Giddy-up – the fantastic Minilogue just arrived for exploration! Been eyeing this for a long time now. As this is my second poly synth modding project only, this might be expanding slowly/see some revisions over time. I also posted an open contribution thread on Modwiggler and hope some of the good folks there will chime in with ideas.
Some demo of Minilogue with boosted signal path, playing bass arp in context with MC909:
At any rate, here is some demo in which you hear slight stereo panning (modded) and voice intermodulation, where one vco modulates that of another voice and so forth…
PCB Layout and Voice Structure
Service Manual (yes, Korg published the schematics!): https://www.synthxl.com/korg-minilogue/
Before diving into modifications I think it might be helpful to provide some rough voice labelling for orientation. The main PCB carries your four voices, with components densely populating both sides of the circuit board. While for some modifications you need to take out the entire PCB, many tweaks can be done with the board in place.
Modifications
What I’m after here is sound modifications rather than changing the external aesthetics, especially since the latter has been done so excellently already! At the moment I am planning for two different general areas – permanent sound tweaks aimed at changing the general timbre a wee bit (I like the “hard” sound of the Minilogue in general, but want one or two things for the filter), and one being aimed at more flexibility/modularity in terms of sound combinations. The latter area especially might need some more development, as IC switches and/or digipots will be needed.
A list of what to find below: Audio output per voice, then some stereo panorama mod for voices (repurposed phones output). For for techno mayhem you get some increased delay feedback, and for otherwordly sounds increased XMOD amount, increased Ringmod, and, fasten your seatbelts: voice-intermodulation (one voice frequency modulates or ring modulates VCO2 of another voice) – does any standard analog poly have this option? VCF input boost and some level drop compensation on higher resonance setting, as well as a new, switchable bandpass filter response spice up your VCF, and VCA boost and VCA overdrive are there too.
VCO
Increased cross modulation amount: one user on modwiggler was interested in boosting the maximum fm amount in the VCO section. you can easily do this by reducing each R351 (47k). On mine, ill halve that value by wiring another 47k resistor across the existing one, so that the old maximum is reached when the pot is turned halfway up.
Crossmod voice-intermodulation: remove the 47k input resistor of the xmod VCA (R351), then use a x4 A/B switch and new 47k resistor for toggling between the own VCO of each voice and that of another voice.
Ring modulation voice-intermodulation: same principle, except for ring modulation.
NB in all demos of Ringmod you hear I changed the input resistor for SND1 (that’s R333) from 1M to 47k. This happen on account of a mistake forst (working on the wrong switch…) but I liked it so much I’m keeping it!
VCF
The VCF of the Minilogue is built around a LM3900, a quad norton amp with each cell set up as a -6db lowpass filter (similar to the Arp 4072 VCF). The -12db and -24db outputs are buffered with simple op-amp repeaters and then selected per IC-switch. Further down the line I’ll do some test with mixed filter responses (bandpass) and replacing the -12db with a -6db low pass.
Resonance is realized per simple VCA (like the VCO Waveform VCAs and the main VCA) for purposes of CPU control.
VCF Input Boost: The sound of the Minilogue filter is sometimes called flat, which is an impression I share partly though not wholeheartedly. Some zest can be added by driving the input a bit harder., which is a modification as easy as wiring a resistor across each R424 (47k). There are four of these (one per voice), and you find their location in the picture above. As for resistor values, I find half of the original value hits the spot (wire another 47k across each R424).
Passband level loss compensation (AKA reso “bass drop” compensation): When turning up resonance, the Minilogue filter loses level (often perceived as bass loss), as do many other filter of this type. Thankfully, this can be counteracted by a super easy modification that piggypacks onto the resonance VCA.
Simply wire a 68k resistor between the top terminal of R424 (labelled RC-A above) and the top terminal of R425 located at the resonance VCA (R425 is not populated on the final PCB).
What happens here is that when resonance increases a portion of the filter input signal is fed into the feedback path, and this in proportion to the amount of resonance. More reso = more input signal = more consistent level behaviour. For my taste, 68K is just about right, yet if you want to have a more pronounced compensation, use a 47k resistor, or go suuuubtle with 100k.
Turn -12db response of low pass to -6db response: Basically, this is as easy as rewiring two IC pins, yet, here comes the challenge – you need to solder some wire directly to two tiny IC pins… Hence, only do this if you are confident at such a task and maybe best don’t do it after watching Stranger Things or something… There’s a quite visible trace between IC28, PIN4 (LM3900, OUT2) and IC7, PIN10 (that’s the +IN of the op-amp buffer of the 2pole output). Cut this trace and rewire IC7, PIN10 to IC28, PIN5 (LM3900, OUT1). NB: I have not tested this (might soon).
Mix additional -6db stage into VCA for bandpass: this is easier than the above mod since all you need is a x4 switch and four 47k resistors. Simply wire the -6db pin of the LM3900 (see pic above) via a switch and a 47K resistor to the left hand terminal of R352 (on the back of the PCB, the terminal facing the dual transistor). BY mixing the -6db stage with the -24db stage, you create a sweet sounding bandpass. This sounds good in -12db mode too! Here are some audio demo with various sounds:
As you can see, this mod can be done on the back of the PCB where you don’t heed to solder onto IC pins. Here are some polyphonic sounds in the following demo clip:
VCA
At its core, the main VCA used in the Minilogue consists of a differential pair of PNP transistors and an op-amp. The audio input is attenuated by means of a 47k resistor (R352, located on the backside of the PCB).
Increased audio input level: Reducing the value of R352 bumps up the incoming audio volume. Here’s how it sounds when you halve the value of that resistor:
Personally, I think I keep VCF boost but leave VCA boost aside, or, if at all, use it very sparingly, mainly because the former introduces more character along the obvious volume changes. Should you want to do both modifications, however, be mindful of your overall headroom lest you want to rename your Minilogue to “Screecho” or “Mc Creackle.” Rearranging the resistor array Ra1 to higher value resistors, say from 10k to 20k, should counteract unwanted distortion at the summing stage though!
VCA Input overdrive: Now, when I just wrote VCA boost does not introduce a lot of character, this may count as a lie when you increase the volume input by a lot – oh yeah, baby, when I write a lot, I mean *a lot.* The following audio demo gets VERY LOUD after some 15 seconds, please listen to that one accordingly – don’t damage your precious ears!
Will need to toy around with this one a bit more, but would be nice to have this on a digipot.
Increased VCA control input: oddly, this one is more to my liking than an actual audio input boost. The VCA is controlled by a digital envelope, which, turned into CV, controls the main VCA level. This control signal passes R347, a 2.2k resistor (NB: schematic reads 3.3k). If you reduce the value of that resistor to, say, half of that, your envelope hits more. I did this by simply wiring a 2k resistor across R347.
Delay Modification
I love the internal delay of the Minilogue. Some folks complain that it is noisy, and yes, it is, but hey, I like it… If you need a “that’s so raaad” spin: Befaco Crush Delay, Makenoise Strega, Soma Lyra 8 – whooooo.
One wonderful aspect of this digital BBD-emulating delay is that you can drive it beyond nice & halfway acceptable into dirty and glitchy territory. I’ve only just begun toying around here but found some fun things. For further inspiration, check this wonderful information resource on the PT2399 by Electrosmash.
Increase delay input/feedback intensity: By reducing R155 (10k), you bump up the input/feedback loop, turning a moderately chosen feedback amount into a self-oscillating abomination. Wire a 20k resistor across R155 and you have a sweet range – do this either fixed or per switch (fixed version will change your preset settings, of course).
Change minimum/maximum delay time: easiest way to set delay time in the PT2399 is per resistor between PIN6 and ground. If you want variable delay, a potentiometer is in order, of course. (As for times and values, best check the PT2399 datasheet.) One drawback here is that when powering up the resistance between PIN6 and ground cannot be lower than 2k – otherwise the delay doesn’t work. Yet, 2k gives you a longer delay as you want for flanger/chorus type things. Hence, many implementations use a so-called latch-up circuit, a little contraption that sets your R at 2k or over for around half of a second and then allows for lower R. The Minilogue does not have such a latch-up circuit.. So we need to be a bit careful.
For changing the delay time span, look for R173 on the upper left hand of the main circuit board. Increase R173 (1k) per switch to 10K and you get wonderfully long and grimey feedbacks.
If you want to go higher speeds to, for comb filter effects, switch R173 to a 100hm resistor (that’s the lowest value that has an audible effect). Problem here is that when the delay time setting in the preset to which your unit powers up is at zero, we have a latch up and the delay doesn’t do its thing. Easiest solution is to set the delay time in your program at halfway up and save that. Units powers up, delay works safely…
Dry signal left – delay signal right: On my unit I eventually decided against the panorama modification (see next section) but also against he conventional headphones output. So I repurposed that on efor having clean and delayed signal on separate channels. Qick and dirty version: chuck out IC 30 and chuck out R5 and R19. Rewire the top terminal of R142 (see delay circuit in Minilogue schematics) to the junction IC30-R6 (10ohm) and the top of R143 to the junction IC30-R20 (10ohm) – boom.
Voice summing and main output
Increased main output level: without any other modifications, the Minilogue can sound a tad quiet. If you like a good racket, simply increase the value of R3 from ca. 2k to, say, 10k.
Decreased output volume: if your main level goes through the roof on account of all them wicked mods, try replacing Ra1 with 20k resistors. This, however, is a bit fiddly, so a more sensible option might be to just use a cooling tent, like we did yesterday with the heat and all…
Individual output per voice: before entering the final summing stage, each voice output signal passes a 10uf capacitor. For a non-switched output, simply tap the signal at each capacitor (the terminal facing each voice number label; see below) and feed it through a 1k resistor to your individual voice output socket.
Here’s a demo of two sounds played with single outputs – we’re listening to three voices, panned left, middle, right. Now, imagine this with a good musical line and some little FX on top, hmmmmm….
If you want switched outputs (any voice fed to single out is subtracted from main mix), you need to break the connection between capacitor and Ra1 and feed that via a switched output socket. Safest way to do this, probably, is to lift the capacitor terminal facing the respective voice number label on the PCB. If you have never desoldered one specimen of this capacitor type, NB that you want even heat distribution on both terminals and take things slow, as these little caps can easily tear off some PCB terminal when going to fast!
Panorama modification: You heard about this modification when your mom‘s dad did this on his old Prophet 5, and you can do it too on your Minilogue 4! Here, I describe a version with fixed panorama settings, but variable panning is doable too.
First we tap each voice at its 10uf output buffer capacitor and feed the signal to two resistors of different values one resistor goes to a LEFT summing node, the other to a RIGHT summing node. All voice signals LEFT and all RIGHT are then mixed and amplified with an opamp stage and then sent right to the headphones output. Using a TRS to TS splitter cable, you have now stereo voices in addition to the mono output. Sounds like this:
In my version I repurpose the dual opamp (IC30) originally used for headphone amplification. For doing so, remove IC30 carefully, then cut the following traces:
Now put IC30 back in. Then wire to each 10uf voice output buffer cap another 10uf cap as shown below. To your new cap, wire your LEFT and your RIGHT resistors.
All your LEFTS and all your RIGHTS are then wired together. The two nodes you thereby created are connected as follows: one to Ic30, Pin3 and the other to IC30, Pin5. Panning is determined by the value of each resistor: for centre panning use 11k on both resistors. If you want to tilt, use one resistor with 8k and the other with 14k. Why is 11k at the centre? Both amps at IC30 have a 22k resistor in their feedback loop, and with 11k as half of that you get an amplification of x2 there (suits line in level better than lower signals).
The only thing left to do is to reconnect C11 to IC31, Pin3 (main output amp), and to connect one 100k resistor between IC31, Pin 3 and ground, and another 100k resistor between IC30, Pin5 and ground. Done!
Some variations to this: if you want to have your panning variable by means of internal trimmers or little pots on the back of your Minilogue, try the following. Wire the voice output buffer cap to the wiper of a 10k trimmer and pins 1 and 3 of that wiper via 5k resistors to the RIGHT/LEFT nodes. (not tested that one). Another mod would be to make the volume of your pan-output variable. In this case you’d need a dual gang 25k potentiometer replacing R19 and R5 (both 22k).
Placing controls and extra parts
Inside the Minilogue is quite some space for an extra PCB, even with some vacant spacer for placement.
As for placing controls, there is less space on the top panel. You have some space on the sides if you clip off some plastic from the control PCB and don’t mind drilling holes into the metal plate.
If you keep the keyboard, you will find space for quite some extra controls at the side panels! There is space for fitting things vertically and or horizontally.
Moreover, there are some recessed bits in the control board that fit smaller parts. Matching a corresponding hole in the panel needs quite some accuracy though!
