Frak, Jack and Roll

I promise I'll try to make better titles in the future.

Bolstered by my success with the ARF VCO and Fry's actually having some parts I want for a change, I decided to "finish the job" of re-jacking all the banana plug connections.

This is a view of the half and half stage, with the Noise module completed, but the LFO still rockin' the banana plugs. I started off with the noise module because it's the one I could most afford to mess up. As a side note, it, like most of my modules, is based off of the simple reverse biased transistor noise source, amplified and filtered to get "pink" noise and a random wobble. The "Volts" knob is just a 0v-5v constant source, which is sometimes useful for biasing a CV parameter or testing something out.

That reminds me - I did some tests with my oscilloscope to make sure I had a good understanding of what voltage levels Wally uses internally. My tests show that I do recall correctly - Most of my signals are in the +/- 5v range, with the envelope generator at maximum output swinging from +/- 10v.

Just so you can get an idea what my late-90's DIY work looked like on the inside, here's a shot of the rear of the LFO module. The orange caps are power supply bypass caps, and as you can see, this is a picture from before the banana plugectomy.

Finally, here is a close-up of the finished DIY part of the FracRack, now completely 3.5mm connected. The "Output" panel is not really a separate module. It's connected to the VCA's output jack. With the Vol knob at 12 o'clock, Wally's signals come out about line level, so he won't blow away my mixer's input bus.

I had a brief scare with the VCA. It stopped working during the middle of some testing. It seems that one of the ICs had a faulty solder joint. Hopefully, I won't have any more trouble with it.

Here's a kind of blurry (sorry) picture of Wally, sporting his new chrome-colored connectors. I still have the "1/4 to banana" adaptor panel (right above the ARF VCO) because I found that with cross-patching the banana plugs, I can use it as a 3-way multiple, and feed pitch CV into both VCO and VCF. I'll have to make a real multi at some point, but for now, this hack will work.

So right now, Wally has 7 rack spaces of usable modules, and a total of 9 individual functional units, if you count the dual A/R generator as two functions. Power distribution takes up 3 spaces, and there's a double blank in the FracRack and a single space blank right above the PSU.

With his current configuration, he can make some musically useful sounds, some cool wind effects, and some cat-destroyingly horrible screeches. When I get some time, I'll try putting together a video or something.

My next project will be to rework my power distribution scheme. Fry's had the molex connectors I need for the project, but I have some thinking to do before embarking on that project. Largely because the FracRack also has a PSU in it, and the MIDI->CV converter runs from a separate wall wart. Ideally, everything should run from the same "on" switch.

One other consideration - Ikea makes a nightstand that is exactly the right size to use for an 6 rack-space high enclosure. It can go to 8 spaces if you mount the bottom shelf lower. That would be a lot nicer than the open frame metal thing I'm using now, and for about $30 total, I can add some protection to Wally, as well as hopefully a place to mount the PSU that doesn't take up front panel space.

ARF VCO - No Bananas

Last night, I went ahead and performed the "bananaectomy/phonejackoplasty" operation on my ARF VCO. Sorry for the quality of the pictures. I'm not sure how to insert proper links to the larger images. Blogger seems to just want to use the thumbnail version.

The original panel was hand-drilled, hand-labeled, and fitted with banana jacks. Since this is a "free rack" module (that is, a 19" rack unit, not a FracRack or other sub-rackspace format), it's destined to get 1/4" jacks.

The conversion was both harder and easier than I was expecting - the banana plug mounting nuts were a bear to get loose until I remembered I had an appropriately sized nut driver in the garage. The wires however, were pretty easy to get loose. I wound up labeling them with masking tape so I wouldn't get confused about which wire went where.

I also performed the conversion one column at a time. Here's the view after the first column was replaced.

Grounding is always a problem with 1/4" jacks. Though they were mounted to a metal face plate, and that provides a common solid grounding point, I also decided to go ahead and wire each ground lug together. I "snaked" the ground through all the lugs, and attached a fly wire from the last one back to the PCB's ground area, approximating a star grounding configuration. I suspect this step was unnecessary, and my multimeter's continuity tester showed that the panel was already electrically connected to the module's ground line, but since my other ARF module wired all the ground lugs together, it seemed appropriate.

Here's the final back panel. After hooking everything back together, I returned the module to its home, and everything is working fine! I really do like the feel of phone jacks over bananas. It's so easy to repatch things that nothing gets in the way of sonic exploration.

The whole project took around two hours, including time for trying to find hookup wire and nut drivers. Not too bad, though now I'm out of 1/4" jacks, and future work will require actually spending some money. Alas, can't be a cheapskate forever. ;)

Here's the VCO patched up for a simple "play me a square wave mixed with a triangle wave, with vibrato added from an LFO" voice.

The one down side to this scheme is that I lose the color coding for the SYNC plugs, which are not V/Oct control voltages.

Stop Jackin' Around and Give Me Some Power

So my conundrum about how to reconcile the fact that Wally uses pretty much every different connection standard ever gained another twist.

I asked my wife (who is way more practical-minded about, well, everything than I am) for some advice. I'd been trolling the PAiA and Blacet websites, ogling FracRac format modules I can't afford (don't judge), when it occurred to me that most of my modules are actually 3.5mm jacks, and so are the jacks on some of the other gear that I've been lusting after (Arturia MicroBrute, Korg Volca series). Well, that's not strictly true, most of my jacks are Banana Jacks, but I'm almost positive that I want to move away from those.

"So," says my wife, "why not just use 3.5mm jacks for everything that's Frac format, and 1/4" jacks for everything else?"

Why not, indeed?

So that means I can convert the LFO, VCA and Noise modules in-place by swapping out their jacks. The ARF VCO will still get 1/4" jacks to match the ARF VCF and A/D Generator.

I'll need to get more 3.5mm -> 1/4" cables. I already have 3, and they work fine. I'll also need a fist full of 3.5mm patch cables. Blacet makes a cable kit I could use to get both, but I should shop around some. I also just discovered Stackcables, which makes banana jacks even less desirable. Though at over $8 a pop, I won't be getting a ton of these. :)

This leaves the question of power distribution. At the moment, the FracRack itself has a simple but effective +/-15v supply in it. Power from that is distributed by basically the same sorts of connections we used to use to power disk drives in computers. It turns out that that's a TERRIBLE way to distribute power to analog equipment. But what can I say, I was young and needed the milliamps.

My "big boy" power supply uses three screw terminal blocks (+V, Gnd, -V), and expects each module to have a 3 wire connection that can clamp down. Electrically, it's sound enough, but it's relatively hard to attach and remove things.

My old MOTM stuff had this great block sort of like this one, and lucky me, I found the extra one I had! So, clearly, it's time to retrofit my system to use the more stable and convenient connectors. Blacet sells a kit of headers and wire which I can use. I might be able to get equivalent parts from Mouser or Jameco, so I'll be doing some shopping.

So there you have it - 3.5mm jacks for my FracRack stuff, 1/4" jacks for my "free range" modules, and MOTM-style power blocks for everyone!

Time to start saving my pennies.

Touch Keyboards and Step Sequencers

Aside from module renovations, I'm also planning on creating some "modular synth" style controllers. I do have a MIDI interface which works well enough, but I've always been intrigued by step sequencers and "Serge Touch Keyboard" type controllers.

I'm still very much in the brainstorming phase, but essentially, I have two projects planned - a simple 8 stage step sequencer and a capacitive touch keyboard controller (TKB). I'll have a lot more to say about these in the near future, but for now, consider this a sort of "statement of intent".

I've also started a YouTube channel to put up some video reference for these projects. So far, it's just me running my finger over the pins of a touch sensor chip and a few linked videos of devices like I'm considering. I plan to shoot a similar 10 second video of the 4-stage prototype sequencer I breadboarded over the weekend. It's also just some flashing lights at this point, with a rate controlled by a knob, but it's progress.

My sequencer is based off of a 4017 decade counter and a 555 timer. Hardly a creative and unique design, but I need something simple for my first time. I'm planning to base it off of the version in Barry Klein's book, which in addition contains some logic to support external and manual stepping controls. I'm going for a single CV output, portability and the ability to run off of a 9v supply.

The TKB is similar - this first version is 8 stages with touch-capacitive plates to activate each stage. The stages will be individually tuneable, so the keys won't correspond to piano keys, unless you tune them that way of course. So it's a very small keyboard at the moment, but I'd like to save the expense of a larger keyboard for after I've discovered what's horrible about my design.

The TKB is going to be controlled by an Arduino, since I'm better with software than hardware. The sensor I'm using is an AdaFruit CAP1189. I'd like to build some sequencing features in as well, so you can either play it like a keyboard, set it running, or both. The Korg Volcabeats is an inspiration, as is of course the Serge TKB.

A side note about the CAP1189 - it seems from the datasheet that I could use the built in LED driver hardware to trigger the key stages, as the chip can be set up to output whatever values you want from those ports instead of echoing the touch plate sensors. I had originally planned to use 74HC595 shift registers, driven by the Arduino using data pulled from the CAP1189. But if I can use the CAP1189 for both duties, then I can save some board real estate and part counts.

Sounds like a good next step for the software part of that project.

A quick note about schematics

I'm not much of an analog circuit designer, though I did spend a few years in Electrical Engineering before giving up hardware for software. So mostly I adapt existing circuits. I do generally understand the principles behind what I'm building though, and I've made a few modifications here and there to basic designs.

Aside from the commercial kits, I got most of my schematics from the web and books like Terrence Thomas' "Sound Synthesis: Analog and Digital Techniques" and Barry Klein's "Electronic Music Circuits", which was heavily influenced by the classic newsletter, Electronotes. If I were going to suggest a single book to someone looking to do modular synth DIY, it would be Barry Klein's book. I've noticed there's a PDF of it that Google finds easily. I'm not sure if it's a legitimate copy though.

My dual Attack/Release module is almost exactly Barry's version, except that I put a Serge Modular inspired inverter/scaling control on the output. (It's a knob that takes the AR's output and lets you vary it from, say, a +5v signal all the way down to a -5v signal. So the AR's can subtract or add to whatever you send them to. Pretty neat.)

My LFO is a straight up copy of Barry's LFO With Variable Triangle Offset and Symmetry, with a two color LED mounted in it so you can see it swing positive and negative.

And my Noise circuit, which is really nothing special, is also clearly derived from the Electronotes simple noise source, which also appears in Barry's book.

I used Thomas' design for the octavizer/pulse divider module, with an added comparator to condition the input. It's still in the "half constructed phase", though a breadboarded prototype worked pretty well.

First Steps

This post is a more rambling discussion of the basic steps I need to get what I have working better. I naturally have tons of ideas for new projects related to Wally, but I'll do my best to stick to maintenance and renovation of Wally's existing modules. This is nuts-and-bolts stuff, but it needs to be said before I can get to more interesting topics.

Clean and Repair

Since Wally was in boxes for over a decade, and those boxes were stored in my garage, some parts of him aren't working at 100%. Pretty much all of his potentiometers are scratchy, and a number of wires had come loose.

The VCA and LFO are in the most serious need, with the other modules being only intermittently scratchy. Some DeOxIt should fix things up.

The Noise and Dark Star modules actually had some broken wires in them. I've already re-soldered them, so they're working well enough.

Side note: actually, the Dark Star wasn't working at all. In fact, I'd never gotten it to work, since it was one of my earlier projects. With a little help from the good people at Blacet Research (who made the Dark Star) and some diligent circuit tracing, I think it's mostly working now. It's a strange enough module that I'm not entirely sure. But it responds more predictably to its controls now, and I've managed to CV all the parameters successfully, so I'm going to say it's working now.

Tuneups

The MIDI->CV converter and my VCO both have a lot of trimmers in them. It seems like almost all of them need to be trimmed.

The MIDI->CV box has a trimmer that tunes 1v/oct sensitivity. If you know anything about guitars, it's almost the exact same thing as setting intonation - the pitch of each octave relative to the other. The goal is to get exactly one octave for exactly one volt. The control is VERY touchy though. I got it pretty close the other day, but it might still be off. It would be much easier if this trimmer was a multi-turn trimmer, but it's not, and the PCB isn't designed to make it easy to swap one out.

The VCO has trimmers of its own for mostly the same purpose. That's right, both the MIDI->CV and Oscillator can go out of intonation, and they both need to agree with each other. In addition, this VCO has a radial octave dial that lets you pitch it up or down an octave relative to the input control. At least one of these is out of tune with the others. That will be "entertaining" to adjust as well. Fortunately, Ric built these all as multi-turn trimmers, so I should be able to get very precise trim.

I suspect the VCF has a similar v/oct adjustment, but given my current inability to put pitch CV through to both VCO and VCF, it hasn't come up yet. :)

Circuit Upgrades

The Attack/Release module, based off of a modified circuit in the Barry Klein book, is pretty good, but the timing control pots are linear 5M ohm pots. While they have VERY long decay times, their precision for short times is touchy. They need to either be replaced with Audio taper pots or converted with that "add a parallel resistor" trick. Since the pots in that circuit are just "rheostat" type variable resistors, and I don't mind having my resistance cut in half, I think I can get away with it. Or maybe I'll just buy 4 5Meg Audio taper pots and be done with it.

Another upgrade is to replace the second AR's gate input jack with one of those "switch normal" jacks, and permanently hook the "normal" connection to the first AR's gate. That way, I can run them both with one signal, and only break the second one free if I plug in a second signal. I've hacked this with an alligator clip, and it's very useful. I need the different jack/switch part to make this a permanent change though.

I plan to do more with this sort of "normaling" in other places.

Much Ado About Jacks

I need to deal with the heterogeneous jacks issue. I can't really upgrade the 1/8" jack modules to 1/4", but I can replace the banana jacks in my VCO. The panel was already drilled for the larger jacks, so it's really just a monster soldering effort. There are 15 jacks to upgrade. Fortunately, I have 20 good quality jacks left over from the Old Days, so I can do the upgrade for nothing but time. I'll admit that the job scares me a little, since I've never been good at panel hookup.

The FracRack banana modules are a bigger problem. There just isn't the room to upgrade them in-place. I might be able to 1/8" jack them, but I think I might want to re-panel them.

FracRack Modules

These modules, specifically the Noise, LFO, VCA and Output modules, present a problem. For one thing, since they were the first things I had ever built, they're cobbled together on perfboard and not well mounted to their panels. They also have banana jacks, and can't possibly fit in their same sizes with the larger jacks. The FracRack is pretty tight for space as it is.

Consequently, most of this work seems like it's going beyond First Steps and into Next Steps. I'll need to keep these modules around, since really all of them are essential to Wally's utility as a musical instrument.

State of Wally, October 14th, 2014

At the moment, Wally consists of these units, from top to bottom:

• PAiA FracRack containting
• PAiA MIDI->CV8
• Blacet Research Dark Star Chaos
• Blank panel
• Noise (White, Pink, Brown, and a knob/jack that puts out a solid voltage)
• VCA (CA3080 based)
• Output (just a volume knob, decoupling cap and an output jack)
• 15v bipolar power supply.
• Juice Goose power strip (master control for Frac, MIDI->CV8 and other power supply)
• 1/4" to Banana Jack interface. Also has a simple CV pedal and foot switch interface.
• A.R.F. VCO - CEM based oscillator with a unique "wave slicer" that allow you to combine top and bottom halves of different waves. (Note, ARF is gone from the web, but here is a link to a guy with a lot of Ric's stuff.
• A.R.F. VCF - CEM based 4 pole low pass filter
• Dual Attack/Release generator with "+/- output" (knob that varies output level from positive to negative) Custom design from Barry Klein's book.
• Blank Panel
• 15v Bipolar Power Supply (much beefier than the little one in the FracRack, originally used to power the MOTM stack.)

I also have a Ring Modulator that is finished and an Octave Divider PCB that's finished and mostly just needs to be wired to its panel. The RingMod was made back when I had plenty of space - spending a whole rack space on three jacks and two knobs seems extravagant now.

With that stack, I can (and have) wire up a simple mono-synth configuration to play around with. I'll try to get some short video together sometime.

The frame I'm using is a 12 unit high "rack mount" frame, which is very cramped. Ideally, I'd have a box with front/back mounts, and move the power supply stuff to the rear.

Internally, my modules mostly follow the "Electronotes" standard - 15v bipoloar power, control voltages in the +/- 10 volt range, and audio in the +/- 5v range, though as audio, it will often have much lower range. I don't have a decent power distribution scheme yet - I just run 3 wire bundles to a trio of terminal blocks. The MOTM scheme was great, but unless I find a cheap alternative, it's too expensive to re-adopt.

One concern is that at the moment, the system is about half and half banana jacks and 1/4" jacks, with a few 1/8" jacks thrown in for good measure. This makes it very trying to connect things together. I'm disenchanted with banana jacks these days, and I like the feel of 1/4", so even though they take up a ton of panel space, I'm considering standardizing on them.

Why so many jacks? Historical reasons. Originally, I was planning to do everything with banana jacks, since they're cheap, reliable, and can "stack". But MOTM was all 1/4", and when I went to that system, I converted a few of my favorite modules (VCF, A/R). The VCO is a prodigious amount of soldering though, so I hadn't gotten around to that one before the Dark Ages struck. Likewise, the PAiA and Blacet stuff was 1/8" only, since it was designed to fit in smaller spaces. The FracRack was mostly put aside when MOTM moved in, though ironically it's back to being very important in the new system.

Introduction - Hello Wally

The purpose of this blog is to help me chronicle my efforts to create my own modular synthesizer - "Wally", short for "Wall of Knobs".

To be more precise, I began working on electronic musical synthesizers back in the late 1990's, and managed to DIY a few modules as well as built a number of Synthesis Technology's "MOTM" modules. By 2000, I had lost the drive, and a few years later I sold my MOTM modules and packed away the rest of "Wally".

Recently, I've experienced somewhat of a Renaissance in my interests, and I re-assembled what parts of Wally I still had around, just to see what I could do with him. The results were encouraging, and I've decided to resume work.

Here is what Wally looks like right now, spread all over my parlor:

Wally, a Theremin and some old TAMA drums, posing in front of the piano.
Excuse the mess.

Not much of a "wall" anymore, that's for sure. Maybe a nightstand of knobs. He was a lot more impressive when he was 3x that tall in a large rack enclosure with two rows of MOTM modules. The link above has a picture that's closer to the mark - the tall skinny one second from the right.

The "ARF" modules were built for me by a fellow named Ric Miller, and they're the majority of the working parts of this synth - the VCO and VCF. The PAiA MIDI->CV8 converter I built from a kit, and the rest is my design, done on a variety of perfboards and copper-clad boards.

In future blog entries, I'll discuss my plans, dreams, and progress. I hope you'll join me for the ride.

(And equally, I hope the ride actually gets out of the driveway.)