The Green Machine is a primordial project originally built in the 1990s that predates my conception of electroniums. But it is a sentimental favorite, and I recently decided to rebuild it. So there’s a story there. (comme toujours) It’s housed in an old plywood drawer from a desk I’d used since grade school. It’s covered in a collage of photos that used to clutter that old drawer. The inside is spray-painted (very badly) with the paint I used to paint my favorite Big Daddy Ed Roth hot rod model, ‘the Beatnik Bandit’. I briefly considered calling this monstrosity ‘beatnik bandit’ because of its bias toward rhythmic ostinati. But when you look at four square feet of a bad fluorescent green spray-paint job, it just seems silly to think that any other name than ‘The Green Machine’ would stick.
It ain’t no invention either. Under the hood is a clumsy frankenstein of an Arp Axxe and a Roland TB303*.
It was first assembled in the mid 1990’s. At that time I was much less experienced and proficient at electronics. I lacked the confidence to build my own audio circuits and so I stayed in the shallow end of the pool, modifying existing devices. Mostly, these devices were broken analog synths that acquaintances would give me for salvage. I am too embarrassed to reveal the number of Arp Odyssey’s I was given AND the number of them that survived my ‘repairs’. This thing is homely in every way. All the ‘dramatic’ imagery in this post is an ironic parody of ubiquitous tawdry synth-porn on the web. This old beast looks like crap on the surface, and under the hood there is a wealth of evidence of how much I DIDN’T know when I started taking a can-opener to these old synths.
And yet…
It sounds GREAT. And for Bass this unlovely miscreant is peerless. It’s a tribute to Mr. Pealrman and Kikumoto-san that my maladroit misadventures in the guts of these synths could not diminish their seismic bass mojo.
On that note, I want to say a word about the video content in this article:
I urge you to listen to sound at least once with some really good speakers or headphones. I managed to capture a hint of the bottom-end and little transducers just won’t reveal that. Also, my protestations aside, the backlighting feature looks as cool as I hoped. The ghostly circuit traces and lurching shadows are vivid . But my ability to capture the detail is not up to the task and the videos only hint at the way it looks in real life.
So with those introductions and disclaimers declared, on with the show….
The Axxe in the Green Machine got off easy. Most of it works. The biggest casualties in the audio signal path were the pulse wave converter in the VCO and the VCF. Both were broken when I got the Axxe. The VCF is, of course, a rather crucial failure and my solution was to buy a DIY replacement VCF submodule from Phil Cirocco at Discrete Synthesizers . Phil is an elite, legendary designer and an expert on Arp repair and restoration. If you need an Arp resurrected send it to Phil (and not me!)
The VCF replacement module Phil created makes the Axxe sound better than it ever did and the module is switchable from Low-pass to High-Pass mode; a capability that the Axxe never had.
The TB303 was also broken when it was given to me. The Green Machine was first built in the mid 1990’s. Vintage analog cognoscenti might remember that that period was exactly when the Acid House fad was peaking, and prices for used TB303’s were exceeding $1000. So I’m keen to declare that when the owner of the broken TB303 gave it to me, I stridently pointed out that these flimsy little plastic boxes were worth serious money, even when broken(!). The owner was fixated single-mindedly on cleaning out his basement and assured me that he had no time to waste on trying to sell any of his junk.
The primitive sequencer processor in the TB303 was failing intermittently and the memory chips were completely faulty. But the analog sound engine was intact. And I could work with that.
In its original form the Green Machine saw a lot of action in the 1990’s but it fell into disrepair and I shelved it for about fifteen years. I’m happy to say that I learned a lot in that decade and a half. And I decided to fix it up again and bestow the benefit of what I’d learned. Now the old green beast is back.
In detail let’s look at a block diagram…
A few comments about the block diagram itself:
First off, I found the original hand-drafted block-diagrams from Arp and Roland on the internet. (the internet is evil, but also wonderful…. like beautiful villainesses). This bygone aesthetic charmed me so I appropriated the elements into the big messy amalgamated diagram above.
And, technical cognoscenti will notice that the signal paths depicted in the diagram are, themselves, messy. This is usually an indication of an edifice that was not conceived as a whole and erected from the ground, up. … but rather, a rambling cluster of elements that were built at different times and with no intent that they’d ever be integrated with anything else…much less EVERYTHING else…..And that impression would be correct…
…because that’s exactly what the Green Machine is. But I will do my best to outline a ‘big picture’. The Axxe and the TB303 are both basic single-VCO, analog monosynths. with a general topography like this:
Note that this classic fundamental design is designed around a keyboard. In the case of the Axxe the keyboard is obvious. Onthe TB303 the ‘keyboard’ is a row of little buttons arranged in standard Faber keyboard form. These buttons can be used to play the TB303 in real time or enter a repeating sequence.
As any one who knows me will attest, the doctor does not consort with keyboards. So ,at the most fundamental big-picture level. I implemented external inputs for pitch control voltages and gates. Then I normalled random voltage generators (blue lines) and an internal clock-pulse generator (red lines) to them respectively. And added a mixer (purple lines) to combine the two voices at the end of the signal chain.
So the big picture is two parallel voices driven by a single clock and a single random voltage source.
As described in the illustration key of the block diagram, red elements indicate where I added circuitry. You can see, at an glance, that I added a lot of features to the Axxe voice.
This is for two reasons:
Arp intended the Axxe to be an entry-level instrument, so they used existing designs from their more-developed instruments, like the Odyssey and the 2600 but they left many features unimplemented. This is catnip to me because expanded capability is just sitting there under the hood. And all I have to do is build a way for the user to get at the extra features.
And … The physical circuit board of the Axxe is large, robust, and easy to interpret and modify. (They REALLY don’t make like that anymore!)
This is why there’s a lot of red on that diagram above. Two extra features that really appealed to me were: adding an independent LFO dedicated to modulate the VCO’s pulse width, and souping-up the LFO by adding modulation of the LFO rate and changing the simple square wave output to a random step voltage.
I was very happy with the sound and there’s no reason that shouldn’t have closed the case and ended the story. But this thing is big. And due to it’s rather ungainly size, I’ve always felt an obligation to do something visual with it. When I first built it. I had delusions of grandeur about LEDs moving behind the somewhat-translucent Arp circuit board. That was a pipe dream. High current would be required in order for light to be more that feebly visible through the circuit board, and high current, shared grounding, motors, and other requirements were simply beyond my skill level back then. So I covered the back of it with a collage of personal memorabilia…… (Note that it can practically camouflage itself among the workshop clutter)
…..and got a can of fluorescent green spray paint and settled for that.
But the old dream persisted.
And I had acquired experience with high-current power supplies, and optical isolation techniques that would allow voltages in the audio circuitry to Interact with a lighting display on a separate power supply. But, making the lights move without wearing-out or twisting-up the wires that power them is not a trivial engineering problem.
My solution was this.
The lights would stay put. But I could move the light around with mirrors on a motor.
It looks like this in action…
The little pyramid of mirrors is turned by a stepper motor that moves one step with every gate-pulse so the light behind the circuit boards moves at the tempo of the envelope gates. the direction of each step is derived from a complex algorithm that does a boolean NAND of the clock divider (Block Diagram <7>) and the state of sample/hold #2.
The assembled overall effect looks like this:
If there is a moral to this story, it’s Never Give Up On A Good Analog Circuit. Sometimes it just takes patience for your skills to catch up with the potential.
That’s the high minded version. I think if these circuits could talk, They’d say: ‘Hook me up to a big-ass speaker with some power behind is and I’ll shake every last boo-tay in this here room’
And finally deep respect to the designers who made the elements of this beast.
Alan Pearlman (the Axxe)
Tadao Kikumoto (the TB303)
Phil Cirocco (the VCF replacement circuit in the Axxe)
and, as always, Grant Richter, my ‘Mr. Miyagi’ who taught me every good trick I know by the wax on…wax off method.
*the story of how I acquired this TB303 a colorful but LENGTHY tale; too lengthy for this article