Sunday, 30 December 2018

Power Supply For Korg Volca

For the last post of 2018, I wanted to write about the making of a small power supply for 4 Korg Volca.

Power supply inside
A Korg Volca can be powered by batteries or by a dedicated power supply : the Korg KA-350.  Not only is that power supply expensive compared to the affordable Volca, but it is over-dimensioned and you end up using one per box.
Moreover, Korg uses a slightly different connector than most 9V guitar pedals power supplies and the polarity is reversed (positive inside).  They correspond to the EIAJ-02 japanese standard.  Size is 4.0 x 1.7mm.

Power plug assembly in progress

There are some solutions out there, like alternative power supplies, adapters, daisy chain cables, etc.  Nevertheless, I couldn't resist to make my own power supply.

Bob screws the cables
This is a classical dual regulated power supply.    I used a design from the excellent Sonelec website.  You'll find the layout below.

The transformer is a 220V to 2x 12V transformer.  Bridge rectifiers are DB205.  Regulators are LM317.  The power plugs are PP-014 from CUI.  The enclosure is a 137x97x67mm ABS Polycarbonate Power Supply Case.

I bought the 5 connectors from Mouser and the plastic enclosure from Banzai Music.  All the other parts are leftovers from previous projects : I usually buy a little more each time to cover losses, destruction or to simply fill my drawers just in case...

Each one of the two outputs of the power supply is split in two with splices and a bit of heat shrink tube.

The transformer is only rated for 12 VA, which means 500 mA per output.  Volca are reported to draw about 80mA. This means a total of 160 mA per output.  That's enough safety margin for me.

That's it.  Let the fun begins.

Circuit and layout

Tuesday, 4 September 2018

Venus Cloud City

I took the opportunity of August's KVR Music Cafe contest to put the last upgrade on my small modular synth to good use.

The piece combines 6 tracks, out of which 4 are from the modular.
The first part is made entirely by the modular.  Then, enter the drums, made of TR-808 samples from NI and a synth pad from Korg Wavestation softsynth.
The main melody is played by the Korg SQ-1 in random mode and relies heavily on Rings.
Winds track consists of three sounds recorded separately and stacked.  They are made thanks to a white noise source, the filter from the DIY-101 module and the phaser.
Bass drone and the other bass pattern use the Doepfer VCO and the filter.  Rampage was used as  envelope generator.  The  movements effects on the bass drone are performed thanks to TAL Filter 2, great little toy.

DAW : Reaper
Synth : Modular Synth Suitcase, Korg Legacy Collection Wavestation, NI Kontakt 808 kit
Effects : ReaEQ, ReaXComp, NI Supercharger, TAL Dub II, TAL Filter 2, TAL Chorus-60, Voxengo OldSkoolVerb, Thomas Mundt LoudMax,

Photo : HAVOC project NASA / SACD

Saturday, 1 September 2018

DIY-101 synth voice - part 2

This is the second part of the making of my Doepfer DIY synth based module.  You can find the first part here.

Finished module in action.
Now is the time for some cabling.

There are 7 connectors on the circuit, numbered JP1 to JP7.  Moreover, I have to connect the 5V and common ground.
With the synth, I  ordered the set of dedicated flat cables to ease the interconnection.

I started with Ground, 5V and all connections that were not linked to a specific connector.

Before cabling connectors.
Pin 1 above
Then I did cable connector after connector, cutting the wires, splitting the flat cable, stripping with my nails and finally soldering … One wire at a time, in order not to lose track. 
I started with JP2 and JP3, as I reckoned they connected mostly to pots and jacks below the circuit and that would prove difficult if I finish with them.
All ground pins on those connectors were ignored as I chose to have a common ground from the 5V power connector.

Pin 1 below

Thanks to the schematics from M-19, I noticed the dot indicating the first signal (red line) on the connector was sometimes above or below in the Doepfer documentation tables.   I fell in the trap on my first schema.

Ongoing cabling

I also verified twice each wire connection before soldering.  And I did well. I noticed a couple of mistakes on my schematics this way.  I annotated my drawings along the way.
Annotated schematics

Bob checks a solder joint

 Bob managed to get some useful Quality Control time. 

My first tests showed I made some minor mistakes :
- switches pins are reversed with regards to the switch handle position;
- a couple of signals were wrongfully soldered on the switch pin of the jack, instead of the tip;
- I detected one or two bad solder joints that escaped Bob's control.

Here is the finished product.


Bob fixes knobs.

Finally, I let Bob place the knobs.

Tuning the VCO proved problematic.  Even after I let the board heat up for 20 minutes, I really had a hard time approaching the 1V/oct slope.
I'm not 100% sure yet, but the VCO is not very stable and slowly drifts, even with the tempco option installed.  Is the circuit defective ?  Is the design problematic ?  Was it damaged during my early tests with a home made power supply ?  I don't know.  The rest works like a charm.

Tuning the beast

Finally, all the Eurorack power flat cables I got for my DIY modules were 16-pin to 16-pin.  And JP1 is aligned in a way it is not possible to use a 16-pin connector properly.

I had to divert the 10-pin to 16-pin cable from Rings to put the module in place.

Time to buy some 16-pin to 10-pin cables next time I guess.

Finally in place.   Ready to wiggle.

Sunday, 26 August 2018

DIY-101 synth voice - part 1

I bought a Doepfer DIY synth nearly two years ago.
It was lying on my shelf ever since.  
This month I made a eurorack module with it.  I'll document it's making in two parts.

Close-up of the finished panel

The Doepfer DIY synth is an assembled printed circuit board that includes all the necessary functions to build a basic analog synthesizer: VCO, VCF, VCA, EG, LFO, you name it.  You can design your own synth.  All you have to do is to provide a case and all the connections to the mandatory controls : jacks and potentiometers. 

Before I could get to a finished product I'm proud of, I went into a lengthy thought process.
My first idea was to make a semi-modular synth whose functions are as close as possible to an Roland SH-101.

I tried several panel designs before deciding on the final one.

XAOC Karl Marx Stadt module provided the necessary faith that I could build a 42 HP (about 21 cm) module to house the Doepfer DIY synth.

Moreover, I was inspired by the Motherfucker-19 work from YokoBoko.  It helped me making some decisions on jack placement and avoiding some mistakes during cabling (not that I didn't do any…).  You can find description and instructions on the module modulagrid page.

As usual, I prototyped the panel on cardboard to check my placement hypothesis.  It seemed critical here.  I wanted to specifically check  the density of jacks on the low part and the room for the spacers to hold the board.
Back side of the early cardboard prototype.

Annotated paper and cardboard proto
I used my traditional way for panels : punching to locate the holes, drilling, then printing the panel, laminating and glueing it.

This time, I chose to drill the holes with a wood block beneath the panel.    It proves to be way more comfortable than my usual technique of maintaining the panel over the edge of the bench.  The panel is firmly clamped and I move the drill freely to get to the next hole.  I know, I should get a drilling press…
Holes are 6 mm for LEDS, jacks and switches and 8 mm for pots.  It is a bit large for the pots.  But it proved helpful later.

Bob cleans-up
This generated a lot of chips and dirt. Fortunately, Bob helped me to clean the mess.

Given the panel is 3U 42HP, or 212 mm x 128,5 mm, it fits nicely on an A4 sheet of paper.
It's a large sheet.  So placing it parallel to the panel edges is essential.  A small deviation could give an awful result.
I'm quite happy with the final result except the paper is one millimeter too much on the right and could have been half a millimeter lower.

Bob cuts holes

I cut the holes with an X-Acto knife.   Here again Bob gave me a hand.

This is the time to place all pots, jacks, switches and LEDs.  There are 51 of them (23 pots, 23 jacks, 3 switches and 2 LEDs).
As the holes for pots were a bit larger than needed I corrected the most obvious alignment problems by moving the pot half a millimeter here, a millimeter there.

Sorted potentiometers

Here is the final panel before cabling.

Final before cabling
Go to part 2.

Friday, 27 July 2018

Repairing my VCO

A-110-1 standard VCO

Maybe you followed this blog for a while and wondered why I didn't write more about the Doepfer A-110-1 Standard VCO that is shown on some photographs.
Or maybe you don't care.

This module is the first module I bought new  It's also the first I fried : I plugged the connector in reverse during one of my trials, before I have a case.

The module began to act odd.  Plugging its output to Rings input halted the later.  It became very hot.
I stopped everything and began to investigate.

 -12V pin of the module connected to an unconnected pin of the bus board, presumably close to ground voltage.  Gnd pin was at 12V and 12V pin at 0V.   I headed to the A-110 service manual from Doepfer to find out which parts could have been stressed.

I spotted two components that would certainly suffer to be powered in reverse (inversion of the 12V and gnd pins) : the CA3046 transistor array (IC3) and the 78L10 10V regulator (IC5).

The two components I changed.

I reckoned the two TL074 op-amps wouldn't have been stressed as they were powered between 0V and a floating voltage that would most certainly be close to 0V.  I removed them and tested them on a breadboard : they were OK.

I also checked the TL431 Reference Voltage and it gave me the proper 3V it should.  So it was OK as well.

Maybe other transistors were stressed but I have no idea how to verify without desoldering them.  The plan is to change the CA3046 and 78L10 first and test the module.

Calibrating the module.
After some difficulties to get a CA3046 (they are reported as obsolete in DIP14 package by main distributors) and de-solder the 78L10 (I guess my de-soldering braid was oxidized), I finally replaced the two parts. 
Et voila.  A quick check at the oscilloscope showed all outputs seemed fine.  Maybe only these two components were affected after all.

Finally, I re-calibrated the module, following the service manual, skipping step 1 though (too complex).

I can finally use my VCO.  End of the story.

Thursday, 31 May 2018


The second kit from Befaco I built is the Rampage.

And it is not for the faint of heart.

Befaco Rampage

The kit is good quality (as was the A*B+C).  However, with 227 parts (about 560 pins to solder), 3 connectors and around 40 jacks, pots and various switches, I wouldn't quality it as easy.

Rampage kit

I chose this module as an alternative to Make Noise's Math, one of the most recommended module in forums.  With some exceptions, Rampage fills the same roles.  Moreover, it is available as a kit which fulfills my desire to build my synth as much as I can.

Main board close-up

According to Befaco website :
"The module is built around two voltage controlled integrators. Integrators (also known as Lag Processors or Slew limiters) allow you to process your voltages, converting sharp square waves in ramp-like waveforms.
These lovely devices can, for instance, convert a gate signal into a simple envelope, or achieve a “portamento” or “glide” effect when applied to pitch CV."

A properly patched Rampage can assume more than 20 different functions : envelope generator, oscillator, comparator, polyrythmic gate generator, trigger delay, etc…

Before mechanical assembly
Assembly instructions are similar to the previous kit : same advantages, same drawbacks.

With a kit of that size and density, you have to be systematic in order to make sure to be "good first time".  Debugging such a circuit afterwards was not in my intentions.

Calibration of the triangle shape
I estimated that the two Schottky diodes of the kit were too large for the PCB footprint and I replaced them by some in my stock.

I also found an oddity in the presence of a solder joint below one of the potentiometer.  I reckon there should be enough room between the back of the pot and the joint when the pot is tightly attached to the panel, but I decided to mask it anyway with some insulation tape.

Better safe than sorry.

Still, I implemented the mod presented in this Muff Wiggler thread to ensure reliable gates by soldering 1M pull-downs resistors at the outputs.

Additional pull-down resistors

Finally, a photograph of the Befaco boys in action and a bit of sound.

Befaco boys in action

Here above, Rampage channel A is programmed as an LFO which frequency is determined by the output of the S&H and cadences the patch.  Channel B is the envelope triggered by channel A and whose attack is also determined by the ouput of the S&H.
A*B+C plays the role of the VCA.   The sound source is the pink noise from the S&H module.
The sound is coming out via Rings (whose parameters are, ... guess what ? ... controlled by the output of the S&H) and the phaser.
Additional effects : NI Supercharger, TAL Reverb II & LoudMax

Friday, 11 May 2018

A * B + C

To complete my synth, I recently bought 2 DIY kits from Befaco.

The first one is called A*B+C.

Finished module

The kit is  good quality and rather simple.

Before soldering

The A*B+C is a dual, four-quadrant multiplier with VC Offset.

It can perform different operations on its inputs :
- amplifier and inverter of A input.
- VCA on A, with envelope CV on B.
- VC offset generation with C input.
- attenuation/inversion of C input
- ring modulation of audio signals in A and B.

Moreover, the output of the the first channel is normaled to the mixer of the second channel.  So the module can serve as CV/audio mixer.

Before assembly

Assembly instructions are good but could have been clearer, especially for the final assembly.
A couple of photos wouldn't have been too much.  But I didn't encounter any major issue.

My only gripe was about some parts that had a larger size than their footprint on the PCB.  I had to replace the two Schottky diodes with those I had in stock and solder 2 couples of resistors vertically. One resistor even had to be soldered on the other side of the board.

Once again, nothing major in my view.

Profile view.  You can see the vertically mounted resistors

Finally, the module suffers from a high level of background noise as indicated in this Muff Wiggler forum thread.   Anyway, it does not seem prohibitive.   There is a suggested mod but it looks like it modifies the potentiometers behaviour, so I didn't want to perform it (and I didn't have the necessary 1/8 W resistors).   Time will tell.