Just in time for Spring Break I scored a second hand PWM Malevolent, the owner of which forgot to mention that they use quite a lot of cologne… This incident compelled me to disassemble and clean this synth but also changed the name to “male vole N.T.” in my head. (Please, voles, and please Paul, forgive me for this association.) Anyhow, this thing rocks! Far from being yet another mono synth, this is more like an alternative universe instrument, where King Tubby and Arturia teamed up before Dub was invented, except Arturia aimed for better build quality.
Circuit Tour
The Malevolent by PWM and Future Sound Systems is really efficiently designed. Looking like a typical Arturia or Behringer style injection mould plastic chassis + wobble keyboard on the first glance, the PWM shows a couple of custom design choices that are not cheap at all: metal faceplate, slanted spacers, custom-ish LED soft buttons, Cliff(?) jack sockets, spacers for accessing the trimmers (awesome measure to prevent damage by slippery screwdriver), and more. The keyboard is still not to my liking, though, so I chopped that one off. Shame about the joystick, but space is limited and all…
The PCB is quality and sports some less usual circuits that finally give some variation to the wealth of same-same mono synth designs. Right in line with the whole Hugget CMOS design legacy, you have some VCOs based around CD4046 Phase Locked Loop ICs and a state variable filter that utilises Opto FET cells. The used components are standard (maybe except for the H11F3M opto cells, and rather than using a ground plane on the outer layers, the component/backside of the PCB show only thick and well visible signal/supply voltage traces, which renders little incisions an easily accomplished task.
The I/O board is solid and has what you need (except midi via USB, right?).
The VCOs
Both VCO’s are designed identically and, as mentioned, based on a CMOS phase locked loop chip. One famous example of using that IC for multi-waveform VCO duties that require only few external components and some clever trick for exponential control is Thomas Henry’s X-4046 VCO (also have a look at Eddy Bergmann’s DIY page for audio examples).
There is an expo converter using one OPA stage, including an expo for compensating thermal drift, and a dual NPN, driving the 4046. While the proper PLL VCO output is a square, you can tap various pions of that IC and make composite waveforms, which you then offset and mix in order to gain TRI and SAW. Inherent in this design is that SAW sounds one octave above TRI and PULSE (reminds me of the SAW trick on the syncussion oscillator, which also uses only a half cycle TRI for SAW). In comparison to Thomas Henry’s VCO, the oscillators on the Malevolent allow for additive waveform switching (done via a CD4066) and continuous waveform shaping via a dedicated knob or CV. Neither do they offer linear FM nor VCO synchronisation. All waveforms are buffered before going to the outputs, and the VCO/ext/noise sum is mixed at a simple OPA stage.
The VCF
Is a state variable Opto FET filter, something you don’t see often! The Opto FETs drive the integrating capacitors, for which you would usually take OTA cells or VCAs, and in contrast to conventional SVF configurations, where you use one input and tap three different outputs (LP, BP, HP), you feed the input signal into three different places in the circuit to obtain a LP, BP or HP response on the common output. This is something we usually see in the Steiner Parker design. Also more than usually the case, this filter is extremely sensitive to input levels. with only one VCO mix potentiometer pulled up, the resonance is drowned out completely, while it simply bites your ears off at lower input levels. This can be harnessed in wonderful ways, using the waveform shape modulation (which affects fundamentals/overtones of individual waveforms and with that, of course, median waveform level).
VCA, Envelopes and LFO
In contrast to VCOs and VCF, the VCA is a pretty conventional transistor-differential OPA setup. Sounds good and does the trick though! Equally, the ADSRs (based on an NE556 dual timer IC) and the LFO (classic TRI/SQUARE relaxation oscillator) are straightforward and well established designs. While waveform looping and LFO waveform shaping would have been cool, the modular design of the Malevolent makes it easy to patch up something, so fair enough!
MODIFICATIONS
VCO SYNC (2<1)
Thomas Henry’s design shows how to do things properly, but I am not always proper, so my version simply switches the capacitor between pins 6-7 of PLL2 to the output of PLL1 via a simple fast switching diode. Every time VCO1 output (C27, where it connects to U7) is zero, it pulls VCO2 capacitor C53 (top terminal) empty through the diode (1N4148, cathode connected to C53). The cool side effect of this way of syncing is nice glitches when VCO1 frequency is higher than that of VCO2. If you were to implement VCO sync input sockets for use with external VCOs, use Henry’s design and avoid mine, since the CD4046 is powered between VCC and Ground (i.e. negative voltages can do harm).
LINEAR FM
I’ve not tested this, but you could very easily implement LIN FM (dc coupled or not) at the -IN of the CV summing OPA via a (capacitor) and resistor. On he VCO2 Core picture above, I marked the point to which you can apply linear FM. For how to set up the peripherals, best consult the Henry X-4046 VCO schematic, sheet 1.
VCF INPUT LEVELS
Since the VCF is crazy cool as is already, I simply bumped up the BP and HP input levels a little by changing the input attenuating resistors from 4k7 (BP) and from 10k (HP) to around half of each value. For HP, around 7k5 would be best, I guess, but I don’t mind that response being a bit louder, so I kept it at 5k.
FILTER FEEDBACK VIA AUX IN
Since the AUX in jack socket is a switched socket, and since the NC lug is not on ground, you can easily pre-wire some connection there that is broken when you plug in a cable. And since the Malevolent gives some nice feedback overload when you pull overdrive in the upper third and feed the main output back into the filter, I just wired PIN3 of the volume knob via a 1k resistor to the NC lug of AUX IN.
VOLUME CONTROL for MAIN OUT socket on top
I don’t like that the 3,5mm output jack is always fully on while the main volume knob only affects the out jack on the rear, so I rewired that front socket from VCA out to the middle pin of the volume potentiometer. You need to cut a trace leading to the output socket (which you can’t miss) for doing that.
