Random*Source 4U Serge Project

I built this large Serge Modular system for a friend, Kri Samadhi. Serge is one of my favorite modular synthesizers. So when Kri started asking me for advice on what Serge to get, it eventually came down to Random*Source kits. I’ve built a lot of Ken Stone’s CGS Serge circuits, which are licensed from Serge. However, Random*Source upped the game recently when Serge Tcherepnin, the inventor of the legendary Serge Modular synthesizer system, began working with them to update his circuits.

Panels and Modules

Random*Source 4×4 Serge panels are 4U in height (7 inches) by 4.25 inches wide. Each panel generally houses two or more modules. Here are the dimensions.

Panel Dimensions
Panels and Boards ready for assembly

List of panels

  • 2 Dual Slope Generator MK2; CV Processor
  • 1 Dual Slope Generator XL; Peak & Trough
  • 2 New Timbral Oscillator (NTO); XFader
  • 1 Dual Low Pass Gate; Timbre; Stereo Panning Mixer
  • 1 Sequencer/Programmer 8
  • 1 Smooth & Stepped Generator (SSG); Random Source
  • 1 Pulse Divider; Boolean Logic; %N Comparator
  • 1 Triple Wave Shaper+; New Ring Modulator
  • 1 Wave Multipliers; Resonant Equalizer
  • 1 Variable Slope Filter (VCFS); Variable Resonance Filter (VCFQ)

Parts Sourcing

The kits come with circuit boards and a panel. All but the Dual Low Pass Gates board were fully populated with SMT components. The builder has to source the remaining parts, including pots, knobs, switches, trim pots, banana jacks, LEDs, and some other parts such as capacitors not surface mounted. Just making a Bill of Materials for this 12-panel system was quite a bit of work, and I made a few mistakes requiring re-orders as I discovered them. The BOM supplied by Random*Source often lacked specific part numbers. I spent a lot of time figuring out which super-bright LEDs to use. I tested them to determine the needed resistor (instead of using trim pots to adjust the brightness). Moreover, a few of the build documents and BOMs were out of date and referred to earlier versions of the boards. I had to carefully examine each board to figure out what it actually needed. I got the pots from Thonk, knobs from SynthCube, Vactrols from Modular Addict, and the remaining parts from Mouser.

Color Choices

Having to source parts lets you decide on jack and knob colors. I found the black, white (and sometimes blue) color scheme for jacks used by Random*Source (and the original Serge) to be confusing, because the same colors are used for both input and output, with White indicating generally uni-polar voltages and Black bi-polar. (And this isn’t strict, since for example the DSG input is Black, but the main output is White, yet the output can be bi-polar if the input signal is.) I decided to use different colors for inputs and outputs. I came up with a transformation from the R*S color to my selected color.

White inputs remained White, unless they are CV inputs, which I made Blue. Black inputs remained Black. White outputs became Green. Black outputs became Gray. Red pulse outputs remained Red. Gate and trigger inputs became Yellow. I kept the traditional Violet for the NTO Sync inputs and picked Orange for the ‘Hot Coupler’ output on the SSG.

Knob color is Black, but for CV attenuators I used Blue. I spiced it up a little by using Red knobs for bi-polar functions such as fading, panning, and biasing.

Power Supply

I built a DIY case for a Power One 1.7 Amp supply (HBB15-1.5-AG Bel Power Solutions). It is rated at 1.7A of plus and minus 12 volts. Even though R*S gives power consumption specs on only some modules, I figured it would be enough. It turned out just right, since the +12 Volt current consumption of this system measures 1.17 amps, about 70% of rated load. However, the supply gets very warm.

Power

I used remote sensing for this project. An eight-conductor cable runs to the case with only one connector (on the case) in the path. All other connections are soldered. I put a cable gland strain relief on the supply box to secure the wires. The power box also has banana jacks to access the outputs, plus LED indicators. The conductors: +12V, 0V, -12V, Earth, and two sense pairs. Earth is connected to 0V at the distribution boards in the case.

Wood Case

Serge Modular Case

Kri and I both worked on the case design and he built it. The distribution boards came from Oakley Modular. The power inlet, plus two TRS jacks for the optional balanced outputs of the two XFader modules, fit on a 6hp Eurorack blank panel. Here’s the case design. (It turned out just slightly different.)

Quirks

A few small quirks may be worth mentioning.

The push button switches (part number given by Random*Source) are too high behind the panel. On the Sequencer I had to add washers behind the pots to extend them up to reach the PC board.

I found high frequency parasitic oscillators on the Gates outputs in filter mode. This was easily fixed with 5pf capacitors across some op amps that were easy to guess (because of the through-hole design). Trouble-shooting is not made easy, since Random*Source does not give schematics.

While most of the calibration instructions were clear, there weren’t any at all for the New Ring Modulator. I tried using the procedure for the Eurorack version of the New Ring, but I could not get the specified behavior. Instead I just set what seemed right, for example, tuning for minimal output in RM mode when there is no carrier signal.

The pulse outputs of the Divider are weak and cannot drive normal gate and trigger inputs. Apparently these outputs are intended to be patched over to the Logic section, which does not load them down and whose outputs (AND, OR, XOR, INV) are able to drive the inputs on other modules.

2 Responses to Random*Source 4U Serge Project

  1. Matthias Herrmann says:

    Nice work.

    Regarding NEW RING calibration, the procedure could be found in the Euro-documentation:
    http://randomsource.net/docs/RandomSource_Serge_RING_Euro.pdf

  2. Richard says:

    Thanks, Matthias. As mentioned above, I did try the Euro RING calibration procedure. No matter what trim pot settings I tried, I wasn’t able to get it to behave as described. Specifically, “(A) Set MODE fully CW and adjust the three trimmers so that the OUTPUT has no DC offset and shows a nice symmetry.” I could get a symmetry OR zero DC offset, but not both. So I made it symmetrical, even though there is a (negative, if I remember) DC offset.

Leave a Reply

Please use your real name instead of your company name or keyword spam.