Fetzog

[Eb7+9]

I'll make this quick, was up all nite - battery light is blinking ... arrfff !

This is what I came up with, it's based on the Garnet Herzog which in turn was loosely derived from the practice of Boogie'ing a Fender Champ into a Fender Twin ... see my pedal page for the AmpLineOut circuit and the accompanying explanation ...

http://www.lynx.net/~jc/pedals.html

My idea consists of producing a signal current large enough to drive a small single-ended coil into saturation - done by paralleling several Bipolar devices into a super-transistor (collector currents simply add through Kirchoff) ...

http://www.lynx.net/~jc/Fetzog.gif

I included a Darlington device to supply multiple Base-current feeds but this could be omitted for the sake of inducing clipping of the previous stage - although I'm not sure if this will even do anything considering the linear nature of base current transfer (at least compared to vacuum tube grid current specs) ...

The main mechanism that is desired here is a magnetic saturation of the output transformer ... this can be achieved by also reducing the amount of core in the transformer - some small audio transformers can be easy to have core links cut through using the Carborundum wheel on a Dremmel ... by progressively (and carefully) cutting through the core one would "mainly" alter the maximum Flux limits whithin the transformer ... FYI: Garnet Amplifier designer Gar Gillies did this in some of his high-powered amps like the well-know SessionMan amp ... so in this way it might be possible to extend the transient emulation of tube circuits (at the very least) into the realm of the Champ and Herzog ...

I know it sounds a little out there, maybe I'm just dreaming ... heh heh ...

... zzzzzzZZzZZzZZzZZZZzZZzzZzzzzz ...
jc

[Peter Snowberg]

Very cool!

Take care,
-Peter

[R.G.]

My idea consists of producing a signal current large enough to drive a small single-ended coil into saturation - done by paralleling several Bipolar devices into a super-transistor (collector currents simply add through Kirchoff) ...

Magnetic cores saturate as a result of volt-seconds. To drive a small single ended core into saturation in one direction at least, all you have to do is to introduce some DC through it.  If you introduce no other limiters, the current ramps up until the only limit on the current is the DC resistance of the wire. In fact, at DC, the *only* limit on inductor (and hence, transformer primary current) is the winding resistance and external current limiters.

Ungapped cores saturate at lower volt-seconds than do gapped cores. The air gap is essentially un-saturable, but has a very low inductance. By cutting some of the laminations, you in effect make the primary inductance nonlinear - and smaller than it was, reducing low frequency response.

Here's how this works:
- an ungapped core (a strip-wound toroid is the best example) will saturate when the volt-seconds produce the max flux density the iron can carry. (Simplifying a bit...) up to that point, the inductance of a coil on that core is proportional to the turns squared times the permeability of the iron, typically several thousand. Big inductance per turn. When the core saturates the permeability of the iron reduces to unity, same as air, and the incremental inductance drops to that of the coil in air.
- a gapped core of the same size has a core constant (inductance per turn squared) determined essentially by the air gap, since the iron's permeability is so high that all the MMF is dropped across the gap. A same-sized core with a gap will have a much lower inductance determined essentially by the air gap. The inductance will be constant until you can induce in the air gap enough flux to saturate the iron in the rest of the path, and then the iron permeability goes to one, and the inductance drops to the air-cored value. But this takes typically ten to a hundred times more volt-seconds to get to, since the air gaps are what must be pushed to a flux that lets the iron "short-circuit" saturate.

This second set of reasons is why single ended output transformers are (a) gapped - to keep from saturating from the necessarily DC current through them and
(b) big! - to keep enough inductance for good low frequency response when the iron is not giving you the huge increase in inductance it would in an ungapped core.

The other problem you'll find is that inductors saturate on volt-seconds. That is, they saturate easiest at low frequencies, since you have to make up in volts what you don't have in seconds as the frequency of the signal goes up. It's almost impossible to saturate an audio output transformer that's otherwise well designed at any kind of treble frequency because of this.

What you *might* do is to leak some DC through the transformer, in fact it looks like you're doing this. Saturate or near saturate it to one side with the DC, and then any signal will be pushing the core into saturation one way and out of saturation the other. This will effectively clip one side of the signal at the transformer output quite softly and give you an OK sounding partially rectified-sounding signal at the output. This might - and probably will, see below - give an OK sounding thing, but it's not primarily a core saturation of an output transformer effect, it's half wave rectification by another means.

It's worth experimenting with, but you do need understand what you're changing to pick the next place to change.

Nota Bene: I've been experimenting for some time now with using *two* transformers for soft clipping by offsetting one core with DC in one direction, the second in the other direction. The clipping is quite soft, and appealing. It's probably too soft for metalheads, but I like it. I expect to have a pedal design posted soon based on that.