I was gifted the above PC board, which is a Fyrall by Ciat-Lonbarde (Peter Blasser). While awaiting the panel for my Dual Benjolin project (previous post), I dug into the documentation for Fyrall. I remembered having looked at it eleven years ago! At that time I gave up on understanding it, because I had no PC board to build it with.

I’ve spent the last three days combobulating myself to the Fyrall, entering once again the mind of Peter B. Finally, I have understood every aspect of the circuit and have started to make a plan for how I want to implement it. I’ll simplify it somewhat by not utilizing all of the nodes. I plan to use color-coded banana jacks for the patching connections.

Going against the spirit?

Peter’s instruments have always been designed for experimentation by non-technical users, who don’t need to know the inner workings. I, on the other hand, being a technician, like to understand his circuit designs. They are especially ingenious at times. By grokking the innards, I can lay plans for customization, as well as being more systematic than Mr. Blasser may have intended. That’s how I feel compelled to approach his work.

Preliminary Findings

You can download the two documents I’m working from, fyrall-papyrion.pdf and Fyrall-Manuel-To-Put-Together.pdf. I’ve included below a few snapshots from those docs to aid in my exposition here.

Short Overview of Fyrall

Peter didn’t really give much of an overview in the two docs mentioned. There is a drawing, but as you may know, his drawings often take some deciphering to make sense of.

Fyrall consists of three main sections, Z, C, and O. Each section contains three duplicates of pairs of circuits. Three times two equals six, which is the overall cardinality of functions. There are two separate-but-related feedback loops inside Fyrall. One loop consists of a cycle of inter-modulating oscillators. The second, much longer loop consists of the oscillators driving staircase generators that also make signals for selecting other signals to go back to modulate the oscillators. I imagine that various substances may have helped Peter with this design.

Section Z – Oscillators

Section Z has six oscillators, three that “stab up”, i.e ramp waves, and three that “stab down”, i.e. saw waves. These come in pairs of up and down stabbing that share a single knob and LED indicator. I like to start with oscillators when describing a complex design, because they originate the driving signals. A pair of oscillators he calls a “King’s Face”.

Now put three King’s Faces together in a spiral.

The six oscillators are modulating each other in a ring. Why does he say they soak orrangers? Because the outputs from section O (below) come back to impact this modulation.

The ramp and saw waves go out to become inputs to the Orranger. Each pair of oscillators also generates two pulses for the section C, which we’ll come to next. But first have a look at Peter’s schematic for a King’s Face.

Section C – “Counten Temples”

Here’s the schematic for section C.

Section C consists of three pairs of chips, a CD4093B binary up-down counter, and a CD4077B quad XNOR logic. As he says in the pic above, the up and down count pulses come from section Z, and in fact they come from the pairs of oscillators in each King’s Face. Thus, the rates of the oscillators sets the speed of up and down counting. A simple resistor ladder network converts the 4-bit output into a staircase wave.

These outputs each go also to an XNOR gate, which inverts them. The XNOR gates second input is grounded, but can be affected by patching into a node. The XNOR logic outputs go through another ladder network to produce a different staircase. Furthermore (!) these outputs are routed to the preset inputs of the next 4093. See, the 4093 can not only count up and down, it can have its inputs forced immediately to the output, by pulling down the PL pin, which is normally held high but is brought out as yet another node, CL, allowing you to mess things up. Peter says, CL is a sensitive node which causes the counten temple things to freak.

Besides the six staircase wave coming out of section C, come six “Select Joint” bits, in three pairs for controlling the switches of section O.

Section O – Orranger

Each one-third of section O consists of a CD4052B.

CD4051B, CD4052B, CD4053B (Rev. G)

The CD4052B is a differential 4-Channel multiplexer having two binary control inputs, A and B, and an inhibit input. The two binary input signals select 1 of 4 pairs of channels to be turned on and connect the analog inputs to the outputs.

Sounds complicated, but the 4052 is just a dual 4-1 switch, controlled by two bits that in this usage select one of four inputs to be routed to the output, which goes back over to section Z to control the inter-modulation of the oscillators.

Here’s a diagram of Section O.

Each switch takes four inputs and there are six switches, making for 24 inputs to be switched (whew). Peter gives the builder a choice from 12 signals to hard-wire into the switches. (Each signal goes into two switch inputs.) His advice is to randomize them, which is a nice, John Cage method of choosing.

Well, I want to know what was being routed where. So I traced it out and drew the following diagram.

On the right, from top to bottom, the following signals are made available. Staircase and XOR signals come from section C. The oscillators are from section Z, in order of their linkage in the spiral. (I picked an oscillator to be #1 arbitrarily. Same for the three part of section C.) I also traced where the O outputs go to, and which Counten Temple the Select Joint bits come from. All this tracing allows me to prepare the U-strip wiring non-randomly for maximal diversity. See my notes on the pic, above.

• Oscillator 3
• Oscillator 6
• Oscillator 1
• Oscillator 4
• Staircase 1
• XOR 1
• Staircase 2
• XOR 2
• Staircase 3
• XOR 3
• Oscillator 5
• Oscillator 2

Now, here’s the schematic for one-third of section O.

Includes notes about my plan for node usage. (Note that the LED driver differs from the published schematic. My schematic is traced from the actual board.)

Let’s talk about the Select Joint bits. These are brought out of section C and used to control the 4-to-1 switches. So, oddly enough, section C not only provides six staircase signals to be switched, but also six control bits to do the switching!

A small, optional modification

Here’s where grokking a circuit can give you ideas. The switch outputs are buffered by op amps, and these points can be brought out as the nodes OZ. I noticed a similarity to the design as used in the Quantussy (center of the Cocoquantus). If you like, you can add a 100nf capacitor to ground at node OZ.

[EDIT] CAVEAT: This probably wouldn’t work, for reasons detailed in my comment, below.

Node ON provides access to the Inhibit input of the 4052B. If this is pulled high, both outputs go high impedance (i.e. not connected to any input). Without the capacitor, unless you patch a signal into OZ, the input to the buffer will float. Kinda weird. But, since I’m adding that capacitor, when node ON is pulled high, the last voltage will be held — a sample and hold! Cool, eh?

Spesal Cuck and Mido modulation

Everyone who’s built a Ciat-Lonbarde DIY board knows that Mido is the +4.5 volt bias which the internal signals center on. For Fyrall, Peter implemented Mido modulation. Why this is called Spesal Cuck is anyone’s guess.

With no input Mido stays at 4.5V. Patching a (DC-blocked) signal into S.C. lets you waver Mido around, with the amount of modulation set by a panel pot. It’s important to note that Mido is used only by the oscillator section Z in Fyrall. The only thing you’ll be messing with is the oscillators — all at once! So, if the inter-modulation of the oscillators, with interaction by signals coming from the Orranger, isn’t enough for you, just patch something into the Spesal Cuck.

Other circuits on the PCB

There are two Sections P, which each consists of two more oscillators that take signals out of Orranger and use these to modulate their frequencies. I’m planning to not build these, but possibly repurpose the op amps for something else.

A speaker amplifier also exists on the board, but I won’t use it.

Finally, there are two output amps, called “Crystal Transfer Station in the Formation.” These take any input from Fyrall and buffer it for a Studio Out signal, as well a producing a “Stab” output, which will pull down any signal patched to it. I will use these.

The Future

This analysis was my first step with Fyrall. Eventually, I will get around to planning my peculiar implementation of it.