Following The Stages Of A New Module Development Process
This is the first of a string of articles following the development of a new module in detail throughout the entire process. its an open source, diy learning/experience project to allow some who are planning a carreer in the industry get a taste of whats necessary and gain experience with no great risk of loss financially or as a reputation killer due to failiures etc. Hopefully it will bring to fruition a new, useful, desirable module that actually finds a place in peoples rigs and proves to teach those interested what they need to learn so they can move on to designi ng their own ideas and being independent.
STEP 1: CONCEPTUALISATION
The absolute most utter start point is an idea. that moment of inspiration that hits at 4am in bed when you should be sleeping but instead are reading wikipedia on 1960's british mixing desk circuitry... and suddenly 'what if this sound were available in eurorack format?' or 'everytime i see this particular module mentioned someone has this issue - it needs a little 2hp expander to solve that'. this is something that either happens for you or it doesnt - you cant learn inspiration, you cant force ideas and you cant take a pill that engnders creativity - but if you have those ideas and can bring them to fruition then every other part of this process can be learned, practised, bought or undertaken from someone in a team.
in the case of the module were are gonna follow the development, the initial idea brewed from realising how many people srtruggle with problems arising from poor clock/trigger/gate outputs/inputs from eurorack modules - there's a need for a utility that would act as a general purpose toolbox for fixing all the issues frequently encountered in one nice compact simple module. thus the raw idea of 'Lockstep' is born.
the first step to take is some very crucial research:
- Does the module already exist?
- If so, has any functionality been missed that means its still realistic to make the idea?
- If there are exisiting equivalents, and no improvements can be made, can it be done cheaper or higher quality?
- If none of the above make sense, is there any point in carrying on down this thought path at all?
its important to keep in focus what the main aim is and what you want to get out of the process - if at anypoint you are deviating from these things and going off on a tangent due to getting hyperfocussed on solving some tiny issue or making some tiny profit or not enjoying the process then you should be gaining perspective or listening to others and realising youre going wrong.
in the case of this module i quickly browsed throuhg the entire of the ModularGrid listings on utility/clock/trigger moduifier etc type modules and looked to see if they were overlapping this idea significantly and then considered that since it seeemed unique enough, that was plenty enough to justify using this as a learning process since it doesnt matter if its massively profitable or brings great success to a brand trying to attract attention etc.
the next thing to do was a basic google search t get an idea of the scope o requirements the module would have - just a basic search using keywords covering the original issues inspiring it such as 'clock, signal, weak, issues, spike, slewed, ground, etc.' . This provided a significant set of results listing peoples forum and server and manufactuirer requests for support on a number of module combinaions that suffered from relevant issues. having noted what the issues were and what fixes were suggested to solve them, it bwecame clear that the main issues were from either clocks/trigs being poorly shaped [having slewed rising edge, spikes on beginning or end, being weak, not returning to ground properly etc] or gates not syncing /being wrong length.
this suggested that there needed to be two main halves of the module - a conditioner that would take near any input signal, whether clock/gate/trig or even unexpected waveform shapes and non standard signals; and then convert them to perfect square edged trigs/clocks/gates of sufficint amplitude to make the destination module happy. the second half needed to be a path that would allow adding variable delay of a signal to sync it to something with latency issues along with something to alter length such that a gate source could be altered to satisfy needs of a destination module not getting the required gate to perform correctly.
this set of info derived from the searches that suggested the above also gave suggestions of how to fix such issues - using a Doepfer A-160-2 (sync issues), amplifier like the A-183-3 (weak clock signal), an active buffer (ground issues) or a trigger modifier like the A-162 (slewed or otherwise invalid clock signal). hence rather than having a bank of these kinda 4-6hp utilities it seemed there was a sensible avenue to make a single 6-10hp module that could do all this and preferably in a dual or multiple set up so modern sequencers and clock sources with multiple outs can be treated.
This research also sparked the idea of collecting together as many known issues of certain module or module pairings into a database that wopuld help potential users diagnose probnlems and so decide whether or not the 'Lockstep' module we were designing could help. this also makes a good sellin point for the module - a good manual with a database in it is a great source of info for the user that puts it ahed of other modules and also attracts good attention toward the idea. this resulted in just these few problems from the intial quick gooogle search:
- sq1 slewed clocks
- es pulses spikes
- marbles when supply voltage stabilizes slowly to its target voltage, something that depends on which PSU you use, where your module is located, and how other modules on the bus board behave,
- PNW seems to be one that doesn’t like to be slaved
- something with Rene?
- slaved Metropolis often needs a dedicated reset signal that is different from (earlier than) the reset bus all other slaved seqs get
- erogenous tones structure not good with clocks
Lockstep development process continues
STEP 2: BASIC DESIGN CONSIDERATIONS;
from the initial research done during conceptualistation it became clear the emost likely requirements for the module would be:
- two channels of clock conditoning that would take near any signal and produce a nice square correct output according to doepfer standards of clocks/trigs/gates. i.e. two input jacks and two output jacks.
- two channels of trig/gate delay that would allow changing timing of signals for sync purposes and alter length of high state to shift signals from a trig right through to any desired gate.
using something that is required if youre going to design modules - i.e. electronic engineering we now have the bare minimum to set out the next stage of the process - circuit design. tedchnically it would be cheapest and easiest to run up an smd based mdoule with an arduino ic doing microcontroller programmed processing - but the aim was to make an open source project people could learn from, DIY easily and do with little as possible knowledge/experience/sklills.
Hence the clock conditioners i knew were simplest done in TH by usign a cd40106b cmos logic ic - its a hexinverter with schmedit trigger, which means tho you could just build an opamp based kinda discrete custom schmidt trigger, two stages of 40106 [twice inverting to return to correct polarity] with the 40106 using a 5V supply will output a perfect square rising/falling edge clock/trig/gate, lengh dependent on input and some simple circuitry. whilst doing this initial circuit design i had a brief moment of further inspiration and thought 'it's be useful to have a choice of jumper selectable output voltages in case people needed a variety rather than just 5v. the final major part was to buffer the output to allow convenient multing with no voltage droop.
Likewise, trig./gate delaay and gate length control can be done very simply by a basic 555 timer circuit that leads on to using an ne556 ic for each channel - dual 555 ic's that therefore only need two ics to provide the whole logic necessities and then the same voltage selection and buffering circuits. to top off the initial circuitry would be needed the power [already designed from previous modules that had current overload fuse resistors, reverse polarity diodes, and decoupling caps], all the relevant power decoupling for the opamp and logic ic's and the tie offs to ground for any unused logic inputs [the 40106 ic has two spare in this instance].
having penciled out these basic necessary requirements, progress may enter basic testing and prototyping with via CAD, simulation, breadbaording or even straight to pcb if you have access to cnc machinery. personally i like to draw out pencil sketched basic ideas into a clear readable cad drawn simulator package to help others discuss and learn from and try out extra ideas/easily, change as well as simulate just for certainty and troubleshooting in parallel with breadboarding - personally i believe in breadbaording p much everything so that real life conditions can be checked to avoid the inconsitstencies and weirdnesses of reality vs. simulation can be ironed out.
the result was this:
Lockstep design process continued...
Technically some of the stuff discussed so far was done a few years back as a proposed RYO module but abandoned as not suitable for a profit driven business. As a learning process though, it is ideal, so when a chance came to involve som people who could actually contribute useful input and would be enthusiatic and inspired by it as well as truly benefit it was a perfect moment to open the floor. as a result this caused the developemnt process to return to the more advanced parts of the basic design phase and good thing it did because there was still much to be done:
PART 3: DETAILED DESIGN AND FINALISATION PRE-PROTOTYPING
having shown the current proposed idea schematics it was time to start discussing everything that could possibly be thought of prior to breadboarding - this wiould allow a full BOM to be made for ordering breadboard/prototyping parts and agree on any undecided issues that would require changes or experimentation.
- how big should the modules be - therefore dictating whethere minipots would be used [decision was minipots too frail and wobbly, as well as resulting in finger-cramping poor ergonomy].
- how many/where/what colour leds should be used to indicate any inputs, outputs, delay times, gate lengths etc [decided on input signal present led and output led that would show delay presence as well as gate length].
- shouold the module generate its own 5V supply or derive it from the busboard [decided to include internal 5v gen so those with no 5v bus presence wouldnt be inconvenienced]
- testing of delay ad gate length ranges are good - what size caps would be best ][decided also to consider a switchable cap choice so each delay/length channel could have range switches]
- are the 5v/10v suitable voltages for the output options [are they needed?if so what for?]
[Within the A-100 usually any voltage beyond about +3V is treated as "high" (e.g. +5V, +8V, +10V, +12V will work) and will trigger an ADSR or any other module with a Gate or Clock input (e.g. trigger delay, sequencer, clock divider, clock sequencer).] from 0 V to +8 V for ADSRs; Trigger, Gate or Clock Signals typical voltage levels of 0/+5 V.
- a database of known clock/gate/trig issues that are due to reset/clock/midi issues is def a good idea, i.e. note voltage/spikes/slew/gnd issues to help those who might think they need this module but it wont help,
can post soltuions given in forums to save people time - that'll make us popular as a rource and goto info source on clocks etc - to go in the back of the manual perhaps?
- Who was gonna do the breadboarding/CAD translation to get the experience under their belt and therefore plan discussions of talk through the early schems and familiarise others with the necessary info.
its also worth mentioning that some of this process is eased by prior experience from being part of a commercial industry designer - i specifically design to use as smol a number of component values as possible so that
a) it's less complex a build
b) it's less complex kit to pack, check and therefore avoid mistakes etc.
c) it costs less when bulk ordering components from factories in mass production [one change ina resistor value could change the bom cost by hundreds or even thousands on major scale]
d) tolerances of components have less overall effect because they tend to balance out when there's lots at the same tolerance
e) - it's easier for humans to build [finding the right component takes ess time the less dif ones there is] - it's easier for production pnp machines - cheaper the less reels have to be changed or the less different ones you use
f) shipping requires less packaging from component source and kit packaging needs less bags etc.
g) there's less colour codes etc for a beginner to learn
h) speeds up and reduces risk of placing a wrong value component the simpler you keep it
[ok thats all the reasons i can think of for now but i'm sure there's more]
as you can see, therefore there's a good chance you may find some things can be easily built of existing parts you have - some r, c and d values/types are ridiculously common in all electronics - way more than most others, fundamentally though there's always gonna be things you havent come across before or you dont have any of left because you only had one since they aint used much...
also certain components just plain cant be substituted - for example although this design makes it possible to change values of resistors in the output buffers to get the same amplification using a number of different combos, it's not possible to substitute the 100uf electrolytic cap on the 5v power converter - that's just straight out strictly required according to the datasheet, and if you aint got one then you buyin one! [most of the components in this thing cost like 0.1c each, it's only pots and the power cable that are around the coupla $ mark]
BOM for as yet unnamed 'clock conditioner'
type part id VALUE pcs note
---- ------------------------------ ----- ------ --------------------------------------------------------
resistors:_
r 1,2,9,13,14,27,31,43,45 100k 8 all the same pretty much…
r 4,5,6,7,9,11,12,16,17,18,19,21,
22,24,25,26,30,37,41,42,54,56,
57,59,60,62,64,66,67,69,70,72 47k 32 all the same pretty much…
r 8,20,58,68 30k 4 all the same pretty much…
r 51,52 10r 2 all the same pretty much…
r 3,15,53,63 120k 4 all the same pretty much…
r 10,23,61,71 330r 4 all the same pretty much…
r 28,32,44,46 220k 4 all the same pretty much…
r 29,36 2.2k 2 all the same pretty much…
capacitors:_
c 9,10,17,18,19,20,23,24,25,26,
31,32,33,34,35,36 100n 17 ceramic c0g 2.5mm or equiv
c 1,2,27,28 22pf 4 ceramic c0g 2.5mm or equiv
c 3,6,11,14 1n 4 ceramic c0g 2.5mm or equiv
c 4,5,7,8,12,13,15,16 10n 8 ceramic c0g 2.5mm or equiv
c 21,22 10uF 2 electrolytic, 25v, 2mm pin pitch, 5mm dia, 5mm height
transistors:_
t 1,2,3,4 BC457B 4 to-92 packacge
diodes:_
d 1,2 amber? 2 LED, 3mm
d 3,4 SB130 2 or 1N5818, schottky power
integrated circuits:_
ic 1 40106 1 cd4xxxb cmos logic hexinverter w/ schmitt trigger
ic 2,3,6,7 TL074 4 opamp
ic 4,5 NE556 2 dual timer
___
p 1,2,3,4 1M log 4 rotary pots
___
j x jacks 8 3.5mm mono, un-switched - like PJ301BM or kobicon/clone [not too important for bb]
___
[for 5v supply if not taken off bench psu]
vr x lm2391 1 5v fixed voltage regulator, to92 package
c x 100n 1 ceramic c0g 2.5mm or equiv
c x 100uF 1 electrolytic, 25v, 2mm pin pitch, 5mm dia, 5mm height
not needed for breadboarding - only for pcb
connectors:_
h x 10pin 1 10pin boxed power header
cn 1,2,6,7 3pin 4 3pin jumper header
cn 1,2,6,7 2pin 4 2 socket insulated jumper
ics 1,2,3,4,5,6,7 14pin 7 ic sockets, 14pin
___
cbl x 1 10-16pin eurorack power cable
___
also will be needed: knobs, screws, pcb, faceplate,
Lockstep development process continued...
STEP 4: A NEAR FINISHED SCHEMATIC / PROTOTYPING CONT.
to do module development requires a buncha tools and supplies - heres a list that would cover p much any possible ciruit or basic need you'd ever encounter:
TOOLS
Side Cutters
Mini 125mm Long Nose Pliers
Automatic Wire Stripper
Adjustable temperature Soldering Station
Desolder Pump/Desoldering Braid [I prefer the pump]
Solder
Stainless Steel Tweezers
"Helping Hands" type Assembly Aid, with magnifying glass
optional: Handheld magnifying glass if you can't see the heccin resistor colour codes, & diode and capacitor labelling
Digital Multimeter: measurement of Vdc, Vac, Iac, Idc, Resistance, Capacitance, audible continuity [diode test, transistor test - optional]
optional: oscilloscope - I use a rigol ds1052e, possibly the most ubiquitous scope in the industry, nay the world!
IC Extractor Tool
ICs
cd4017b decade counter used for sequencers & more
cd4066b quad switch
cd40106b hex inverter with Schmidt triggers
LM13700 transconductance amplifier - a good way to cv stuff
TL072 dual opamp
LM311 dual comparator
555 Timer multipurpose tool
PT2399 delay [echo]
RESISTORS
carbon film resistors
10R, 100R, 1k, 10k, 100k, 1M 10x [1/4W 5%]
15R, 150R, 1k5, 15k, 150k 10x [1/4W 5%]
22R, 220R, 2k2, 22k, 220k 10x [1/4W 5%]
47R, 470R, 4k7, 47k, 470k 10x [1/4W 5%]
56R, 560R, 5k6, 56k, 560k 10x [1/4W 5%]
68R, 680R, 6k8, 68k, 680k 10x [1/4W 5%]
82R, 820R, 8k2, 82k, 820k 10x [1/4W 5%]
1M5, 2M2 2x [1/4W 5%]
CAPACITORS
Ceramic Disc Capacitors 20x [+/- 0.25pF]
1pF, 10pF, 33pF, 47pF, 100pF, 470pF, 1nF, 210nF, 100nF
Electrolytic Capacitors, Radial
1uF 50V 10x [85 deg]
2.2uF 63V 10x [85 deg]
4.7uF 40V 5x [85 deg]
10uF 25V 20x [85 deg]
22uF 25V 5x [85 deg]
47uF 25V 10x [85 deg]
100uF 25V 10x [85 deg]
220uF 16V 5x [85 deg]
DIODES
1N4148 Signal Diode 4x [100V 300mA]
3.3V BZX55 Zener Diode 1x [500mW]
3.6V BZX55 Zener Diode 1x [500mW]
5.1V BZX55 Zener Diode 1x [500mW]
5.6V BZX55 Zener Diode 1x [500mW]
6.2V BZX55 Zener Diode 1x [500mW]
6.8V BZX55 Zener Diode 1x [500mW]
9.1V BZX55 Zener Diode 1x [500mW]
12V BZX55 Zener Diode 1x [500mW]
Transistors
BC547B Transistor 1x [NPN TO92]
BC557B Transistor 1x [PNP TO92]
BS170 N-channel MOSFET 1x [60V 500mA TO92]
BF245 N-Channel J-FET 1x [30V 6-15mA TO92]
HARDWARE
3.5mm Mono Jack Socket PCB mount, Vertical or Horizontal
Switched contact optional. 10x
9mm with round 6.35mm metal shaft. For pcb and/or panel mounting.
10k Linear 2x
10k Log 1x
25k Linear 1x
50k Linear 2x
50k Log 2x
100k Linear 5x
100k Log 2x
500k Linear 1x
1M Linear 1x
optional:
Selection of various Horizontal Finger Adjust trim pots.
trimmer pot adjustment tool.
Miniature Toggle Switch SPDT, 2x [5A/125Vac 2A/250Vac.]
(choose type as seems most useful: e.g. ON-ON, ON-OFF-ON etc.)
OPTO-ELECTRONICS
Selection of 10-15 various colour 3mm LEDs, any luminous intensity [mcd]
optional: NSL32 - Optocoupler with resistive output.2X [Observe polarity
H11L1M Optocoupler, High speed optoisolator with Schmitt trigger output. 6-pin DIL
PROTOTYPING
2x Solderless Prototyping Breadboard 172mm x 63mm, 840 tie points. [Multiple boards can be locked together for larger prototyping areas.]
Approx 100 male/male jumpers for use with breadboards, various lengths with solid tips.
…
FOR POWER
You'll with need a lab/bench-top type psi, or alternatively, you can get a smog eurorack psi like the tiptop happy ending kit [hek].
The lab psi's have pro's: stable, accurate power to be dialled in at whatver voltage and amperage you choose
The hek has the pro's: provides actual real life eurorack module power to see the effects of actual setups effects.
Bonus pro's: holds maybe 7-10 eurorack modules in a little desktop skeleton rack, allowing a set of vco/vca/eg/lfo/filter/etc as test modules!
(See how your breadboard circuits behave with and respond to real life eurorack modules!)
FOR RYO PROTECTED BREADBOARD POWER [optional:]
Solderless Prototyping Breadboard 80mm x 60mm, 390 tie points: 29 columns in 2 banks of 5 plus 4 rows of 25
if using an hek or similar to power the breadboard:
8-Way single red-stripe ribbon cable approx 30cm length,
female 16-pin IDC connector,
(optional, and highly convenient: idc crimping tool).
if using a bench-top/lab psu to power the breadboard:
lab/bench-top PSU,
Approx 30cm of 4 x Dif colour solid core wire to connect psi to the breadboard.
10r Resistors, carbon film 2x [1/4W 5%]
5x 100nF Ceramic Disc Caps 5x [+/- 0.25pF]
2x 10uF Electrolytic Caps, radial 2x [85 deg]
1x 100uF Electrolytic Cap, radial 1x [85 deg]
LM2391 Voltage Regulator 1x [TO92 package]
2x 1N5818 Diodes Schottky Rectifier Diode. 2x [30V 1A 0.5Vf]
Lockstep Development Process Continued...
STEP 4: A NEAR FINISHED SCHEMATIC / PROTOTYPING CONT.
after some discussion here's the current state of affairs ready to be breadboarded when components arrive:
