Here’s a first playlist: https://soundcloud.com/uibkmedan/sets/b2600-modifications
Yup, another Behringer for modding… As for modifications, I’ll stick to the “commonly-known” material and will include some aspects of a list of very useful 2600 modding suggestions given by KSS in the muffwiggler forum (comparator at the envelope follower, eswitch ext clock, and a couple more). There are many great findings and descriptions in the context of TTSH builds, yet I would deem it inappropriate to “milk” these without explicit consent. In some cases I might indirectly point to respective TTSH findings.
Great ARP Resource: https://github.com/CreativeInquiry/ARP-2600
Fair use and support:
The information gathered on this page is for fair use only. This is not because I find my stuff so great that I am worried about being ripped off – this page is for synth/DIY enthusiasts and the work put in here built on trust and enthusiasm. If you lack the technical skills and hire someone to mod your 2600 based on the info here, great! If you mod Behringer 2600s for commerical use and base your mods on the info here, you donate 10% of your profit per unit after the second unit sold to any of the collectives/organizations in the following list: https://drlizdobson.wpcomstaging.com/2018/02/18/feministsoundcollectives/. If you like these mods and want to support this page, donate any amount you see fit to any of the collectives/organizations in the same list.
Behringer 2600 info & mods
CPU: is an ARM GD32F510
Portamento with analogue control (on/off, mom, time) but implementation in software. I personally find the lower portamento range has some headroom for improvement in that you move from nothing to just a bit too much too quickly.
1V/OCT CV FROM CPU
The internal midi interface of the Behringer 2600 offers two 1V/Oct lines for duophonic play. The signal form the CPU is amped and filtered at ICs 18 and 26 for the “KBD CV” line and ICs 17 and 25 for the “Second Voice” line.
MASTER CV INPUT MOD (tested, works well): If you want to sequence the 2600 externally via CV, each VCO needs a dedicated CV input, which, when played monophonically, eats up a multiple and three extra cables. If you want to keep the flexibility of dedicated CV per VCO but also make use of quick and simple patching, you can easily insert a “MASTER CV IN” that breaks the connection from the internal midi converter and all “KBD CV” sockets. Simply lift R408 (100ohms) and wire the top terminal (leading to IC17) via a 100ohn resistor to the normal lug of a switched jack socket. The tip lug of your new socket goes to the lower PCB terminal for R408. The ground lug of your new CV in socket must be on ground – just use any ground lugs of the existing jack sockets – otherwise you’re in FM garble hell (or heaven, depending on how you click).
VCOs are as per 4027-1 schematic. I need to trace more, but some parts are mapped in the pictures. All pp-amps used here are LM4580. Sync mod seems to be like Odyssey-derived classic suggestions with an NPP and a FET. No mods of my own planned.
Notable differences between ARP schematics and the layout on the 2600 is the option to use SMD or through-hole filter capacitors and the LM4580 op-amps replacing LM301. In the 4012 type, the AD3958 dual FET is replaced by two single FETs, T15 and T16 (each labelled “XY”). Moreover, in the Behringer the filter input signals are actively mixed.
Input signals are summed via R301-105 and pass an inverting amplifier (IC 20-B), are inverted again at IC20-A and passed through to the 4012 type via R144 (44k) and to the 4072 type via R172 (44k). FC CV is summed and inverted at IC21-B, and inverted again at IC21-A. For the 4072 type, FC CV then passes IC22-B (=ARP Z2A). IC11-B is inactive.
The filter type select switch toggles between output/feedback path PINs as equivalent to terminals “2” and “5” on ARP picture below:
The filter type not presently selected is still active as signal input and FC CV are always present on both. The output of the filter type not presently selected is active too (can be tapped at IC11, PIN1 for 4012 and IC22, PIN1 for 4072) but at zero resonance.
Accessing both filter types simultaneously (nope): Since resonance control is shared, an independent setup would require decoupling the type select switch and adding a sub circuit that duplicates the setup of R161, C62, and R160 as in the ARP schematic above. Yet, this would still mean that signal inputs and frequency CV in are identical on both types. The latter could be changed by inserting switched jack sockets by means of which one can break the internal signal flow, yet overall this seems too much effort for little gain.
Pre-filter overdrive mod: Driving both filter types with a hotter input mix signal can easily be done by lowering the value of R191 (74k), which is located between IC20-B and IC20-A. At the moment I prefer a switch for this (over a pot ) that roughly halves the value of R191.
Eventually, I decided on 150k across R144 and R172 each to keep levels matched. My impression is that the 4012-type retains its “woody” characteristic for longer when overdriven, while the 4072-type starts sounding clippy earlier.
Direct filter input mix out (not implemented on mine): since the input signal sum is buffered by an opamp, you can use this for a mix out and/or even an inverted mix out. This could be interesting for creating different filter responses. Tap the outputs of IC 20-A (inverted) and/or IC20-B (not inverted) for mix signals.
Resonance level loss compensation (tested, works): Background: some filter lose level on high resonance, and nifty engineers like Yusynth, Émilie Gilett and others offer solutions for compensating this. The trick is to tap the resonance slider and mix an amplified version of this signal to the final filter output – the more resonance, the more additional signal… As I yet feel reluctant to drill additional holes into the chassis, my solution is to route the level compensation signal to the VCA IN2 slider. This breaks the internal connection to the ring modulator, but since that one has a dedicated output socket, nothing is lost.
How to do this: tap the wiper of the resonance slider (at the filter type switch, see pic below), and feed that signal into an inverting it op-amp (74k at -in and 47k feedback resistor), then feed this into the normal lug of the VCA IN2 socket via a 1k resistor. Break the internal connection between the normal lug and ring mod output by cutting the trace leading to the normal lug (see pic). If you don’t use VCA IN2, your opamp buffer needs to be non-inverting (signals into VCA IN2 are invetred, so you need to invert in the above case to end up with the correct signal polarity). NB: the VCF output socket remains unaffected by this- if you wanted that, you’d need a switch/jumper for com on/off and sum your signals differently.
FILTER RESPONSE MODIFICATIONS FOR THE 4072 TYPE
-6db and -12DB OUTPUTs for the 4072 type: tap the pole outputs and amplify the signals by a factor of ca. 9.5.
-6db/-12db on the 4072 – lazy type: On my unit I did a bit of a crazy thing and send these outputs via 12k resistors and a select to the preamp input. With preamp gain factor X10 and the volume slider fully up, the pole output level matches that of the normal filter out well enough that you can create bandpass filter responses when mixing -6/-12db with the –24db out. Preamp in needs to be rewired to a switched socket (the normal lug of the existing socket is wired to ground). This is probably not fully respectable audio engineering, but at the moment I fail to see any harm in this. The AC coupling capacitor at the preamp input (C24, 100nf) filters some lows but I tested with/without and find it actually very helpful in limiting resonance screech on -6b/-21db. At any rathe, here’s how it sounds:
Uses one half of a TL072 as op-amp. Most notably, the High Frequency Reject trimmer (labelled “High Frequency Resect” on the Behringer) and its surrounding circuitry (Behringer R364, C206, C208) are not populated. No mods of my own planned.
ADSR and AR
Triggering/gating ADSR and AR independently: The Behringer is different to the ARP in that you can trigger/gate the ADSR and AR independently. This means, apart from the timing capacitor (spped switch), this is a modification we get ex factory. Since the setup is potentially confusinge, here’s a quick description:
Set switch labelled “1” to upper position. Send gate (+5V or higher) to socket labelled “2” – this engages the AR. For engaging ADSR, you need to send another gate to sockets labelled “3” and “4” simultaneously (use a multiple or stackable). This second gate needs to be +5V or higher. Confused myself here for a bit, since this is quite different on the ARP/TTSH but B2660 quickstart guide (page 17) confirms gate in threshold of +4v and trg in threshold of +5v.
AMS Synths pointed out in the comment section here that also the AR circuit was adjusted to have a faster attack time, a mod done by using a transistor instead of a diode. (BTW it’s great that Robert Keeble is pointing out technical aspects more concretely.)
Boosting AR Level with external gates of +5V: When gated by the pushbutton or the internal SH clock, the AR produces a higher level than when gated externally by +5V gates. This is because the AR is not a proper “envelope” but a lag processor that slews the incoming gate voltage. In the following clip you first hear internal gate, then an external 5V gate, then a boosted gate with AR controlling VCA volume:
If you use +5V more often than +10V gates, this might be a mod for you. No trace cutting is needed and there is no interference with the internal wiring, i.e. gates by pushbutton and S&H remain the same.
An external gate at AR GATE IN opens Q1, which then passes through +10V to the base of T82 (T82 is the previously mentioned AR timing mod). D1 protects Q1 from negative voltage, R2 is for protection against shorts (maybe not needed). D2 makes sure your added AR boost does not interfere with internal gates. On the following picture points for soldering on the back of the PCB are labelled:
Alternatively, you can use external gain boosters (Ladik Gain Up, or Doepfer 183-4) or check out my utility module suggestion on the bottom of the page.
Normal ADSR Gate in socket to TRG in socket: this saves you a multiple when triggering/gating the ADSR. Wire the tip lug of ADSR GATE IN to the normal lug of TRG in via a diode (cathode to normal lug of TRG in). Plugging a cable into TRG in breaks this connection.
ADSR TRIGGER THRESHOLD: Should your external sequencer/synth not deliver the necessary voltage to trigger the ADSR of the 2600s, you can lower the voltage threshold for triggering by reducing the value of R427 (120K, equivalent of ARP R20).
Fix envelope clicks: This may be nitpicking, but the envs are clicky – again (again, as in this is not a Behringer first). Yes, this is related to more aspects than just the envs, but these also click on my Dominion1 and Neutron, so the first mod I did was a 220nf between AR jack socket tip lug and ground. For location see reso level compensation picture above. [Might do 47nf-100nf on ADSR also.]
No mods planned, but the LFO retriggers on gate on, and a potential mod would here be to either use an on/off switch for this or feed the trigger voltage through a normalled jack socket (internal trig, dummy cable in = no retrig, external trig). T103 would be the respective FET responsible for retrigger.
No mods planned for this one.
Not really any mods planned here for the noise circuit itself.
LF Noise to LAG PROCESSOR IN: KSS mentioned having modded some ARPS with direct LF noise outputs, which is a neat way of having a dedicated modulation source available if you use white noise for audio purposes. (Think using the SH regardless of noise level fader setting).
Since I don’t feel like drilling holes yet, I just broke the normal connection between Envelope Follower out and LAG in and rewired LF noise from the Noise Colour Fader via a 1K resistor to the normal lug of LAG in. With lag slider fully left, you have LF noise and with increasing lag, you filter the noise even down more. Gives you a secondary noise modulation/audio source without much effort!
Here’s some demos of LF out into LAG
CV for S&H CLOCK SPEED modification (general suggestion by KSS): I installed a CV in socket for controlling the speed of the SH clock. Wire a 15k resistor to the collector of T99 and a diode from ground to the junction of your input socket tip lug and that 15k resistor to clip any incoming negative voltage. [15k impedance is low – eurorack “standard” would be 100k), yet T99 receives 0-15v via a 40k resistor, R534, so unless you send in some 100volts, this should be okay. Belts and braces would be using an opamp buffer, I assume.]
The electric switch is set up as per schematic, so I just describe the classic ext clock mod here. This mod is not new and comes up in many discussions, and while it is straightforwar, it opens up a whole lot of additional possibilities.
The electric switch on the 2600 is a useful circuit for all sorts of applications, from clocked distribution of signals to creating a sub oscillator, and more. The switch is bidirectional, i.e. you can use it to toggle between two signals present at “A” and “B”, which are then sent out at “C” in an alternating fashion, or you can use “C” as an input, which is then alternatingly sent out of “A” and “B”. The toggle control of that switch is the S&H clock, yet, unlike the S&H, the E-SWICTH toggle is not affected when plugging an external clock source into the EXT CLK IN.
Fortunately, you can easily DIY this and rewire things so that the E-SWITCH toggle can be externally clocked. You need to cut one trace as marked on the first image below (NB your cut needs to the left of the tiny circle in the trace, otherwise you cut the internal clock to SH connection!). Then connect terminal “MOD X1” on the first picture with wire to the terminal “MOD X2” on the second picture.
Alternatively, you can also wire up a dedicated E-SWITCH clock input (you’d need to drill & place a socket for the latter), so that you can clock the S&H and the E-SWITCH independently with external clock sources. In this case, cut trace as shown and wire normal pin lug of EXT CLK IN to a switched socket (normal lug), which then (tip lug) goes to the “MOD X2” point. The existing setup around C383 converts gates and longer pulses into short trigger pulses, so no extra circuitry is needed. Once I decided on a layout for additional controls, this will be the first hole to be drilled on mine.
“Lazy” version of this mod (since you do not need to lead wire around from the back to the front of the PCB) is as follows:
Cut the trace as in the other version, but since we’re bypassing two resistors on the front of the PCB wire using the through VIA solder point, we need to add them to the back. Now, here’s something weird: as per ARP schematic, we’d need to wire the Clock socket lug to the junction of a 10k and a 22k resistor, with the 22k wired to ground and the 10k to our re-entry point (junction Behringer R708-C383). On the Behringer, however, the values of these resistors are slightly lower, and with the extra setup, internal clock signal is too low to get the E-Switch going. I tried different values and ended up simply wiring a 1k resistor between the tip lug of the EXT CLK socket and the junction of R708-C383. Tested and working well with internal clock and VCO1 pulse wave into EXT CLK IN.
Normal +10V to ESWITCH “C” SOCKET: this was suggested by KSS and saves you cable plugging. Tap +10V from the voltage processor and wire it via a 1k resistor to the normal lug of socket “C”. Especially when using external clock, this is fun! Feed in VCO pulse to clock, route another waveform af the same VCO into the filter and route ESWITCH “A” to the filter too – subosc is ready. At lower/internal clock rates, your ESWITCH works as a squarewave LFO. [Behringer often keep the non-prewired normal lugs of their sockets on ground, yet those of ESWITCH “A”, “B”, and “C” aren’t.]
DIGITAL FX (XMAS-TREE VERSION)
The digital spring reverb is based on a Coolaudio V1000 Multi-FX DSP chip and a Coolaudio V4220M for AD/DA-conversion. Since the V1000 offers several pre-programmed FX algorithms that can be addressed by external voltage, you can do a couple of nifty things without much effort.
Reverb Spring Rattle Mod: This one is silly but fun. Spring reverb tanks rattle when shaken, and using this for musical effects has become some stock element in “rock and roll” gestures (high on the list with keytar banging and wiggling your fingers when playing vibrato per aftertouch). Some FX boxes, like the Vermona Retroverb Lancet, even offer options to do this rattle by the push of a button, and this mod here emulates that. Wire a momentary-on pushbutton between IC6, PIN7 and a 1k resistor. The other terminal of your resistor goes to ground. When you push the button, the voltage on PIN7 of IC6 is pulled below +2V, causing the FX DSP change from program “Spring3b” to program “Chamber2” (which is a 7-second church reverb). If you push that button quickly the resulting FX sound is at least a bit reminiscent of rattling springs.
FX Program Rewiring: As per V1000 datasheet, you have 16 internal FX programs (i.e. algos) available, including reverbs, delays, phaser, flanger, and chorus. Programs are selected by a simple on/off-state combination of four IC pins (see pic above – green = “prog0”, blue =”prog 3”). On/off, in this case means if a voltage between 0 and +0.3V is present at the pin, it’s off, and if a voltage between +2V and +5V is present, it’s on. In the case of the spring reverb prewired on the 2600, this would be, from “blue” to “green” – off, on, on, on, and accordingly, you see IC6, PINS 7, 6, and 5 connected via 10k resistors to the top terminal of L4, which carries around +4.7V, while IC6, PIN8 is wired to ground via a 47ohm resistor. Rearranging your presets is as easy as removing R80, R81, R84, and R85, and reconnecting the program pins according to which FX Program you would like to have. Keep a 10k resistor between each ON pin and L4, and a 47ohm/100ohm resistor beween OFF and ground. Suggestion: use a DIP switch or, if you want to be fancy, a switching IC. Punk suggestion: route an LFO (via a 1k resistor) to any of the program Pins. For available programs, see V1000 datasheet, page 4.
Here’s some delay, chorus and reverb switching:
UTILITY MODULE FOR ACCOMMDATING FURTHER USEFUL MODS
If you use a modular synth and have some space in your rack, here are some utility module suggestions I find helpful as 2600 aides.
KSS’s suggestion of an additional comparator as a 2600 extension is, bluntly put, the bomb, so I included one (derived from yusynth). In brief, this helps you creating loopings envs, delayed triggers/gates, syncopated rhythms, but is also useable in audio range as a sine/tri/saw to square converter that even allows for variable pulse width. There are dedicated Eurorack modules too (check Ladik’s comparator, for instance), but if you like DIY, go! Input A is measured against voltage B (as set per pot) and the upper output spits out +10v if A>B and the lower spits out +10V if A<B.
Next, I implemented a slew limiter (based on haraldswerk, in turn, inspired by Yusynth) with an up/both/down switch and a momentary on button. Since the 2600’s portamento only works with the internal midi-cv converter, I found this a helpful addition when not using that internal conversion.
Finally, I added two transistor-based switches that spit out +10.3V when an incoming voltage of >+2V is present. This way, my Erica Black Sequencer is able to properly gate/trigger the 2600 with the same levels as the internal midi-2-CV would. The circuit is simple (based on Graham Hinton’s suggestion here: https://www.muffwiggler.com/forum/viewtopic.php?t=193340).
R4 sets the input threshold, D1 protects against negative voltage. Positive voltage at input opens Q1, which, in turn opens Q2, passing through 10.3V as set by the voltage divider R1-R2. R5 pulls down output to ground if no gate is present.