Bias Shifter from Patent US4995084A, Inventor Eric K. Pritchard

[Vivek]

Bias Shifter
from

Patent US4995084A
Semiconductor emulation of tube amplifiers
Inventor Eric K. Pritchard

Feb 19, 1991

https://patents.google.com/patent/US4995084A/en




" At low input signal levels the operational amplifier 136 works with resistors 130 and 131 only. But when the output exceeds the voltage drop of diode 135, then capacitor 133 charges up through resistor 134 and the diode 135. Resistor 132 then puts the resulting off-set on the input and consequently produces a level dependant offset that will work with a symmetrical clipper to produce both even and odd harmonics."


I chose some values of components to get desired input impedance, gain, ratio of DC shift, time constants :


I see output varies with time as follows :



If we zoom in a bit, we can see the time constant. Normally something like 50 to 200ms is OK.



Zooming into the first 5 ms show symmetrical output



But later on, we get Asymmetrical clipping due to the DC bias shift:


Compare this with the Levin Amp (Bias shifter and Distorter in one, mentioned in the Pritchard patent):
https://www.aronnelson.com/gallery/main.php/v/slackers-stuff/album170/levinamp.jpg.html?g2_imageViewsIndex=1

[EBK]

This isn't too far off from the SLM/Crate patent 5,032,796 method.
I built a circuit out of it that had the following simulated characteristics (multiple traces show different gain settings):

First, the initial response:


Next, the response after bias has fully settled:

[Vivek]

Yes that SLM patent schematic is also implied in the other patent.

I too had simulated the SLM patent as wel !!!

I did not build a circuit yet.

I will need to add pre and post frequency shapers to the Bias Shifter / Bias shifter + Distorter to make it sound Amp like.


Did I understand it correct: One does Symmetric clipping first and then leading to Asymmetric

Other does Asymmetric clipping first, leading to Symmetric

[bowanderror]

This is a cool find! It looks like this was the guy behind Pritchard Amps, makers of sophisticated, but not always universally-loved solid state amps. I believe this patent was one of many that went into the somewhat controversial 1990s PRS Harmonic Generator amps (like the HG-70).

It seem the basic concept is the use of resistor + diode (or transistor) to emulate the three halves power law of a triode. The Harmonic Generator amps had one stage for generating even harmonics and another for generating primarily odd harmonics. These stages interacted with a lot of bandwidth-limiting circuitry as well, so you're probably right to give that a go.

[EBK]

There is another simple way to emulate bias shifting, which I think was done in some Peavey amps, by adding a diode to the gate terminal of FET stages.  The details are fleshed out in what I call the Fetzer2 paper (it's a follow up paper to the one ROG got their Fetzer valve idea from).  (scroll down on this page): https://www.researchgate.net/publication/290534843_Triode_emulator_-_Part_2

[Vivek]

Also this : http://www.sugardas.lt/~igoramps/article68/article.htm

[Steben]

The transtube preamp uses BJT stages with diodes

[Vivek]

Figure 5 of the Pritchard patent shows a circuit the begins as symmetrical clipping but progresses to asymmetrical clipping when the bias shifts


The SLM Crate patent shows a circuit that does it the other way around: initially asymmetrical clipping leads to symmetrical with bias shift


What's the difference ?

Which will be more pleasing ?