Tube Amp Forum: The Ultimate Tone

Full Version: zener diode bounding circuits
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I'd like to know more about the variations of the zener diode bounding circuits that are peppered throughout TUT and seen in some popular modded-Marshall preamp designs (Jose/Cameron/Friedman/Fortin). Here are some specific questions for starters:

1. Sometimes there is a 10k resistor (and large-value coupling cap) off of the cathode follower. Why is this used, and what are the drawbacks of adding it?
2. Some incarnations of the circuit replace the usual large-value coupling cap with a 100n cap as part of the shunt circuit (that is, in series with the compliance resistor and zeners). I get that this cap would keep DC off of the zeners, but I'm not sure how the value of the cap would be chosen. 
3. Some of the modded-Marshall designs listed above put the Master volume immediately after the clipping circuit and before the tone stack. Others put the Master after the tone stack but put a 1M resistor in parallel with the zeners (or in this specific case, transistors). What is the rationale behind this load?

I could draw out some schematics tomorrow if need be.
Hi jmcd

The resistor in series with the diodes is for voltage compliance, making the clipping softer. You can zero that resistor and have a harder distortion and either way is quite useful for different kinds of music.

If the bounding network is placed at the output of a follower, as in a Marshall, a DC blocking cap is an absolute necessity. The cathode sits at 1-200Vdc and were you to add a pair of 10V zeners in anti-series, the cathode would be pulled down to about 11V. Remember that while one diode "zeners" the other is forward biased and acts like a standard diode, otherwise the series connection would not work.

Placing the cap in series with the bounding network DC-isolates the network but retains the connection between the follower and the EQ. This is purely aesthetic and is probably the way to go in a modification. As a new build, the coupling cap between the follower and EQ allows the builder to use lower-voltage caps in the EQ. In this scenario it is a good idea to add a leak resistor to ground on the EQ side of the new cap. This assures a charge path for the cap regardless of how the bounding network is switched (if it is).

You can use BJTs as low-noise zeners, but the clipping is then at about 7V - no freedom to change this.

Placing the MV after the bounding network is logical because there must be significant signal for the diodes to clip it. The EQ works regardless of its position, although being fed directly from the MV is not a great choice as the source impedance into the EQ is variable. It is one thing for it to be dynamic, as with plate-drive, but off the MV it goes from zero to half the MV pot value, to whatever the impedance is driving the pot. The tone versus MV sweep could be very changeable and unpredictable.
Lots of good info, Kevin. 

One clarification to my original post: in addition to the compliance resistor, I've seen a resistor placed in between the cathode follower and the bounding network (see the R with the asterisk below). The value of this resistor is often 10k. I presume this is serving a different function and that it might also reduce the signal somewhat, depending on the values of it and the compliance resistor, slope resistor, etc.

[Image: KaStH19l.jpg]
Hi jmcd

Yes, the series R has three functions.

If you leave out the bounding circuit, the resistor increases the effective source impedance driving the EQ and is sometimes called a "build-out resistor". This makes the drive of the EQ a bit less stiff and is a method unto itself of trying to give the tone-crippling CF more of a dynamic quality like a plate-driven EQ would provide. It allows a tech or hobbyist to have better tone without taking the plunge of removing an iconic circuit portion that "certainly" must still provide some voodoo?

With the bounding circuit in place, there will be a voltage division and thus an attenuation, working against the diodes and the compliance resistor.

If the bounding network is switchable, then each of the above operating modes work in the alternate switch positions.