M O D E L 1 0 1 1

D i s c r e t e V o l t a g e C o n t r o l l e d O s c i l l a t o r

Construction & Operation Guide

R E V A - F O R P C B V 1 . 1

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

S P E C I F I C A T I O N S

PHYSICAL

FORM FACTOR: Loudest Warning / 4U

WIDTH: 3NMW / 75.5mm

HEIGHT: 175mm

DEPTH: ~40mm from panel front inc. components

PCB: 70 x 150mm, Two-Layer Double Sided

CONNECTORS: 4mm Banana

IDC power connector pinout. ELECTRICAL

POWER: +12V, 0V, -12V

CONSUMPTION:~40mA +12V Rail, ~30mA -12V Rail

CONNECTOR: IDC 10-pin Shrouded Header, Eurorack Standard or MTA-156 4-Pin Header

I/O IMPEDANCES: 100K input, 1K output (nominal)

MTA-156 power connector pinout. INPUT RANGES (nominal) 1V/OCT: +/- 10V

FM: +/- 5V

LOG: +/- 5V

SYMMETRY: +/- 5V

SYNC: +/- 5V (falling-edge trigger)

OUTPUT RANGES (nominal)

OUTPUT A: +/- 5V

OUTPUT B: +/- 5V

SUBOCTAVE: +/- 5V

Specifications 2

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

T A B L E O F C O N T E N T S

SPECIFICATIONS Specifications / Power Requirements 2 INTRODUCTION Introduction 4 CIRCUIT OVERVIEW Circuit Overview 5 Exponential Converter 6 Sawtooth Core 8 Triangle / Shapers 10 Pulse / Suboctave Shapers 10 Output Mixers / Amplifiers 12 CHOOSING COMPONENTS Bill Of Materials (BOM) 14 Choosing Components 15 Transistor Matching 16 CONSTRUCTION Construction Overview 18 Physical Assembly 20 CONTROLS Controls 21 CALIBRATION Calibration Overview 22 CV Scale 23 CV Offset 24 High Frequency Compensation 24 Triangle Adjustment 25 REFERENCE PCB Guide - Lower Board 26

PCB Guide - Upper Board 27 This document is best viewed in dual-page mode.

Circuit Overview 3

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

I N T R O D U C T I O N

The Slightly Nasty Model 1011 is a voltage controlled oscillator that's a little bit different. Despite featuring a host of functionality including four mixable waveshapes, suboctave, linear and logarithmic FM, pulse width , and hard sync, inside it you won't find a single IC opamp or OTA. What you will find is no less than 41 discrete transistors flying in close formation, doing their best to output useable musical tones.

The Model 1011 has been designed from the ground up to use modern "jellybean" components that can be cheaply and easily obtained from most electronics suppliers. Despite the unusual implementation, the architecture is actually a very traditional sawtooth-core design that will be familiar to most people who have worked on VCOs before.

Three outputs provide mixable sine-triangle, saw-pulse-suboctave, and suboctave square respectively, the pulse also featuring both manual and CV-controlled symmetry (pulse width). Aside from the usual 1V/Octave input, there are also separate inputs for both linear and logarithmic FM, each with input attenuators, as well as a hard sync input. The exponential converter is temperature compensated for better thermal stability and the sawtooth core features high-frequency compensation for better pitch tracking.

The Model 1011 uses the Loudest Warning 4U format for the front panel, and follows Eurorack electrical and power standards. All front panel components are PCB mounted for easy wiring-free construction. The front panel is available in two finishes - satin anodised and gloss white powdercoat, both on 2.5mm aluminium with robust UV-printed graphics.

Introduction 4

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

C I R C U I T O V E R V I E W

For full schematics, please download the separate schematics PDF. Excerpts shown in this manual may be outdated and are provided for reference only.

While the fully populated PCB of the Model 1011 can look quite intimidating, the circuitry can actually be broken down into a set of relatively simple subcircuits that each handle a very specific aspect of the module's operation. Overall, the 1011 has a fairly standard architecture consisting of the following units:

1. Exponential converter - this allows the use of 1V/Octave pitch CVs by taking a linear scale voltage from the CV input and converting it into an exponential scale current to feed the sawtooth core.

2. Sawtooth core - this is the sonic heart of the module, generating the base sawtooth signal from which all other waveshapes are generated. Sync is also implemented in this circuit.

3. Triangle/sine shapers - These convert the raw sawtooth signal into triangle and sine by first folding the sawtooth into a triangle shape, and then soft-clipping that to create a pseudo-sine.

4. Pulse/suboctave shapers - These create the by feeding the sawtooth signal into a comparator, using the symmetry controls to set the threshold level. The pulse is then used to clock a pulse divider to form the suboctave square.

5. Mixers/output amps - These allow the blending of the various as well as converting the different levels and offsets of the various raw signals to match the +/-5v expected at the outputs.

Block diagram of the Model 1011. Circles marked "A" are attenuators.

Circuit Overview 5

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A Circuit Overview 6

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 21 02 13 11 A s y m mDei ts rcirce tSel eOws cLiilml ai t oe r

E X P O N E N T I A L C O N V E R T E R

Undoubtedly the finickiest part of most VCOs, the exponential converter in the VCC 8 Model 1011 is essentially a discrete reimplementation of the opamp-stabilised 3 + V+ 1 transistor pair found in countless other designs. This circuit works by using the 2

- V- GND naturally exponential relationship of a transistor's base-emitter voltage to its 4 output current, using two matched transistors to mostly cancel out each others' VEE R3 thermal effects and keep the conversion stable across different temperatures and 3 3 currents. A feedback-stabilised current source on the shared emitters of the 1V/OCT R1 1 1 2 2 transistors holds one transistor at a constant current, causing the exponential GND TEMPCO current caused by changes to the input voltage to appear at the collector of the R2 other one. A temperature-sensitive "tempco" resistor provides additional GND correction to the aspects of the circuit's thermal response that are not cancelled VEE TO_CORE by the matched pair.

The exact operation of this sort of converter is a bit too involved to get into in this Traditional configuration of PNP exponential converter with opamp manual, but an excellent rundown of the basic principles can be found on René current source. Schmitz' website at http://schmitzbits.de/expo_tutorial/index.html

In the 1011, the exponential converter can be broken down further into three basic sections. There are the frontend buffer/amplifiers that combine the various CVs and panel controls into a single pitch voltage; the exponentiator itself, in the form of the matched pair; and the feedback controlled current source, which consists of a differential pair controlling a current source tranistor. The bulk of the exponentiator is single rail and works between 0v and +VCC.

The input buffer/amplifiers are essentially just crude emitter followers, and consist of transistors Q501, Q502, and Q510 along with their respective passive components. The output of Q501 and Q502 are both combined and go through the voltage divider comprised of RV505 and the tempco resistor R522, in order to reduce the level to the small voltage swing needed for the exponentiator. Because the circuit is single rail, Q503 provides a buffered offset voltage so that the resultant scaled CV is centred near the 1/2 VCC mark. The scaled CV is finally buffered by Q506 before being fed into the exponentiator at Q507.

Q507 and Q509 comprise the matched-pair exponentiator, and share a common emitter. Q507 takes the scaled pitch CV as input at its base, while Q509 has its base held at a fixed voltage around 1/2 VCC. The exact voltage at Q509's base can be trimmed with RV506 in order to offset the CV response and get the desired centre frequency for the oscillator (usually middle C).

Finally, the differential pair of Q504 and Q505 along with the current source transistor Q508 form the feedback-stabilised current source, which would normally consist of an opamp in a circuit of this type. Q505 is referenced to 1/2 VCC via the voltage divider of R513 and R514, and like an opamp the circuit will try to get the opposite input (the base of Q504) to match this level. That base is connected to R518, on the collector of Q507, and so the circuit will try to hold the voltage across

Circuit Overview 7

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

R518 at 1/2 VCC - and consequently maintain a constant current through both it and Q507. It dœs this by controlling the current that feeds the exponentiator pair through Q508. C503 performs a similar role to bypass found caps in opamp feedback paths - preventing oscillations that can develop due to various phase effects. The buffered linear FM input is AC-coupled through C501 and feeds directly into the base of Q504 along with the feedback signal.

S A W T O O T H C O R E

The sawtooth core in the Model 1011 basically consists of a timing capacitor with a discharge FET across it, and a reset comparator. The core of the reset comparator is formed by Q204, Q205, and Q206 - with the first two once again forming a differential pair and the latter serving as the gain stage / output. One input of the comparator (the base of Q205) is connected to the timing capacitor, and the other (the base of Q204) is fed the reset threshold voltage set by the voltage divider formed by R206 and R205 (these are chosen to get a ~6v P-P on the sawtooth).

Q202 and Q203 are used to pull down the threshold to a lower voltage when activated, in order to implement the hysteresis needed in a relaxation oscillator. When the capacitor passes the threshold voltage, the comparator's output gœs Voltage across timing capacitor (top) high and simultaneously switches on the discharge FET Q207 and the threshold vs. voltage on the base of disharge FET Q207. Blue shows charging pulldown transistor Q203. This means that the capacitor now needs to discharge period and red discharging. Note down to the new, lower threshold voltage before the comparator output gœs back that the discharge time is exaggerated for clarity, and in most to low and completes the cycle. cases can be considered virtually instantaneous. The second threshold pulldown transistor Q202 (and its inverting input buffer Q201) is dedicated to the sync input, and triggers a reset cycle whenever a sufficiently powerful falling edge triggers it. The necessity of a second dedicated pulldown transistor for this is due to the possibility of the comparator being knocked into an undesirable region of operation where instead of acting as a comparator, the feedback path formed through Q203 causes the circuit to turn instead into a voltage follower, tracking the threshold voltage and locking up the oscillator. To prevent this, the feedback path has to be kept from reaching the metastable point near the threshold voltage, and so the sync input is given its own transistor outside the feedback loop.

RV202 and D201 form the high-frequency compensation circuit (a.k.a Franco compensation). This works by using the voltage developed across RV202 by the charge current to trigger the reset slightly earlier as the current increases (and with it the pitch). This means that higher frequency cycles are shortened slightly, and thus increased in pitch to counteract the droop caused by other effects in the circuit, such as capacitor discharge time and so on. D201 limits the maximum compensation effect of the circuit, to prevent excessive drops in sawtooth

Circuit Overview 8

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A Circuit Overview 9

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

amplitude at very high frequencies where pitch is less discernable.

Q208 and R213 are the high-impedance output buffer for the sawtooth core, and prevent the downstream circuits draining charge from the timing capacitor and affecting the frequency.

T R I A N G L E / S I N E S H A P E R S

The triangle shaper in the Model 1011 takes advantage of the normally undesirable behaviour of an inverting transistor amplifier when the input and output signals "cross over". Because the transistor's base voltage can't be higher than its collector (in the case of an NPN transistor), once the output at the collector drops too low it "collides" with the base voltage and can't go any lower - causing the output to follow the base voltage instead. We can use this to "fold" the sawtooth over on itself to create a just by offsetting the input sawtooth by the right amount.

Q301 is our inverting unity-gain amplifier, which is AC coupled to the sawtooth signal so that the input can be offset by the resistors R301 and R302. C301 helps to shape the inescapable glitch in the triangle at the sawtooth's reset point so that it can be smoothed out more effectively by the lowpass filter R305/C302. The pair of Principle of triangle formation by inverting amplifiers at Q302 and Q303 amplify the triangle signal to the desired folding . Intermediate stage is shown for clarity, in reality amplitude. this is a single-stage process. The triangle signal is attenuated through the R312/R313 voltage divider before being soft-clipped by the pair of diodes. This distorts the triangle into something approximating a sinewave, which is then amplified back up to useable levels by Q304 and Q305.

P U L S E / S U B O C T A V E S H A P E R S

The pulse waveform of the Model 1011 is generated in the same way as most VCOs - by feeding the sawtooth signal into a comparator and varying the threshold voltage to implement pulse width control. The basic circuit is the same three- transistor comparator used elsewhere in the module, taking the sawtooth signal as one input and the summed "symmetry" CV and front panel voltages as the other.

The suboctave is slightly different to the other waveshapes, in that it isn't formed by just shaping the sawtooth in some way. Instead a classic two-transistor Pulse formation with comparator, multivibrator circuit is clocked from the positive-going edges of the pulse showing PWM via varying threshold. waveform, creating a square output at half the frequency. C601 and Q604 convert

Circuit Overview 10

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A Circuit Overview 11

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

the pulse signal into buffered positive-edge pulses, feeding it into the multivibrator circuit via C604 and C607.

Both the pulse and suboctave circuits output unipolar 0v to VCC waveforms, in the case of the pulse this isn't an issue as it is eventually mixed directly with the likewise unipolar sawtooth wave, however the suboctave is AC-coupled through the large 10uF capacitor C610 in order to centre it around 0v.

Suboctave formation, showing the O U T P U T M I X E R S / A M P L I F I E R S positive edge pulses generated by C601 / Q604 causing the multi- vibrator to flip state. At this point, all of the various waveforms in the Model 1011 are at various and offsets, and the role of the mixers and output amplifiers is to combine these disparate elements and ensure that the final outputs are in the +/- 5v range expected in most modular systems.

The two amplifiers are built around what are essentially discrete op-amps, comprising a differential pair for input, a single transistor gain stage, and a two- transistor push-pull output. The suboctave signal gœs through a much simpler push-pull output buffer that dœsn't need to worry about linearity or crossover distortion on account of it being a purely squarewave signal.

Amplifier A takes the triangle and sine signals which are already at matching levels and combines them via the mix pot RV403 before feeding them into the amp's positive input. The amp output is fed back into the negative input via R418, and so the circuit operates as a standard voltage follower opamp circuit.

Amplifier B is only slightly more complex, it has an additional network of voltage dividers before the mix pot to match the levels of the sawtooth and pulse signals, and the suboctave signal is fed into the negative input via an attenuator. Because the negative input is no longer just connected directly to the output feedback, the gain of the amplifier actually changes as the suboctave attenuator is adjusted, in order to keep the output level within +/-5V regardless of how much suboctave is added.

Finally, all outputs go through a 1K output resistor to protect the amps and buffers from short-circuits and provide the expected 1K output impedance.

Circuit Overview 12

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A Circuit Overview 13

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A B I L L O F M A T E R I A L S

RESISTORS 10R 2 R101, R102 240R 1 R523 390R 1 R519 1K 7 R311, R318, R320, R417, R420, R424, R524 1K 3300PPM/C 1 R522 1K2 2 R306, R520 2K2 11 R207, R208, R213, R307, R310, R321, R511, R515, R521, R614, R617 3K3 2 R401, R603 5K1 2 R504, R507 9K1 2 R206, R405 10K 29 R203, R204, R205, R210, R211, R214, R215, R317, R319, R404, R409, R410, R413, R414, R509, R513, R514, R516, R517, R518, R525, R526, R602, R608, R610, R611, R612, R615, R616 12K 2 R402, R510 15K 1 R313 22K 2 R209, R505 33K 1 R305 51K 6 R506, R508, R512, R607, R411, R412 62K 1 R606 91K 1 R316 100K 19 R201, R202, R212, R303, R304, R312, R314, R403, R406, R407, R415, R418, R422, R425, R426, R427, R501, R613, R618 150K 2 R601, R604 200K 4 R301, R421, R502, R503 270K 1 R309 300K 1 R605 560K 1 R308 1M 6 R302, R322, R408, R416, R419, R423

CAPACITORS 10pF 2 C402, C403 33pF 2 C301, C503 220pF 1 C302 1nF 3 C201, C604, C607 10nF 1 C601 10nF (C0G/NP0) 1 C206 (Optionally use standard 10nF film capacitor) 100nF 9 C202, C203, C305, C401, C404, C405, C501, C502, C602 0.68μF 2 C303, C304 10uF Electrolytic 2 C306, C610 100uF Electrolytic 2 C101, C102

SEMICONDUCTORS 1n4148 7 D201, D401, D402, D403, D404, D601, D602 1n4007 2 D301, D302 BC550C / BC547C 26 Q201, Q202, Q203, Q204, Q205, Q301, Q303, Q304, Q401, Q402, Q403, Q404, Q407, Q409, Q411, Q501, Q502, Q503, Q504, Q505, Q510, Q601, Q602, Q604, Q605, Q606 BC560C / BC557C 11 Q206, Q302, Q305, Q405, Q406, Q408, Q410, Q412, Q506, Q508, Q603 MATCHED BC560C / BC557C 2 Q507, Q509 2n7000 2 Q207, Q208

POTENTIOMETERS 1K 25-turn 1 RV506 5K 25-turn 1 RV202 10K 25-turn 1 RV505 10K Linear 1 RV502 100K Linear 8 RV401, RV402, RV403, RV501, RV504, RV503, RV601, RV602 100K 25-turn 1 RV301

CONNECTORS Banana Socket 8 P102, P103, P104, P105, P106, P107, P108, P109 IDC 10-pin Header 1 P101 (Option 1) MTA-156 4-pin Header 1 P101 (Option 2) 10-pin 2.54mm pin 3 Use standard breakaway pin strip. header 10 pin-2.54mm female 3 pin header

HARDWARE M3 x 20mm Screw 4 M3 Washer 16 M3 x 10mm Threaded 4 Metal Hex Spacer M3 Nut 4

Bill Of Materials 14

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 21 02 13 11 A s y m mDei ts rcirce tSel eOws cLiilml ai t oe r

C H O O S I N G C O M P O N E N T S

Selecting the right components for the 1011 is fairly straightforward, with only a couple of parts needing any special attention. All resistors should be 1% tolerance metal film types, most capacitors are standard rectangular film caps and electrolytics. Three types of transistor are used, the bulk being BC550C and BC560C (Edit: As these have now been discontinued, they can be substituted with BC547C and BC557C respectively), with a pair of 2N7000 FETs used in the sawtooth core. Diodes are mostly the ubiquitous 1n4148.

In the exponential converter there is a 1K 3300ppm/C tempco resistor that will need to be bought from a supplier that specialises in synthesiser components, such as Thonk (www.thonk.co.uk) or Modular Addict (modularaddict.com), among others. The exponential converter also requires a matched pair of the BC560C transistors (see the section on transistor matching over the page), which can be selected from your inventory of transistors using a simple matching circuit.

In the sawtooth core, the main timing capacitor responsible for generating the sawtooth wave can be either a normal film capacitor, or a more thermally stable part if greater stability is desired. Traditionally, polystyrene caps were used for this role in VCOs, but as these are now becoming rare and expensive a much better option is one of the new generation of C0G/NP0 ceramic capacitors.

The sine shaper uses a pair of 1N4007 diodes instead of the usual 1N4148s to get a slightly better sine shape, though 1N4148s will work also.

The module is designed to use either side or top-adjustment 25-turn trimpots for calibration adjustment - side adjustment is usually the better option as it means the unit can be more easily calibrated when connected to the rack's power bus.

The front panel PCB fits Alpha brand 9mm vertical-mount round shaft potentiometers, these are widely available from stores such as Thonk, Tayda, Smallbear, Mouser etc. The module should fit a number of different banana jack sockets, but the "correct" parts are the Cinch Connectivity range of jacks.

The intended knobs are Davies Molding parts - the 1913BW, 1910CS, and 1900H - though given the outrageous pricing of the actual Davies 1900H I'd strongly recommend using a good quality clone. Avoid the cheaper clones without an internal brass bushing - Thonk sells an excellent brass-bushed 1900H clone for a very reasonable price that I use in all of my own builds.

Alternatively, feel free to use any knobs that have similar diameters and will fit the Alpha round shaft pots. The Davies parts are 29mm, 19mm, and 13mm respectively, and many other manufacturers make knobs of similar sizes. The classic silver top Moog-style knobs actually work quite well also for the larger diameters.

Choosing Components 15

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

T R A N S I S T O R M A T C H I N G

The 1011 uses a pair of matched BC560C/BC557C PNP transistors in the exponential converter to ensure a stable and reliable conversion across different temperatures and pitch ranges. These transistors need to be matched to ensure that they have the same VBE (base-emitter voltage drop) at a given temperature, which requires testing a number of transistors to find ones that have the closest Vbe values.

A common mistake made by inexperienced builders is to match the transistors using a multimeter transistor tester, using the transistors that show the best matching values. This will not work. The transistor tester built into many cheaper multimeters measure the hFE (current gain) of the transistor, and not the base-emitter voltage that we are interested in. Testing for VBE requires setting up or building a small test circuit to allow measurement of the difference in VBE between transistor pairs.

Recently a number of small, cheap component testers have appeared on the market that do measure VBE, however while these are handy to roughly check component values and find faulty parts they do not have the resolution or accuracy required for matching exponential converter pairs.

There are a number of circuit designs available for matching transistors, but I personally recommend the Ian Fritz method for its simplicity and reliability. There are a few variations on this method, but the circuit I use is shown here. Essentially it consists of setting the transistors up as diodes with precisely matched resistance on the emitters of each (using a 25-turn trimpot to zero out the tolerance errors of the 100k resistors), then measuring in millivolts the difference between the emitter voltages of the two transistors. The switch shown here swaps the resistors between the two transistors to allow the trimpot to be accurately set. I'd strongly recommend building a socketed version of this circuit on stripboard, to keep on the workbench for future projects that need matched transistors.

When testing transistors I recommend setting up a fan blowing across the test circuit, to ensure that both transistors are kept at an identical temperature. It's also worth leaving each pair for a couple of minutes to allow the transistors' internal temperatures to stabilise. If the temperature in the room is relatively stable, you can speed up the process by leaving one transistor in the circuit permanently, and swapping out the other position one by one, taking note of the voltage difference of each tested part. Once you've found a few transistors that seem to show very close or identical differences to the fixed "reference" transistor, you can take the reference transistor out and test the rough-matched pairs against each other as normal to find the ones that have the closest match. Even if you don't test all the rough-matched transistors, keep them together for future projects, because searching through labelled pairs that are already fairly close is a lot faster than finding matches between random parts!

Transistor Matching 16

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A PNP VERSION

+12V TO SET TRIMPOT: With a pair of transistors fitted, measure the voltage difference while switching between the 100K two switch positions. Adjust the trimpot until the voltage is the same in both positions. 100K 100K 1 3 1 3

DPDT SWITCH Set multimeter to mV range when measuring.

2 2 Polarity is not important as long as it's kept the same when testing multiple transistors.

MULTIMETER

MULTIMETER 3 3 E E B B 2 2 TRANSISTORS UNDER TEST 1 1 C C

0V/GND

NPN VERSION

+12V 2 2

1 1 3 3

MULTIMETER

MULTIMETER 2 2

This is essentially the exact same circuit 1 3 1 3 but with the power reversed and the transistors installed accordingly. 100K 100K

100K

0V/GND

Transistor Matching 17

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

C O N S T R U C T I O N

For the most part the 1011 can be constructed like any other PCB project, but there are a couple of special components that need consideration. The exponential converter uses a matched transistor pair and a 3300PPM/C tempco resistor to achieve a good degree of thermal stability, and these need to be mounted together in a specific way in order to ensure that they are all in close contact and share the same temperature during operation. See the section labelled Transistor Matching for details on how to find a pair of matched transistors, the 1K 3300PPM/C tempco can be bought at synth part suppliers such as Thonk (www.thonk.co.uk) and Synthrotek (store.synthrotek.com) among others.

The majority of construction can be performed like any PCB build, starting with the lowest-profile components (resistors and diodes) and working through to the taller ones (Capacitors, transistors, etc.). The simplest way to populate the board is simply to work through the BOM, doing each component type and value in one chunk before moving on to the next. Avoid fitting the special components for now (Q507, Q509, and R522)

Given the unusual number of discrete transistors in the build, it's worth commenting on how to best populate them without risking damage or ending up with a motley forest of strangely angled TO-92 packages. My preferred technique is to put a batch of the transistors in place and bend the outer legs as usual, taking care to get the height roughly the same between each, and then soldering only the centre leg of each. Once these are all done, flip the board over and use a pair of tweezers to straighten each transistor until they all look correct. Flip the board back over and then solder one of the remaining legs of each of the transistors, then finally go through and solder the final legs once all those are done. This way each transistor gets the chance to cool down between each joint being soldered, which reduces the risk of damage.

When soldering transistors it's important to hold the iron long enough to get a solid joint that extends down into the plated hole, but not so long that you risk thermal damage to the transistor junction. With a properly heated iron, a few seconds on each should be all that's required.

When soldering rectangular capacitors, I like to solder one leg on each, then hold the board in one hand while applying a very light pressure on top of the capacitor with a free finger, using the other hand to reheat the solder joint until the capacitor slides down tight against the PCB's surface. Continue this process for all the installed capacitors then go back and solder the remaining legs. This approach also works well to mount other components that need to mount securely onto the board, such as trimpots, IC sockets and pin headers.

Care must also be taken to ensure that the PCB-mounted potentiometers are mounted as vertically as possible on the board - one option is to click the potentiometers into place, then mount them to the front panel before soldering

Construction 18

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

them. Also note that most potentiometers have a small anti-rotation tab on them that will need to be removed before soldering them into position, these can be cut off with a sharp pair of sidecutters, and I personally like to clean up any remaining protrusions with a few passes of a needle file as well.

The pin headers that interconnect the two boards are another component that needs a bit of additional care when assembling to ensure correct aligment. The best course of action is to solder one side of all the interconnects (either the pins or socket) into place, being careful to get them straight and flush with the board. Then connect the other halves onto them, lay the other PCB in place over the top (I would even recommend mounting the boards together with screws and spacers as they will be when finally assembled), and solder all the pins of the other side. Once they are all soldered, carefully separate the two boards, taking care to not bend the headers in the process.

When fitting the matched transistors and tempco resistors, these need to be thermally connected to ensure the best stability. The two BC560Cs should be joined face-to-face with a band of heatshrink tubing (I also like to smear a very thin layer of thermal compound between the two, making sure none gets near the conductive legs). Carefully bend the legs with a pair of tweezers so that they match the hole spacing on the PCB, and solder them into place. Once the transistors are installed, the tempco resistor can be mounted on top, using something like an epoxy or liquid electrical tape to keep it thermally coupled to the transistors and insulated from ambient temperature changes.

Assembly of the matched pair using A NOTE ON POWER FILTERING heatshrink tubing. After fitting the tempco resistor, a covering of non- It's common practice among some builders to replace the 10 ohm power conductive material should be added to filter resistors with ferrite beads, in the belief that this will prevent power thermally insulate the assembly. rail fluctuations under varying current loads while still providing the filtering action desired. This is not recommended. Ferrite beads do not even begin to show reactivity until somewhere up around the 1MHz mark, an order of magnitude beyond the audio range. Within the audio band (and for a long way beyond it) they are electrically identical to a wire jumper.

Construction 19

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

P H Y S I C A L A S S E M B L Y

Assembling the finished PCBs and front panel is very simple. Begin by fitting the M3 hardware to the panel-side PCB. screwing the hex spacer tight to hold it all together. Once all four screws are in place, start fitting the banana sockets into their respective holes on the front panel - making sure to align the flat terminals vertically (if using the Cinch-style sockets). The banana sockets need to be tightened solidly to prevent them coming loose in use, something like a dab of hot glue between the nut and thread can also help prevent loosening.

Make sure that the nuts and washers have all been removed from the PCB-mount potentiometers on the front panel PCB, as well as the anti-rotation tabs on the pots themselves (if present). Now you can join the front panel and panel PCB by pushing the pot shafts through their respective holes, fitting their washers and Connection of the two PCBs using nuts, and tightening everything into place. standard M3 hardware. Washers are necessary on the inside to correctly Now you'll need to connect the banana sockets to the front PCB using either space the boards for the interconnects. Screw head should go some offcut component leads, or tinned copper wire. The simplest way is to on panel side. solder the straight pieces of wire vertically into the pad on the PCB, then bend them over to meet the banana socket and solder that end to the flat side of the terminal. This way they can be easily disconnected for servicing by simply heating the terminal with the iron and pushing the wire away once the solder reflows.

Once the sockets are all connected, put M3 washers on all four mounting screws and carefully fit the second PCB into place - taking care to get the interconnects correctly seated. Until calibration is completed I would not fit the final washers and nuts to allow easy separation of the PCBs when troubleshooting, just making sure to take extra care plugging and unplugging the power connector when the PCB is unsupported.

When the module is confirmed to be working properly you can fit the final M3 washers and nuts and tighten up the whole assembly. Double check that the hex spacers haven't loosened in the meantime as well.

Connecting the banana sockets using an offcut component lead or similar.

Construction 20

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 21 02 13 11 A s y m mDei ts rcirce tSel eOws cLiilml ai t oe r

C O N T R O L S

FREQUENCY CONTROLS Fine and coarse adjustment of inital oscillator pitch.

MIX KNOBS Crossfades the respective output between two waveshapes

A & B OUTPUTS SUBOCTAVE OUTPUT Outputs mixed sine-triangle (A) and Outputs the raw suboctave signal. mixed saw-square-suboctave (B)

SUBOCTAVE LEVEL SYMMETRY Controls the amount of suboctave Sets the inital symmetry (or pulse that is mixed into output B width) of the pulse waveform. Centred is 50:50 squarewave.

INPUT ATTENUATORS Allow 0-100% attenuation of the FM signal, Symmetry CV and Log CV. INPUT JACKS AC coupled inputs for Linear FM and Sync signals, and DC coupled inputs for 1V/Octave pitch CV, Logarithmic FM, and Symmetry CV.

SLIGHTLY NASTY JACK COLOURS RED Bipolar signal output BLUE Bipolar signal input YELLOW AC-coupled input BLACK Logic output WHITE Logic Input

Controls 21

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

C A L I B R A T I O N

Calibration of the 1011 consists of adjusting the four calibration trimpots on the back of the module to set the following values (in order):

1. CV scale - sets the scaling of the pitch CV to ensure that a 1v change in CV produces a one octave change in pitch.

2. CV offset - sets the centred "zero point" for the front panel frequency knobs.

3. High frequency compensation - this allows you to "boost" the CV response at higher frequencies to compensate for the tendency of VCOs to "droop" at higher pitches.

4. Triangle wave alignment - This sets the folding point of the saw-to- triangle shaper to make sure that the reset point of the wave lines up correctly and forms a nice uninterrupted triangle wave.

BEFORE YOU BEGIN

Before powering up the module for the first time, use a multimeter to check the resistances between the three power rails. Make sure that they show a resistance higher than 1KOhm, any lower and it's possible there is a short circuit or incorrectly oriented semiconductor somewhere on the PCB.

If the 1011 dœsn't oscillate when you power it up, first try adjusting the CV offset trimmer in both directions - this trimmer is quite sensitive and can easily push the oscillator outside of its operating range.

Before calibrating the CV response, allow the oscillator to warm up for a few minutes - the frequency will drift in this period as all the components settle into their operating temperatures.

While I've given a specific order to these operations, you can expect to have to go back and forth on some of them, particularly the CV Scale and CV Offset calibration. Also if you notice the oscillator pitch seems way too high or low when you get to the CV Scale step, feel free to adjust the CV offset control to get it in the right place.

Also, you'll want to disable the High Frequency Compensation before you start the CV calibration steps, which means measuring the resistance across the two outer pads of the "HF.COMP" trimmer and adjusting it until it reads 0 ohms (or as low as it will go).

Calibration 22

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

C V S C A L E

he goal with this step is to get the oscillator to respond with an accurate 1V/ Octave response to the frequency CV - so that a change of +/- 1V on the input results in a +/- 1 octave change in the output pitch (ie. doubling or halving of its frequency). We're not really worried about absolute pitch here, only that the amount of change relative to the CV is correct.

Getting the CV scale right is always one of the more tedious jobs when calibrating oscillators, and different people have developed various systems over the years to get the job done. However you choose to do it, I would strongly recommend using whatever 1V/Oct source you will be using when the module is completed, such as a your midi-CV converter or a keyboard with 1V/Oct output.

A first basic step is to either hook up a frequency counter or instrument tuner that shows frequency (if you have one) and your listening system, and play notes on the keyboard that are one octave apart, somewhere around middle C. Adjust the CV Scale trimmer until the resulting pitches from the oscillator are as close as you can get to being one octave apart (the higher note should be double the frequency of the lower one). Because the actual frequencies of both the notes will be changed each time you adjust the trimmer, just be sure to play them each a couple of times between each adjustment to determine what the relationship between them currently is.

Once you're more or less happy with the response over one octave, try playing notes that are further apart, such as the next octave down from your low note, and the next octave up from the higher one. Once again adjust the trimmer until you get the correct relationship between the notes - in this case the high note should be 8x the frequency of the lower one. Make sure to occasionally go back and check the notes that are closer together again, to make sure that they're also staying in calibration (they should if the exponential converter is providing an accurate conversion).

Get the response as accurate as you can, but don't obsess over it yet, because you'll want to fine tune this a little further once the CV offset has been trimmed.

Calibration 23

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

C V O F F S E T

While this adjustment is quite critical in a keyboard synth where the oscillators are expected to have a very specific voltage-pitch response, in a modular where we've got big frequency knobs on the front panel to adjust the pitch, it's really just a convenience. Essentially all you want to do here is apply 0v to the 1V/Oct input (even running a alligator lead from one of the M3 pcb mounting screws to the tab on the back of the banana jack will suffice), set the coarse and fine frequency knobs to their centre positions, and then adjust the CV Offset trimmer until the oscillator is outputting middle C (261.6Hz). Don't worry about getting this exact, because tiny movements of the coarse tune knob will throw this off substantially, and tuning oscillators is a completely normal task when patching modulars to play melodically. This setting just makes sure that similar knob positions on individual oscillators give consistent frequency ranges.

H I G H F R E Q U E N C Y C O M P .

Once the oscillator is responding fairly accurately in the low-mid pitch range, it's time to set up the high-frequency compensation. This is necessary because various electrical effects in the circuit usually cause the 1V/Oct response to "droop" at higher frequencies, meaning that notes will get progressively flatter and flatter (too low in pitch) as you continue up the musical scale. The high frequency compensation circuit adds a boost to the oscillator frequency as the frequency increases, to counteract this drooping and restore the expected 1V/Oct response.

Setting this up generally just consists of playing notes higher up on the scale to see how flat they are, and slowly turning up the HF compensation trimmer until their pitch is adequately corrected. It's worth also playing notes at lower pitches while you're adjusting to make sure that the adjustments aren't upsetting their calibration.

The 1011 generates the triangle wave by folding the sawtooth onto itself, and to

Calibration 24

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

T R I A N G L E A L I G N M E N T

form a smooth and uninterrupted triangle the fold point must be set accurately. Turn the front panel sine-triangle mix knob all the way to the right and scope the output - you should see the triangle wave with a slight glitch on the topmost corners. Use the frequency knobs to set the oscillator's frequency to something comfortable like 200Hz or so, then adjust the trimmer until the glitch looks to be as central in the wave as you can get it.

Once you're happy with how it looks, hook up the output to your listening system and fine tune the trimmer by ear until you find the position where the triangle has the least upper (this is wherever the triangle sounds the smoothest and has the least "buzz" to it).

Triangle alignment. Centre image is correctly aligned.

Calibration 25

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

P C B G U I D E - L O W E R

LOWER BOARD - TOP LOWER BOARD - BOTTOM

Reference 26

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A M O D E L 1 0 1 1 D i s c r e t e O s c i l l a t o r

P C B G U I D E - U P P E R

UPPER BOARD - TOP UPPER BOARD - BOTTOM

Reference 27

S L I G H T L Y N A S T Y E L E C T R O N I C S A D E L A I D E , A U S T R A L I A T R I A N G L E A L I G N M E N T

w w w . s l i g h t l y n a s t y . c o m