Scott Stites Synth DIY
Birth of a Synth


Birth of a Synth | Modules | Buchla 291 BPF Adaptation | VTL5C2 Phase Shifter | Soft Sync | TriPhase/QFG, Sample and Hold,DSC2000 | 291 Filter Meets Dark Star Chaos 2000 | Patches | Schematics | Synth DIY Links | Kits, Modules and Systems | Synth DIY Music | Contact Me

Diary of a mad synth builder.

12 December, 2004

Yesterday my lab assistant and I took a road trip up to Manhattan to visit Jeff. Prophet 5 through the Dimension C = Yummy.

We finalized our plans for the application. We decided the Dim C function alone certainly deserved a rack application.

Currently the plan is to use a dual ganged pot as the input attenuator for both channels, provide a control to attenuate external CV's, yet another to select between mode switch and depth and freq pot controls, provide a switch to select frequency range of the internal LO, another switch to switch between external CV and internal LFO, provide an input jack for remote bypass, and have a dedicated TZF output.

Later, a more specialized TZF device may be in the offing.

Here's a sample I recorded Friday night last - it consists of several tracks of the modular through the Dim C. The Dim C provides great effect for Polyphony, but also, as demonstrated here, is great for multi-tracked monophonic devices. Each track is modified by the Dim C. Various modes of the Dim C were used, as well as mono->mono and mono->stereo operation.

Modular Through Dimension C (3.1 MB)

06 December, 2004

I had a little time tonight to try something else that sprang up on the DIY Stompbox list. Mark Hammer (as usual) put in an interesting post about TZF modulation. Specifically, asymmetrical versus symmetrical modulation.

When I mention 'asymmetrical modulation', I'm in part really talking about how flangers normally work. With a normal flanger, you have a mixed 'dry' signal and a delayed signal that's swept back and forth against the dry signal. Delays can't travel time, so a normal flanger never can get to the point of through zero flanging, but the closer they approach the dry signal's 'time frame', the more jet-like the sound becomes. That's why the A/DA was such a good flanger. It could sweep to a very small amount of delay.

With the Dim C TZF, we don't have a fixed dry signal. We're actually moving two delayed signals inversely in time. It's like two cars driving at the same speed approaching each other, and the point of impact is the point of TZF. But these cars are in different lanes, so they just pass each other =-D. They each arrive at the end of the road at the same time, turn around and pass each other again, ad nauseum. So, the sweep is always symmetrical, with TZF occuring in the middle.

Now, if we put (yet another) switch in there and (yet another) control, we could make one of these cars stop anywhere on the road we wanted it to stop. The other car could then pass it at different points along its route. Say we place our stopped car at one end of road. The moving car would travel almost the entire length of the road, pass the car, turn around, and pass the car immediately again on the way back. In this way, we could control the flanger like a 'normal' flanger and also be able to set where along the sweep the TZF occurs.

But, why stop there? What if you made one of the cars slower or faster, and/or made it move back and forth on a smaller stretch of the highway? This could be done rhythmically, randomly, or on a note-per-note basis.

The way to do that would be to have a switch that allowed you to apply a voltage through a pot to one of the BBD clocks (the switch and pot mentioned above). A switching jack (or more likely another position of the same switch) would allow you to apply a CV through the pot to the clock instead.

This would actually allow one to control the BBD's individually with two separate controls and two separate CV's, or both BBD's with the same CV. All the while, still preserving the original function...hmmm, what was that? Oh, yeah, the Dimension C =-D.

In tonight's experiment, I removed the LFO input from the second BBD clock control input and injected a scaled DC voltage from my mixer module directly to it. I recorded a sample of its operation along with a sample of the 'normal' symmetrical modulation of the Dim C circuit.

The following sample is of filtered noise injected into the Dim C mono input, with the output of it taken from the passively mixed TZF output. This was recorded using the normal symmetrical modulation of the Dim C.

TZF, Mono, Symmetrical Modulation (878 KB)

The next sample is of the asymmetrical modulation, with one of the BBD's under manual control. It starts off with the input at the low voltage range. Listening to it, it becomes obvious that the swept delay comes just to the brink of TZF, then backs away. At around the 00:32 mark, I manually move the fixed delay to a shorter delay (this can be heard as a disruption of the sweep effect). For a while, one can hear that the swept delay sweeps up to the point of TZF, passes it, turns around and comes back. After a bit, I begin slowly moving the fixed BBD back down, stopping here and there. The TZF effect can be heard occuring at different intervals. Note the apparent 'slowness' of the LFO compared to the Symmetrical sample. In actuality, the LFO speed has not changed, but because one of the 'cars' is stationary, the relative speed between the two cars is now lower =0).

TZF, Mono, Asymmetrical Modulation (1.63 MB)

05 December, 2004

With the stereo input function 'in the can', my next area of concentration on the Dim C is "Through Zero Flanging" (TZF for short).

Mike Irwin posted on the DIY stompbox list that if the outputs of both expanders are passively mixed, a pleasing through zero flange effect can be obtained (thanks again to Mike!). I confirmed this, but am exploring the possiblility of making this function actively mixed, so that one of the expander outputs can be inverted, allowing both positive and negative flanging effects to be obtained.

Further, I figure that if two outputs were provided, with one output mixed with an inverted signal, stereo flange effects would be possible.

I mixed the two outputs passively, then ran the same two outputs to my mixer module and actively mixed those, one signal inverted and one signal non-inverted. I ran the outputs of both mixes to the D8. This is a sort of 'on the fly' test, but it appears that interesting stereo results can be obtained. More work must be done to incorporate it in the circuit. One thing to consider is a de-emphasis circuit on the output of each TZF signal.

This first sample is of the DSC2000 noise generator being processed through a filter and VCA by the TZF mixers:

Noise Mono to Stereo TZF (1.03 MB)

This second sample is of a single sawtooth VCO being processed through a filter and VCA by the TZF mixers:

Sawtooth Mono to Stereo TZF (939 KB)

Here's a quick and dirty experiment using regeneration with the TZF.

TZF With Regeneration (1.3 MB)

04 December, 2004

My synth time this week was spent working on a modification for the Dimension C.

While studying the both Dimension C and the Dimension D more closely, I started looking at the differences between the two devices.

As I understand it, the Dimension D differs in these parameters:

1. It operates at higher 'studio' levels
2. It has balanced inputs as well as unbalanced inputs
3. It accepts either stereo or mono inputs (whereas the Dimension C only accepts mono input).

Obviously there are other internal differences, otherwise, their general functionality and theory of operation is identical. My goal is not to build a Dimension D, but at the same time, some of the functionality of the D would be advantageous to a rack mount version of the C.

The main thing that interests me is the ability to process stereo signal inputs. Looking at the Dimension C schematic, I could tell all the hooks were there, it just was not implemented. I would imagine there are several reasons why Boss chose not to put in stereo inputs - BOM cost, the sheer number of components in the original C to fit in a stomp box, and its target market of chiefly guitar players. Stereo guitars do exist, but for the most part, guitars are mono instruments. Obviously where the C really shines is its ability to take a mono signal and create a shimmering stereo field. It also has the ability to process a mono signal input and produce a mono output, something I'm not sure the D was capable of doing (somebody please correct me if I'm wrong).

As I mentioned, the hooks seemed to be in place to modify the C for stereo input. What was needed is the following:

1. A separate pre-emphasis circuit for the second channel (the second de-emphasis function was already in place for the stereo output).

2. A separate compressor for the second channel - this was already in place, but disabled, as the original C mono input only required a single compressor. The unused half of the first NE570 could be used for this.

3. A separate low pass filter preceding the input to the second BBD (again, the second BBD already had a low pass filter after the BBD).

4. A 'slight bass boost' buffer for the dry signal, and an additional path for the unaltered dry signal as well.

5. A means to route the first channel to the second channel when a mono input signal is used.

To make an already long story a bit shorter, I added all of this to the Dim C (using a separate 'satellite' breadboard). It worked perfectly! My main concern was that it would affect the sound and functionality of the original circuit. That was all preserved - it's mono to mono and mono to stereo functions still sound the same.

My other concern was that sending the stereo signal down the separate paths reserved for a mono signal would sound 'warbly'. That was not the case.

Below are two sets of samples to illustrate the stereo input function. To begin with, no piece of gear I have is stereo by default, so I had to create a stereo signal with the modular.

The first set of samples uses a sine VCO modulated by another sine VCO to create a bell-like sound. The signal is routed to two different 24dB VCLPF's. The output of each LPF goes through a separate VCA straight into the stereo inputs of the D8. One LPF is modulated by a slow triangle LFO to create some movement in the stereo field. Both VCF's and VCA's are also controlled from the same EG. The VCO's and the cutoff frequencies of both filters are controlled by a simple eight step sequence.

The samples start out in the bypass mode (dry) for a few bars, then the effect is switched in. At the end of each sample, the effect is then bypassed again, for comparison. There is one sample for each of the four modes of operation of the Dimension C.

FM 'Bell' Stereo Input Mode 1 (756 KB)

FM 'Bell' Stereo Input Mode 2 (761 KB)

FM 'Bell' Stereo Input Mode 3 (852 KB)

FM 'Bell' Stereo Input Mode 4 (841 KB)

This next set of four samples uses the same sequence, but the VCO frequencies have been altered to provide more harmonic output (a 'harder' sound, if you will). The resonance on both EG's has been cranked up, and the filter that is not modulated by the LFO is instead modulated by an inverted copy of the EG signal. This is to provide more obvious movement in the dry stereo signal. Again, there is one sample for each mode of operation of the Dimension C. Also, the same pattern is followed: dry signal, effected signal, dry signal.

FM Stereo Input Mode 1 (801 KB)

FM Stereo Input Mode 2 (818 KB)

FM Stereo Input Mode 3 (818 KB)

FM Stereo Input Mode 4 (829 KB)

27 November, 2004

Thought I'd add a little visual information today.

Below is a picture of the Dimension C on breadboard. It's a lot less dense than I usually breadboard - it's a rather twisty, windy circuit, and I breadboarded with an eye towards correcting mistakes. Actually, it worked right off the bat (besides the fact that the schematic had a wrong connection illustrated, which required the removal of a jumper).

Dim C Breadboard

Click here to enlarge picture.

Next up is a pic of the KS-01 as it sits today (the completed modules, that is). I haven't begun to tackle the issue of front panel legends, so most modules have the functions sharpied in on the front.

From left to right, the modules go as follows:

Ray Wilson Quad LFO with Thomas Henry Sine Shapers
Dual Mixer/Peak Trough/Multiplier (Peak Trough adapted from CGS Analogic)
Blacet DSC2000
CGS Wave Multiplier/Grinder
Rene Schmitz 2040 Filter/Mixer/Mult
Ray Wilson Quad VCA

KS-01 Synthesizer

Click here to enlarge picture.

Here's a pic of the power supply and power distribution board I built. The power supply is a Midwest Analog Platinum supply housed in an old power supply chassis that the company I work for donated (rather than being discarded).

In the lower left of the picture, you can see part of the Thomas Henry keyboard controller, and next to that is a Ray Wilson re-triggerable AR envelope generator - the one used to create 99% of the envelopes in all of these samples. It's been on the breadboard for over two years LOL.

Power supply and distribution

Click here to enlarge picture.

26 November, 2004

I've got the Dimension C connected to the D8 in stereo, and I finally can here what the stereo mode sounds like. It certainly does not disappoint!! This has got to be one of the sweetest effects I've ever run across.

Following are four samples using the stereo operation of the Dimension C. The Yamaha SHS-10 is feeding its 'horn' voice (mono) through the Dimension C. As with the previous samples, the sequence is heard 'dry' then with the effect in operation. There's definitely some great stereo expansion going on with the Dimension C.

Yamaha SHS-10 Through the Dimension C, Mode 1, Stereo (769 KB)

Yamaha SHS-10 Through the Dimension C, Mode 2, Stereo (749 KB)

Yamaha SHS-10 Through the Dimension C, Mode 3, Stereo (738 KB)

Yamaha SHS-10 Through the Dimension C, Mode 4, Stereo (778 KB)

Below is a link to the Dimension C schematic (listed as DC-2):

General Guitar Gadgets - Boss Schematics

25 November, 2004

Well, it's thanksgiving day....and the Boss Dimension C resides on my 'mambo' Jameco breadboard. It is definitely not a dissapointment - I'm very glad to have 'discovered' the Dimension C, otherwise known as the DC-2. I have to hand it to those Roland/Boss engineers - they know what they're doing.

First of all, it's *quiet* - especially considering that there's not only one, but two BBD's running the show. That can be attributed in part to the use of compandors (NE570) and very good filter design. The filter design, as I understand it, also contributes to the sound of the thing.

A bit about the DC-2: it's a chorus device that uses two BBD's modulated anti-phase to each other by a common LFO. The effect of this is, for the most part, to eliminate the cyclical repetion found on most chorus devices. It's the 'little brother' to the legendary Roland Dimension D.

There are four modes of operation that range from a slight effect to a fairly obvious effect. It has mono and stereo operation, though I haven't explored explored the stereo function yet - I'm simply not set up for stereo right now! That's *gotta* change, though. From all accounts, this device really shines in the stereo mode. Having said that, the mono mode is not too shabby, either.

The four modes consist of different depths and speeds of modulation. In the original design, these were selected by four switches. There are four levels of modulation and three different speeds. I'll number the modulation levels from 1 to four, with 1 being the minimum modulation depth and 4 being the maximum modulation depth. I'll do the same with the modulation frequency or speed: 1 will be the slowest speed and 3 will be the highest speed. This is how the four modes work out:

Mode 1: Modulation Depth 1, Speed 1
Mode 2: Modulation Depth 4, Speed 1
Mode 3: Modulation Depth 2, Speed 2
Mode 4: Modulation Depth 3, Speed 3

I built the circuit using mainly the original parts, with a few exceptions.

The first exception is that the original bypass operates off of a BA634 flipflop, of which I could not find. Small matter, really, I think about any CMOS flip flop can be used in its place. What I did in the meantime was buffer 9V through an op amp, and run that through a 470R resistor to the anode of the 5.1V zener diode. I switch bypass in and out by moving the 9V reference to the op amp between 9V and ground. Works on the breadboard for now.

The second exception is that I used 1N5225B and 1N5231 zener diodes instead of the RD3.0EB1 and RD5.1EB3 zeners.

The third exception is that I used a dip version of the M5218 op amp.

And, finally, many of the caps I used were generally not the ceramic versions that I think were used in the original DC-2, and the MN3102 timing caps are polystyrene.

I'd finished breadboarding the circuit Monday night, but it absolutely was not working right - the problem was in the compressor - horrible oscillations occured if the input level was anywhere close to 50 mV. Mike Irwin, to my great gratitude, pointed me to the source of the problem - the schematic actually had a small, but vital, mistake - pins 10 and 9 of the unused section of the NE570 were jumpered together. I raced home Tuesday evening, removed the jumper, and Voila!, it worked to my satisfaction.

Following are some initial samples of the four modes of operation. Sadly, they're mono samples - I'm going to have to scare up some cables in order to make some stereo samples. In any event, these are actually using my Yamaha SHS-10 keyboards flute voice - my VCO's and EG's, which are not integrated into modules yet, were disconnected as the DC-2 is built on the 'master' breadboard that also serves as the power distribution and star ground hub of my breadboard system. I've been playing with the DC-2 so much, I haven't bothered to reconnect them!

The samples each start with the bypass on, so the sound is 'dry'. Then I supply 9V to the bypass system, which defeats the bypass, and the effect is heard in the second half of each sample. By 'dry', I mean going through the effect, but without the BBD's having an effect. According to Mike Irwin, the filtering of the circuit will still have an effect on the source of the sound, though, if that's the case, it sounds rather pleasing to me.

Yamaha SHS-10 Through the Dimension C, Mode 1 (691 KB)

Yamaha SHS-10 Through the Dimension C, Mode 2 (648 KB)

Yamaha SHS-10 Through the Dimension C, Mode 3 (636 KB)

Yamaha SHS-10 Through the Dimension C, Mode 4 (702 KB)

20 November, 2004

Well, it's been a year since I started the "Birth of a Synth" page. After gestating for a year, the synth has a total of 7 modules. But, it's been a fun year.

I'm preparing to begin work on the Dimension C clone, so before tearing down, I thought I'd record the MN3207 comb filter circuit "proper". Below is a multitracked sort of droney composition using the comb filter as the stage between the 2040 filter and PT2399 delay. Four tracks of PW/PWM VCO (two VCO's, one's pulse width is modulated) and one track of saw/triangle (one VCO saw, one VCO triangle) layered with it. The 2040 filter is controlled by a triangle LFO and the EG, the comb filter is modulated by a triangle LFO, the pulse width is modulated by yet another triangle (all supplied by the quad LFO module). The keyboard's modulation has an additional triangle LFO piped into it for a slight vibrato. The four 'pulse' tracks have high resonance and a slow modulation of the comb filter, while the single saw/triangle track has low resonance and a faster, narrower modulation of the comb filter. All tracks go through the sidekick amp where a bit of reverb is deposited on them.

BBD Circuit Composition (4.6 MB)

11 November, 2004

The work on the MN3207 comb filter module has continued, in spite of the fact that my lab assitant, Matthew, and I have both been struck down by strep these last couple of days.

On the Synth DIY list, discussion of BBD's, and the Boss DC-2 Dimension C, a gentleman by the name of Mike Irwin brought mention of Aron's DIY Stompbox Forum, which I've found to be a very, *very* good source of information. In fact, it lead me to a breakthrough in increasing the sweep ratio of the comb filter circuit. I've been spoiled by being the past owner of the legendary A/DA flanger (I loaned mine out years ago, and never saw it or the chap I loaned it to again). A major component of the A/DA's great sound was the sweep ratio, which was something like 42 to 1. Mike Irwin modified the A/DA circuit for use with the SAD1024, and it was from a schematic of this adaptation drawn by Stephen Giles that I found the solution for a very wide sweep ratio. Up until now I've tried everything from using VTL5C2's and VTL5C3's, transistors, and FET's as the variable resistive element in the MN3102's clock. None of these approaches proved to be satisfactory - in order to get a wide range of delay, problems would show up in the upper clock frequencies - jumpy sweeps, spontaneous oscillation, etc.

I've so far based the comb filter circuit loosely on the Morley Sapphire flanger, mainly because it had the circuitry already adapted to supplies and levels required by the MN3207. At least on the breadboard, however, the sweep ratio of the Sapphire Flanger was not anywhere close to what I wanted. The original circuit uses a CD4007 as the voltage controlled resistive element for a CD4047 multivibrator, driven from an LFO through a darlington pair for what I assume to be a logrithmic response on the sweep. This approach eschews the MN3102 clock chip altogether. The A/DA uses the same type of configuration, but the configuration of that circuit as opposed to the Sapphire circuit produces an incredibly wide sweep ratio. With it, I can clock the MN3207 from a delay range of a little under 500 uS to as high as I want, basically (far longer than I would ever want). I kept the the same supply circuitry (9V) and right now I am controlling it with an LFO from my synth. I also have not incorporated the buffer of the A/DA circuit (mainly because I do not have that particular chip) - but it seems to drive the MN3207 fine without it.

To get a hyper-triangular wave, I am processing a sine LFO with my peak/trough module, cutting the waveform and re-aligning the two halves to form the hyper-triangular waveform. Eventually, I would like to adopt Craig Anderton's technique of feeding the triangle wave of a VCLFO back to the expo control input to obtain the hypertriangular wave - this would be handy then for using the triangle for chorus and the hyper-triangular for flanging. That's a ways off yet.

The range of the circuit travels through doubling, chorusing and flanging delay ranges, so ultimately the module will be able to perform all of those functions. I might mention also that I added a negative regeneration circuit to the Sapphire, which didn't have that function to begin with.

I've made some samples of an exaggerated flanging sweep. The samples each start out with no regeneration, then end with maximum regeneration. The first two sample are of the noise output of my DSC2000 going through the circuit, one with positive regeneration and one with negative regeneration. The last two samples are of a single sawtooth VCO being processed by the circuit, the first with positive regeneration and the second with negative regeneration.

Noise With Positive Regeneration (1.05 MB)

Noise With Negative Regeneration (922 KB)

Sawtooth With Positive Regeneration (1.3 MB)

Sawtooth With Negative Regeneration (1.04 MB)

This is the link to Aron Nelson's stomp box page, home of the DIY Stompbox forum.

Aron Nelson's Stompbox Page

07 November, 2004

Here's a sample of the MN3207 based comb filter project. Hyper-triangular wave is controlling the clock, with the same EG that is used to control filter and VCA mixed in with it. Negative regeneration is used (set at max). 2 Sawtooth Vco's->LPF Filter->VCA->Comb Filter->PT2399 delay circuit. More on the comb filter later.

BBD Comb Filter Sample (3.5 MB)

29 October, 2004

Here is a sample of the PT2399 chip being used as a comb filter of sorts.

The minimum delay of the PT2399 is a bit long to do 'true' flanging, but there is a small window of range at the very extreme end of the delay adjustment to do this type of thing. As for chorus, I think the PT2399 is quite capable of it. Modulating longer delays ever so slightly can render very lush echo effects.

PT2399 Comb Filter Experiment (1.4 MB)

27 September, 2004

Last weekend I put together a Midwest Analog Platinum power supply. I mounted it into a military grade power supply chassis that my workplace was discarding. This weekend, I put together a distribution board for it, and Voila - I had power. And a rack of synth modules. I was thinking how close I was to having an actual synth, rather than a bunch of breadboards. And I am close, but the gestation is still ongoing - my rack has no expo VCO's or EG's yet. Guess I kinda work backwards =0).

Anyway, it occured to me that, though I have no expo VCO's or EG's, I do have sound sources and modifiers available. All of the LFO's in my quad LFO module will go up into the audio range, though they are not voltage controlled. The DSC2000 does have a linear VCO plus a noise source, and the discrete SSM2040 filter does self oscillate.

For timbre modifiers I do have the CGS Wave Multiplier/Grinder, the 2040 filter, a ring modulator and even VCA's that can be modulated.

So, I thought an interesting project would be to use *only* the completed modules to produce some sounds and see what I can come up with. Yesterday morning I recorded some tracks of various patches and mixed them together to make the following sample.

KS-01 Synth So Far Sample (4.7 MB)

22 August, 2004

Here's the sample of the XR2206 VCO.

At the time the sample was taken, the tuning had shifted from the initial tuning adjustment from when it was first breadboarded yesterday. I really didn't notice it, but it was pointed out by someone with ears of a higher precious metal than my own. Re-checking it against my Yamaha confirmed this.

So, I'll be putting in a tempco. Another factor of my breadboarded version is that I didn't have the right *type* of timing capacitor handy, so I put in a Mylar (not my first choice by far). That'll be changed out too, on the next sample. Finally, I think the tuning accuracy will be greatly enhanced if I use a multi-turn trim rather than a single turn (especially the one I have, which is vertically mounted and the #$%! thing pops out of the breadboard every time I try to adjust it).

Still, considering it's based on the largely-dismissed XR2206, it performs quite well, still hobbled by the above constraints.

Stay tuned (no pun intended).

XR2206 VCO sample (2.5 MB)

Following this is a sample using the Sample and Hold circuit of the Buchla Model 266 Source of Uncertainty, and the Low Pass Gate. Quite a bit of AD633 ring modulation, too. This one uses VCO3 (unrelated to the above XR2206 VCO).

I call it "Breakfast in Pangea". With apologies to Morton Subotnick.

266 S&H and LPG (5.8 MB)

25 July, 2004

Still working on a big 'what I've been doing' update.

In the meantime, here's a sample of the baby sequencer being sampled by the Buchla 266 S&H circuit, controlling two VCO's and a DSC2000. My first recording on the Korg D8.

Baby Sequencer/266 S&H (4.2 MB)

15 July, 2004

Has it been two months since I posted? Wow! A lot's happened, but that'll have to wait for another day.

I'm posting a sound sample that more or less arose out of a thread on the Synth DIY list. It concerns using a filter to process control voltages.

The filter controlling the CV's is an as yet unpublished design that I've had the pleasure to beta test for Thomas Henry, due to be published in a new DIY book he's authoring. The signal input is being fed a repeating gate signal from another TH design (keep yer eyes peeled for the book). The control voltage input is modulated by a positive going ramp wave from my new Quad LFO module (sigh, I haven't posted anything about *that* either). The output of the filter is modulating the Rene Schmitz/JH 2040 discrete filter, which is set for high resonance. That is the high tone you hear being modulated. The TH filter is also set for very high resonance, and does resonate to give a sine wave along with the slewed/not-slewed gate signal hitting the signal input. As the cutoff frequency increases, the pulsing gate signal 'takes over' more of the CV signal.

As a bonus, the signal input is two Rene Schmitz VCO3's going through a Buchla 194 BPF adaptation built for me by Jeff Pontius. Each output (there are four of them) of this filter is going to a separate VCA (my quad VCA module - new and unmentioned as of yet), then being mixed (my Mixer Model 3 module, again as yet unmentioned) and fed into the discrete 2040 filter. Signals from my quad LFO are modulating a couple of these VCA's. Towards the middle, I crank up the CV input of one of the VCA's, which is passing a particular band of audio and is being modulated by a triangle wave from one of the LFO's and sweep the frequency control of the LFO - the LFO is actually in the low audio range.

I'm controlling the VCO's with the Thomas Henry keyboard.

Filter Processing CV, 194 BPF (1.4 MB)

09 May, 2004

Again I've been very busy on some projects that I've long since wanted to try.

In the past few weeks, I've worked on the following projects:

A simple CD4017 based 10 step sequencer
The Buchla Low Pass Gate
The Korg PS3XXX series resonator
A 24 dB VCF based on the VCF in the Buchla Low Pass Gate

First of all, the Low Pass Gate: Other than the great 'ringing' one gets in the 'Both' mode, I was actually quite impressed with the low pass filter in it - it's a lot more pronounced than I had expected from a 'non-resonant' filter design. I guess what I'll have to try as far as Kitchen sinking goes is to put a resonance control on it - hmmm....haven't tried that yet.

What I did try was cascading just the low pass filter section with another, and adding resonance to that. That project is actually turning into quite a nice little filter. I added negative feedback, with a max gain of five or so, and have it so that it can resonate up into self-oscillation. For control, I borrowed the expo current sink from Rene Schmitz's VCO4069. Rene Schmitz meets Don Buchla!

The Korg resonator: I'm in love with this circuit. I built it pretty much from the original schematic, only I used VTL5C3/2's instead of the original LED/LDR's of the original circuit. Anyone who's familiar with the resonator from the original Korg PS3XXX series, or, for that matter, anyone who's got a MaM resonator or one of the Triple Resonant Filters available from MOTM or Cyndustries, knows what I'm talking about. It's an extremely versatile module that can give phase shifter-like effects, produce formants, or act as a sort of EQ if you like. I haven't started kitchen sinking it a whole lot yet, but I'm sure I'll be putting in the famous JH modifications to the original design.

The sequencer is a simple 4017 Johnson counter with the outputs summed through pots to a mixer.

Of note is the fact that I now have a SynthModules PSIM-1, and it comes in *very* handy for quantizing the sequencer - it erases the combined effects of tuning with trimpots and my tin ear when setting it up =0).

Below is a sample that combines a lot of what I've been doing the past few weeks - it uses the 4017 Sequencer, the resonator (quite liberally), and ends up with a brief sequence of the Low Pass Gate cascaded VCF. It can be quite a ballsy filter. Everything is clocked with the pulse output of my heavily kitchen-sinked Fluctuating Random Voltage circuit, and there's a brief bit of the flat noise output from the circuit run through the resonator as well. VCO's are Rene Schmitz VCO3's, and the VCA is Ray Wilson's, along with the Ray Wilson AR/EG.

Resonator and LPG VCF Sample (3.2 MB)

Below is a sample of the Low Pass Gate, mostly in the 'Both' mode. There are some long, sustained notes so that you can also get an idea of what the thing sounds like when it's not being 'knocked around' by short pulses, which seems to be the content of most LPG samples that are around. Of course, that is a nice effect, and I do knock it around a bit here. For that, the sequencer, which is controlling the VCO's, is also triggering the AR EG, which is set to 0 attack and 0 decay, and that is modulating the Low Pass Gate on the sequenced part. Any other VCA would just register clicks with this type of patch, but the Low Pass Gate 'rings', which gives a very nice, acoustic texture to the sound. On the sequence, two VCO3's are passed through the LPG, along with the Blacet DSC2000, which is triggered on the fifth step of each sequencer cycle. The DSC2000 pitch is controlled by the Fluctuating Random Voltage circuit. Towards the end is a collage of random...err....noise produced by running the VCO's through the LPG in various modes - the VCO's are controlled by the Fluctuating Random Voltage here, and the triggers are produced by the FRV as well for the main sequenced part and the noise part.

Low Pass Gate Sample (3.9 MB)

14 April, 2004

Here is a sample of the Shift Register circuit I worked with last month - it's in the 'random' mode, and being fed with an LFO to determine the high or low state when a clock pulse arrives. More on that later.

Shift Register Random (5.08 MB)

13 April, 2004

I've posted the schematics of the VCO Bank Pitched Noise source *as it was put on my breadboard*. It'll work as is shown in the schematic, but could be much better. It was a spur of the moment "I wonder what it would sound like if I did this" kind of thing. Being low on breadboard real-estate, I put it down exactly as in this schematic.

I think it's a good starting point for going in several directions - the CV control could quite obviously be improved, an exponential sink could be used, individual pots for the oscillator timing resistors could be used to 'tune' the noise, it may work a bit better if values were calculated so that the Vactrol resistors could be placed in series rather than parallel, VTL5C3's could be used instead of VTL5C2's, or Vactrols themselves could be discarded in lieu of some other control voltage scheme.

So, here it is, in two pages.

VCO Bank/Pitched Noise Schematic Page 1

VCO Bank/Pitched Noise Schematic Page 2

I pulled the oscillator bank, including the frequencies used, straight out of Thomas Henry's "Noise Generator Cookbook". This book has proven to be of great value to me - I adapted a different circuit out of it for use in the modified Source of Uncertainty module I'm working on, and there's other great circuits in it, as well as the most easily understandable explanation of what noise *is* and how pink noise differs from white noise (and why)......highly recommended by me, that's for sure! Here's a link to Midwest Analog:

Midwest Analog

11 April, 2004

In the time since the last post, I've been busy breadboarding and playing with a few ideas and things I've wanted to do. Since the last post I've:

Breadboarded the beginnings of a Breath Controller

Breadboarded a modified version Don Buchla's Random Fluctuating Voltage circuit from the 266 Source of Uncertainty

Breadboarded Don Buchla's Voltage Controlled Integrator

Breadboarded a modified version of Thomas Henry's Oscillator Bank pitched noise circuit.

I've also devoted a lot of time researching all the great stuff that Don Buchla put into various modules. I've seen on the lists that Don may be making a comeback with a new synth(?) - if that's true, I can't wait to see/hear the thing!

Details/schematics/explanations will follow, but tonight I've spent quite a long time uploading the following samples, and I've run out of time.....,

So, here are the samples.

This first sample is a portion of one that I mentioned when I first posted about the modified Oscillator Bank Pitched Noise source. This sample has no human interaction, other than the fade in and fade out. The oscillators are under control of the smooth output of the Random Fluctuating Voltage (RFV) circuit. The EG is being triggered by the RFV clock, and the filter cutoff is also controlled by the smooth voltage, stepped voltage, and EG signal. The VCA is controlled by the EG signal. The only tone source used is the Pitched Noise oscillator bank.

Pitched Noise Oscillator Bank/RFV 1 (2.1 MB)

The following sample is a collage of three separate samples using the same oscillator bank/pitched noise source. This sample has the 'gong' sound that I described on the Synth DIY list as well as some voltage controlled gliding controlled by a triangle LFO while the filter and EG are again modulated by various outputs of the RFV. Once again, the pitched noise is the only audio source used.

Bang the Gong (3.35 MB)

Next is the 'cat' sample. This sample uses a sawtooth VCO, the square output of a DSC2000, and the pitched noise source/oscillator bank. Again, it's all controlled by the RFV circuit. The EG signal that is modulating the filter cutoff is being processed by the Voltage Controlled Integrator, which is controlled by the smooth output of the RFV. The caterwauling in the background is the pitched noise source modulated by the smooth output of the RFV. I made this and sent it to my friend's cat =-D.

Buchla For Kitty Cats (582 kB)

And finally, this is a sample of the RFV doing what it does best - it's controlling a VCO, the DSC2000, and the 2040 filter. There's quite a bit of overdubbing in this sample - check it out if you like 'weird' =0). The creepy intro timing is all determined by the RFV - the EG is controlled by the trigger out of the module, and the smooth output is controlling the 'probability of change', which controls the clock frequency and smooth lag time.

Landing on a Different Planet in the 1950's (4.7 MB)

15 March 2004

I've been busy the past few weeks trying a few different things and working some stuff out on paper. My latest dalliance concerns the 4006 shift register.

The circuit uses a comparator to determine the input of the shift register by comparing a signal input (usually noise, but can be any input, really) with a reference signal level. This outputs either a 1 or a 0 to the shift register input. Since the input signal is not synchronized to the clock which is driving the shift register, the data bit that is loaded into the shift register at each clock pulse is more or less random.

I have the for D+4 outputs of the shift register buffered and going through pots to a mixer, with about a 10 to 1 weighting on each input. I then use the pots to set a different level for each bit that is coming out of the shift register. This produces a voltage pattern that can can consist of as many as 16 different 'notes' if each bit is set to a different voltage level going into the mixer. It's much like a random sequencer where the range of notes is set by the trim pots, and the pattern is set at random. Additionally, I can 'capture' a random pattern by inputting the end of the shift register back to the beginning.

The pattern can be influenced by the comparator reference level, a sort of 'probability control' - the lower the reference level, the more likely the comparator will be high when the clock pulse arrives when using noise as the input. I've found that using an LFO for the input can create interesting random patterns depending on the rate of the LFO versus the rate of the clock.

Any of the bit outputs can be used to gate an EG, which gives a whole new perspective on the generated pattern. If Bit 1+4 is clocking the EG, one 'snapshot' of the pattern is taken, a different view than say if bit 2+4 were gating the EG. What will be *really* interesting is to have two or more EG's firing different voices that are controlled by the mixed bit pattern - cool stereo effects are in the offing, I believe.

Finally, the two 'extra' bits, 2+5 and 4+5 add two more bits to play with, which would raise the theoretical number of steps to 64, and provide two additional gate options as well. I have found, though, that the more bits you try to mix, the complexity of mixing the bits to a CV pattern becomes exponentially more challenging.

Setting it up to perform a 'sequence' is very different than a step sequencer - it's like molding clay more than anything, because each pot not only affects the level of its bit, it also influences the mixed output level of the other bits when it's high at the same time. It can, however, be a very creative process.

Below is a sample that was composed using the 4006 shift register, other than the pad sounds and a bit of hand tweaking on filters and such, everything else was produced as a pattern by the 4006 shift register circuit. There's also a little cool noise that I got by clocking the 4006 at an audio rate with a random input. The sample is a bit over 5 MB, but I urge you to listen to it - it's amazing what a little old shift register can do.......

Shift Register Magnum Opus (5.1 MB)

22 February, 2004

A quick post here - I've been back to playing with Vactrols this week. In addition to that, one evening I multitracked a piece using a fairly cool patch.

The patch is really nice as it takes advantage of the Wave Multiplier, Wave Folder, and the MS-20 filter to get a lot more harmonic variation from two VCO's. The MS-20 filter is in high pass mode, and in that mode it's capable of delivering the most pleasing lower intervals. This sample is also the first *stereo* sample to appear on this web site (welcome to the 20th - er 21st century, Scott =0).

I assembled it in the freeware program Audacity. I didn't use any of the built-in effects - this is all KS-01 synth.

Give it a listen, it's a bit different than the other motley samples on this site!

Wave Mult, Folder and 2 VCO's in Living Stereo (1.56 MB)

Here's an illustration of the patch:

Illustration of Patch

15 February, 2004

I'm pretty excited over a new Ken Stone circuit - his 'simple' Wave Folder. It's an incredibly small circuit, given it's sonic punch. It takes only a few minutes to breadboard and probably an eon or two to explore =0). I added exactly one part - a pot set up as a voltage divider with the output summed to the same point that the audio in pot is summed to. With this I inject CV sources to act as an offset voltage, much like Ken's Wave Multiplier. You can also run a VCA before the audio input to control the timbre, which I haven't even played with yet.

Anyway, here's the link to Ken's site - look for 'Simple Wave Folder' and Buy the Board!!

Ken Stone CGS

And here's a sample of what it can sound like (just a few selected timbres). The patch uses a single triangle wave into the input of the Wave Folder, with the envelope generator providing the offset. This passes through the 2040 filter (with resonance set to 0), on through the VCA on out through the PT2399 echo circuit. The same EG is controlling the VCF and the VCA. I have an LFO going through a Motohiko Takeda exponential VCA into the keyboard modulation input. The VCA is controlled with an AR EG for delayed vibrato. For the different timbres, I only adjusted the audio and CV input levels to the Wave Folder and the ADSR controls. I didn't alter any of the filter settings.

CGS 'Simple' Wave Folder (1.8 MB)

Here's a sample I did a few days ago of Rene Schmitz's MS-20 filter in the band pass mode - it's a heck of a filter - I love the resonance of this filter!

Rene Schmitz MS-20 Filter in BP Mode (459 KB)

07 February, 2004

Still suffering a bit from jet lag induced by the last business trip, so I haven't gotten much constrution done since I got back. Today I relaxed with the synth and breadboards and played with the grinder section of the Wave Multiplier, and the other usual toys. I've got a couple of samples here that I find interesting.

'Grinder and More 1' uses one VCO, the 2040 filter, PT2399 delay, two LFO's, the Farm ADSR, a VCA, the Wave Multiplier (of course) and the Blacet DSC2000. I have the 'Eerie Mode' on for this one....=0)

'Grinder and More 2' uses the same setup as above (except the DSC2000 is not used), with the addition of Rene's Late MS-20 Filter in bandpass mode, fed into the 2040 filter. The interesting thing about this sample is the interaction of the VCO and the resonance of the MS-20 resonance - for given pitches, it sounds like two simultaneous notes are played, though my synth is quite monophonic.

Check the samples out - IMO, they're good examples of the textures that the grinder section of the Wave Multiplier can provide when processed by the multiplier itself and utilized with other modules.

Grinder and More 1 (1.76 MB)

Grinder and More 2 (289 KB)

05 February, 2004

It's been a while since the last post - another business trip threw a spanner in the works, so to speak. However, on this trip, I was priviledged to actually meet Rene Schmitz in person in Germany - he was kind enough to have me over one evening, and it was quite an experience. Rene has got a very cool setup, and I learned a lot from him, as always.

Right before I left, I finished Rene's version of the 2040 filter. It's got three signal inputs (multipled so that the signals can be 'passed on'), one V/Oct input and two CV inputs with attenuators (again, mult'ed), a low pass output (multed twice), an output for the three input mixer (multed twice), and an inverted output for the three input mixer (again, multed twice). It's a sweet module indeed....

Below is a picture of the four modules I have completed. There's a drilled panel in there as well, ready to be turned into a module. In four U of space I have the 2040 filter (with three input sub-mixer), a Blacet DSC2000 (fully loaded), a CGS Wave Multiplier (again fully loaded with the grinder as well), and a 45 point multiple. A point about the Wave Multiplier - I've seen that some choose to omit this feature, but I find it to be a pretty useful timbre altering device, particularly when used with the multiplier section of the module itself. Anyway, I've chicken-scratched the modules with legends until I do the front panels proper - it was turning into a necessity - there's already a lot of connectors and controls to keep straight in the head =-)

4 completed modules

18 January, 2004

It's been a long time since my last post here - I've actually been gone for awhile on a business trip. Well, I'm back and regrouping. This weekend I breadboarded the ADSR EG I'm going to be using - Motohiko Takeda's Farm Synth EG. A while back I traded some parts for some PCB's of Motohiko's work. The EG board contains two ADSR's and two exponential VCA's - a very useful module IMO. I'm modifiying it for re-triggering, since my keyboard doesn't put out a gate if another note is held down, but it does always put out a trigger (this arrangement is actually quite useful for playing legato, if you only use the gate). Also, I'll put an inverted output on it, a switch to select different cycle ranges, and I am considering putting an end-of-cycle pulse on it as well; I just haven't sat down to figure out what in's and out's I'll be able to work into the front panel.

I've created some samples to put here - they exploit the CGS Wave Multiplier. In all of the samples, only a single triangle wave is used.

In the first sample, the very first section uses the Rene Schmitz Late MS-20 filter in the band pass mode, and the Ray Wilson AR EG - the triangle wave goes through the filter and then through the Wave multiplier - the sound is basically the triangle with a lot of MS-20 resonance being processed by the Wave Multiplier. The rest of the sample uses the Motohiko Takeda ADSR and the 2040 filter. This starts out with the sound slightly reminiscent of a Buddhist chant, then moves through a sitar-like sound, then an electric guitar type sound. I'm modulating the Wave Multiplier with complementary signals from the ADSR, which always tends to yield very interesting sounds.

The second sample is the MS-20 filter again, in the band pass mode, with the resonance cranked up and feeding the Wave Multiplier.

The third sample uses the triangle through the Wave Multiplier, then through the 2040 filter. It's a demonstration of the milder, less modulated characteristic of the Wave Multiplier.

In the first three samples, the folds ouput of the Wave Multiplier is used. In the fourth sample, the pulse output is used. The PWM is modulated excessively, so that the sound cuts in and out, producing a nice effect with the 2040 filter.

Wave Mult Sample 1 (700KB)

Wave Mult Sample 2 (1.3 MB)

Wave Mult Sample 3 (800 KB)

Wave Mult Sample 4 (494 KB)

30 December, 2003

What a fun time I've had the past few days. I finished module number 3, the Blacet DSC2000. It takes a bit of a different format than my other modules, mainly because I wanted to keep the original pot mounting. In order to do that, I had to offset the pots a bit. I also compromised and didn't put attenuators on it for the attack and decay CV's - just wasn't enough room to do it comfortably in a 1U module. I still managed to get attenuators for the Noise Clock, Noise Filter, VCO Frequency and PWM CV's, though.

The really exciting part of the past few days has been the breadboarding of Rene Schmitz's excellent version of the 2040 filter. It's such a sweet filter - I love it. It's been high time I started messing with four pole filters, and this seemed like such a nice choice. I still have yet to do a transistor ladder filter, but that time's coming.

The other fun thing was my trip down memory lane - when I got back into Synth DIY, I went off on a tangent with the PT2399 delay chip. Then, my attention turned to getting some of the basics down (like, a VCO, for one thing), and got away from it. I pulled out a protoboard I did of the PT2399 echo core back then, and improvised a mixer for regeneration, and instantly remembered why I'm so crazy about the chip. I've posted some samples of it, with the 2040 filter. It's "warts and all" - in the final version, I plan to do a bit better job of filtering, and also to compand it much like JH does his Storm Tide Flanger, using the limiting diodes on the regeneration. Also, I just rigged one mixer to play around with - actually there should be two, one to set the balance between wet and dry signals, which I haven't done on these samples.

Each sample, except the DSC2000 sample, uses a simple VCO3 sawtooth through the 2040 filter, then through the delay. The first two samples have an unusual patch - the delay comes before the VCA. On the rest, the delay is after the VCA.

My original design for the delay uses a FET to control the delay time. I tried putting the expo sink from Rene's VCO4069 on this version, which works well, though I can't seem to pull it to absolute minimum delay - I still have to work that out. But, it modulates the delay quite well. The DSC2000 samples use the VC control of the delay so much, you'll probably be sick of it before you're done listening to the samples =0). In the first DSC2000 sample, the delay is being modulated by the random noise out of the DSC2000. In the second sample, a square wave LFO is pulsing the delay time, and the sound source is the VCO out of the DSC2000 - it's actually being controlled by the keyboard!

2040 Filter With PT2399 Sample 1 (350 KB)

2040 Filter With PT2399 Sample 2 (443 KB)

2040 Filter With PT2399 Sample 3 (1 MB)

2040 Filter With PT2399 Sample 4 (862 KB)

2040 Filter With PT2399/DSC2000 Sample 1(1.08 MB)

2040 Filter With PT2399/DSC2000 Sample 2 (1.2 MB)


21 December, 2003

I accomplished a lot these past few days. A friend sent me some MOTM rack rails, and I was able to use a Schaeffer panel to make a template for drilling the mounting holes of the modules. I built the 'Mother of All Multiples' (second completed KS-01 module, except for panel legend), and I put in some work on the 291 VCF filter board, and tested the core, input mixer, and LP/HP buffers succesfully.

Here's a pic of the KS-01 Mass Manufacturing Center, located in the industrial heart of Kansas.

KS-01 Manufacturing Center

Below are a couple of the usual blurry pics of the KS-01 multiple. Briefly, it consists of 45 connectors arrayed in three columns of 15. Each column is divided up into 5 common multiples, with the lower two multiples connected to the above multiple with a switching jack. A more detailed explanation of it can be found in the Modules section of this site.

1 U Multiple

Below are a couple 'real' pictures of the panel work I've been doing. The pictures show the front and back of the rack, which has the front panel and controls for a VCO3 module, the completed Wave Multiplier module, and the rest are blank panels with mounting holes drilled.

Front Panel

Rear of Front Panel

17 December, 2003

A red letter day - finished the first module, the Ken Stone CGS Wave Multiplier. I've had the PCB done and playing with it for nearly a year, and I thought it was high time it was turned into a module. I still need to put a legend on it, and drill the mounting holes, but it's an operational module, the first for the KS-01!

Here's a sample of the module showing just a tiny slice of the things this wonderful Ken Stone design can do. The Wave Multiplier can be sweet and it can be very nasty, this sample kind of, sort of, covers both ends of the spectrum. Both parts of the sample use a triangle wave as the input. In the first part, there is no reverb. The folds output is used and the 291 Low Pass output is used (unmodulated). The offset and folds CV are modulated with the EG.

In the second part of the sample, the pulse output of the module is used, the PWM and Folds are modulated with two different LFO's. The 291 is set for max slope with some resonance, and is controlled by the EG.

CGS Wave Multiplier Sample (605 KB)

Ken Stone Wave Multiplier Module

Wave Multiplier

15 December, 2003

Home sick today, didn't get much done besides this sample of the VCO3 suboscillator reset function.

The sample is of sawtooth of the VCO3 mixed with either the divide by two or divide by four signal of its suboscillator. The suboscillator is being reset by the pulse of a second VCO3.

VCO3 Suboscillator Reset Sample (710 KB)

14 December, 2003

Worked a bit on the 291 VCF PCB. Put in the IC sockets and power distribution jumpers. I also put in the bypass and feedback caps, and the resistors and Vactrols for the upper left quadrant of the board.


Here's another sample of the AD633 Ring Modulator. Again, two triangle waves going through the 291 VCF. I've been playing with sawtooths, which is way cool, but I didn't make any samples. By law, people like me shouldn't have ring modulators. =-D

Ring Modulator Sample 2 (483 KB)

13 December, 2003

Today was a nice, snowy day to get some DIY done.

I've been thinking about it a while, and I decided to rearrange VCO3 one more time. I'm taking out the multipled waves and adding three more inputs:

1. AC input for Lin1.

2. External input for the suboscillator.

3. Sync input for suboscillator.

I breadboarded a simple means to reset the suboscillator 4013, and it makes a fairly decent sync. I think it will also work as a way to sync the oscillator to the keyboard in LFO mode. I tried AC coupling the reset to get a short pulse for the reset, but DC coupling gave better results, to my ears.

The suboscillator input will override the internal VCO input, and will at that point operate as an individual function separate from the VCO. With three VCO3's, I'll be able to daisy chain the suboscillators to get as high a divide count as 64, if I so choose. Since the signal is first going to the comparator, it shouldn't require a pulse wave to drive it.

I also breadboarded a simple version of the AD633 Ring Modulator - just eight resistors, two trimpots, and the AD633 make up the circuit. I'm pretty impressed with the quality. Below is a sample of it in action with two triangle waves, going through the 291 Core VCF.

AD633 Ring Modulator Sample 1 (486 KB)

11 December, 2003

Another benchmark day - I drilled the first module panel, and successfully! It's a VCO3 panel. I put bullseye circles on the template, and that helped a *lot* in getting the small connectors straight. Below are some pics of it.

VCO3 Panel

Look, Ma, I'm Morton Subotnick!

10 December, 2003

Finished the layout for the triple ring modulator panel.

I also performed an experiment with a scrap piece of aluminum to confirm if I had the right drilling technique and if my format was actually usable in real life. It was pretty much a success. The jacks weren't lined up perfectly straight, but if the results I got tonight are the worse it gets, I'll be pretty happy - in fact, on the overall synth, I'd say it would be unnoticeable. I had hastily tacked down the image, the middle section of a VCO3 panel, and hadn't gotten it perfectly flat. I think I could also improve the accuracy by putting a small bulls-eye on the crosshair in the illustration - that'll aid in telling if the drill is aligned perfectly on the crosshair when drilling the pilot holes (a laser sight would be nice!). All in all, I'm very happy with the results, and the next step is to drill an actual VCO3 panel.

As for the format, it works really well. On the sample panel, I have no problems plugging and unplugging patch cords, and I can tweak the knobs easily.

Panel Drilling and Format Test

09 December, 2003

Today I made tentative plans for the KS-01 Triple Ring Modulator. I drew up a panel for it, too, which I will be posting shortly.

The module will contain three different variations of ring modulator design. One will be Rene Schmitz's MC1496 Based YARM Ring Modulator, another will be an AD633 ring modulator, based on Roman Sowa's design, and the third will be a 'digital' Korg MS-20 type ring modulator. This last ring modulator will have comparators on the inputs to square up the input waveforms, so that any wave can conveniently be injected into it.

In addition to that, the module will have two 566 based V/Oct VCO's built into it. Each VCO will have a tune control, and a modulation scalar/inverter and will have sine, triangle, and pulse outputs. These VCO's will be available for use with the ring modulator, as well as for other tasks such as control with joystick, sample and hold, and sequencer. Each ring modulator will have signal normalled to each of the inputs, The AD633 and MC1496 each will have a sine normalized to each input, and the MS-20 will have a pulse normalized to it. The VCO's will have a common unscaled input that can control both of them. The design can be found in Thomas Henry's "Making Music with the 566", available at Midwest Analog (see links section). The AD633 and 1496 Ring Modulators will have DC or AC inputs.

In other news, tonight I experimented with placing a drilling template over a blank panel. Actually the illustration I used lacks the crosshairs on the drill points, but it was a good excercise before I get started. I think my panel illustration technique will be something like this as well.

Panel Drill Hole Alignment

07 December, 2003

Got quite a bit done this weekend. I worked out the functions on my VCO3 daughter board, tested them and put them to schematic, which I've posted to this site. Received the tools I need to start working on panels from my brother.

03 December, 2003

Finished the fourth and final 'quadrant' of the PCB layout so that Jeff and I can start putting the 291 core VCF to the board. I also adjusted the placement of controls and connectors on the VCO3 layout to ensure that I have 5/8" clearance for mounting to the rail. Also did an experimental layout of VCO3 where I drop the AC/DC switch, add a soft sync pulse input level pot, and change the multiple inputs to the EXP CV input and Linear CV2 input to switched AC inputs. That way, connecting to a connector will provide coupling, but connecting to the other connector will provide DC coupling. They will be normalled for the AC input, and plugging to a DC input will override the AC input.

Last night I added the 'Soft Sync' page to the site, and tonight I added some sound samples to illuminate the function.

01 December, 2003

Finished a bit more of the 291 layout. Also nailed down the soft sync a bit more. Here are a couple of samples of two VCO3's, soft synced, with pulse waves and divide by two pulse out of one of them, through the 291 and then through a BBD delay circuit that I jury-rigged to an optocoupler in order to move the delay frequency around a bit.

2 VCO3's through 291 LP and Short Delay 1 (380 KB)

2 VCO3's through 291 LP and Short Delay 2 (518 KB)

30 November, 2003

Worked a bit on dividing out my drawing of the 291 PCB into quadrants in preparation for the build - putting the drawings into quadrants makes it easier to blow up and see the connections. It's going to go on protoboard that I got from Here's a pic of the board that I'm using:

Futurlec Proto Board

In a bit, I'm going to post the full schematics to this filter adaptation - they aren't the final version (I need to clean up a few things), but they're usable now.

I recorded a couple of samples of two VCO3's, soft-synced together through the 291 filter, using the Low Pass output of the filter. The signal is made up of the modulated pulse from one VCO3, and the sawtooth and divide by 2 pulse of the other VCO3. On both of the samples, the filter is modulated by the AR EG and a slow triangle LFO (the slow triangle is also modulating the pulse of the 'synced' VCO). On the second one, an additional pulse is modulating the FM input to the filter.

Two VCO3's through 291 Low Pass Sample 1 (295 KB)

Two VCO3's through 291 Low Pass Sample 2 (327 KB)

29 November, 2003

I bought a step drill bit today for drilling out front panels. My brother and I tried it out on the drill press today. It works pretty well for the pots, but the right size for the 3.5mm connectors is just in-between steps. A circular file or reamer puts it to the right size. My brother is going to get for me a reamer at just the right size that I can mount on the drill press.

I'll start with 1/8" pilot holes, then use the step drill to get the holes to the right size. For the 3.5mm connectors, I'll have to add one more step to get the holes the right size.

Below is a pic of the scrap aluminum we tried things out on. It's got a couple of jacks and a pot with a knob mounted on it.

Trial Drill

27 November, 2003

Thanksgiving Day, a holiday traditionally spent with family. Still, today I thought out a good way to make a simple 5 octave quantizer. So many projects, just one short lifetime =0).

Added the Dual Mixer 1 schematics to the schematics page.

26 November, 2003

I thought a bit more about the soft sync today - probably because of the values of components I had chosen, there seemed to be more detuning than I cared for, and it was hard to ascertain if I really was soft synced or not. One problem is that this is my first experience with soft sync, so I'm kind of working in the dark - I understand what it is, but how it sounds has always been a mystery to me until this experimentation.

I chose a different, rather more convoluted tack. I used a TL072 dual op amp to breadboard an inverting mixer and a comparator. I used a 2N3904 transistor, 2.2K resistor, and 100pF cap and pretty much duplicated Rene's input to the hard sync, only in this case, I ran the output of the transistor's collector to the input of the inverting mixer through a 100K pot. I connected the output of the mixer to the comparator.

On the other input of the mixer, I injected the triangle wave of the VCO that's being synced. To set the limit to the comparator, I put voltage through a trimpot and adjusted it so that when the triangle wave only was going through the mixer, the comparator was not triggered. Then I applied the pulse wave from the VCO that was supplying the sync signal.

In this manner, the comparator will only put out a pulse when the triangle and the pulse are fairly close together rather than constantly, which you would want with hard sync. That means that the two oscillators must be close together in frequency for the soft sync to 'take hold'.
Adjusting the level of the pulse can increase this 'capture range' since the triangle can be farther away from its peak when mixed with the input pulse to trip the comparator. However, I wasn't able to get a hard sync response by really cranking up the level of the pulse (not that I care, I just thought that would happen - could maybe have to do with the comparator limit I've chosen)

In any event - it works, and I now know what soft sync is like. Basically you can 'lock' the frequencies together, and offseting the frequencies of the oscillators slightly gives a much different effect than if they were unsynched, yet is much different than if you did the same thing with hard sync. This configuration seems to be a lot easier to control and set up the soft sync.

Below is a quick sample of a string pad sort of sound I recorded tonight. It's the sawtooth output from the two VCO's that are soft synced together, with a tad of vibrato on both, through the 291 Core VCF, with the BPF output and very wide bandwidth. Wish I had an ensemble chorus :-\.

Soft Sync Sample 2 (511 KB)

25 November, 2003

Experimented a bit more with the soft sync circuit of VCO3. I found that 470n works as a good value for the input cap. I've pretty much decided to drop the AC coupling switch and add a level control for the soft sync. Feeding the soft sync with a narrow pulse gives the best 'standard' soft sync results - throughout a narrow range when the VCO's are fairly close together in frequency, the lock together for a nice chorusy, ringing sound. But I also find that feeding it a triangle wave or some other waveform can also give some more effects-based sounds. The level control helps to 'lock in' the sync signal to a certain threshold, and driving it with wider pulses can derive some effects that are similiar to, yet unlike hard sync.

24 November, 2003

Not much happening today. I recorded a sample of the soft sync in action. It's two sawtooths, slightly detuned. It starts out with no soft sync. 18 seconds into the sample, I soft sync the second VCO3. Around 33 seconds into it, I remove the alligator clip (thumping sound) so that they're not synced, then clip it back on (another thumping sound) to finish up soft synced. The soft sync signal is a triangle wave from the 'master' to the 'slave' oscillator.

22 November 2003

Today was a good day - I got some things done that I've been intending to do for quite some time. I breadboarded a couple of additions that I want to put into the VCO3 modules - a suboscillator and soft sync. I'd mentioned some convoluted patch for achieving soft sync a while back, and Rene had pointed out that it really would only require adding a capacitor and a resistor to the existing VCO3 =-D.

The suboscillator uses both sections of a 4013 to get divide by 2 and divide by four outputs. Here's a sample I took after I'd finished the divide by two part - it's a single VCO3 sawtooth (IIRC) and divide by two suboscillator run through the VTL5C2 Phase Shifter, through the 291 Core VCF and on out to the VCA.

Suboscillator Sample 1 (483 KB)

Here's a sample of a single VCO3 that has both suboscillators mixed with the sawtooth, through the 291 VCF.

Suboscillator Sample 2 (331 KB)

These next four samples are examples of modulation with the suboscillators, with one VCO3 and the 291 VCF being modulated by the pulse and suboscillators of another VCO3.

Suboscillator Sample 3 (511 KB)

Suboscillator Sample 4 (405 KB)

Suboscillator Sample 5 (522 KB)

Suboscillator Sample 6 (465 KB)

21 November 2003

Today I worked out the faceplate arrangement for a quint AR EG, Quad VCA, and a Quad LFO. Also worked on two different dual mixers and tweaked up the CGS Wave Multiplier layout a bit more.

Started a new 'Modules' section of the site today to illustrate the modules and layout I intend to use.

Below is a physical layout of the knobs on the two dual mixers.

Dual Mixer 1 and Dual Mixer 2

20 November 2003

Worked a bit more on the 2U 291 Core VCF faceplate layout. Instead of the multiples mentioned above, the input multiple I'm going to put on it is for the CF CV. On the outputs, I'll have three jacks per output.

A thought occurred to me that I could easily provide inputs for the internal Inverted Bandpass/Mixed audio crossfader. This will provide a means to use this function in conjunction with the filter, or as stand-alone. Because of the nature of the design (which I'm not going to change), this crossfader will invert the outputs, which actually might be useful for certain applications. The inputs are normalled jacks, so that with no connections, the circuit works in the 'normal' mode. I'm posting an illustration of my faceplate design and adding some text to explain the variations on the 291 Adaptation page.

Also tweaked up my VCO3 faceplate design today. Below is a fuzzy pic of it with knobs and jacks overlayed on it. The blank spot is for a switch which will select between AC and DC coupling on one of the linear modulation inputs.

VCO3 Panel

19 November 2003

I worked a bit on module layout today. I thought for a while that I could get away with a 1U 291 core VCF, but things got too tight to put all of the switches and pots in. A 2U design leaves too much dead space for my taste, so I decided to add the following:

1. A third audio input.

2. Leave the original level control from the Buchla design as a master level control.

3. Add multiples to the Center Frequency, FM, CF/Both/BW, and Bandwidth CV inputs.

4. Add multiples to the BP, LP, HP and InvBP/Mixed Audio outputs.

This will add some versatility and will only require the addition of one resistor to the PCB.

Below is a picture of the arrangement of pots and jacks on a 2U panel for the 291 Core VCF. The blank spot is where the three switches will go.

291 Core VCF Layout

I also checked the layout of my 1U CGS Wave Multiplier. Here is a pic of it, with the knobs and jacks layed over a scale drawing of it.

CGS WM Layout

18 November, 2003

Start of this log. Finished boards include 1 Rene Schmitz VCO3, CGS Wave Multiplier, Keyboard Controller, PT2399 Proto Circuit, Blacet Dark Star/Chaos 2000. Boards in the works: two more VCO3's.

Today, my friend Dan McKay stamped out 18 1U panels, and 4 2U panels out of sheet aluminum - the first real wobbly steps of my 'permanent' synthesizer. Other projects under way: putting 291 Core VCF on proto-board