should try some LTspice experiments on a Germanium type to demonstrate that. What I'd expect to find is that the V_DC1 voltages for the graphs will just shift downward. In other words, instead of no change from 0 to 1 volts (of V_DC1) as it did for silicon, I would expect that changes in frequency response start to occur at some point below 1 volt in V_DC1. This would simply be the result of germanium diodes conducting sooner.

Conversely, if you were to use LEDs instead, the dead zone would be considerably higher than 1 volt for V_DC1 (due to their much higher turn-on voltages).

So in this particular application, I believe there is little to be gained from using different diodes. It will just move your sweet spot for the control signal (which is adjusted out by calibration).

[expanoncolin]

Interesting.  It's too bad that you can't alter the exponential conduction curve by using devices with different characteristics.  Given that, I agree with your intuition that the curves would just be "shifted".

-Colin

[ve3wwg]

I think anyone interested in analog audio processing techniques is always on the lookout for some physical property to exploit.

One of the things on my "to investigate list" is the tunnel diode (or any other device/configuration that exhibits "negative resistance"). The tunnel diode has a rather weird conduction curve:

  http://en.wikipedia.org/wiki/Tunnel_diode

I can't help think that a signal that moves through the region of "negative resistance" to the other side and back, would create a number of nice harmonics. The last time I checked ebay they were scarce and pricey. Checking just now, I see that there are a lot of them available from Russia or the Ukraine. I once bought a Geiger counter from the Ukraine, that works well.  It's very tempting, if I didn't have so many other things to build right now.

[expanoncolin]

One of the things on my "to investigate list" is the tunnel diode (or any other device/configuration that exhibits "negative resistance"). The tunnel diode has a rather weird conduction curve:

  http://en.wikipedia.org/wiki/Tunnel_diode

I can't help think that a signal that moves through the region of "negative resistance" to the other side and back, would create a number of nice harmonics. The last time I checked ebay they were scarce and pricey. Checking just now, I see that there are a lot of them available from Russia or the Ukraine. I once bought a Geiger counter from the Ukraine, that works well.  It's very tempting, if I didn't have so many other things to build right now.

Very peculiar!  I agree that that may create some very interesting distortion.  I played with negative resistances a long long time ago (maybe 2003?) when I didn't know what I was doing (I still probably don't), based on this design: http://moosapotamus.net/THINGS/tri-negistor.htm and some other stuff...

-Colin

[ve3wwg]

Very interesting!  Wonderful to see you don't need a tunnel diode to check this out.

In http://www.keelynet.com/zpe/negistor.htm

shows that a 2N2222 and many other NPN devices can act with negative resistance. The link above shows how to test for that (a little hard to read).

I'll be experimenting with this to be sure.  I have lots of NPN around here to try it on.

I also found it interesting how they explain that you can form an SCR from a pair of transistors.  I've seen references to this in semiconductor lit, but always thought of it as theory rather than something that actually works.  There have been times when I wished I had a low power SCR. But SCRs aren't something I normally stock.

[expanoncolin]

http://www.keelynet.com/zpe/negistor.htm

That article is exactly the other one I was looking at... I actually think I still have a printed version of it in my stack of papers.  Looking forward to seeing what you come up with.

-Colin

[ve3wwg]

I'm thinking I should bring out my homemade diode tracer circuit to see if we can demonstrate that on the scope somehow.  

I need to get some other stuff out of the way first but this looks like fun, to be sure.

Update: I found a better to read version of that PE article here:

http://www.schematicsforfree.com/files/Components/Circuits/Negistor%20Explained%20-%20The%20Mysterious%20Negistor.pdf?action=download

or http://tinyurl.com/87uaads

The other copy was very difficult to read.