AVR-Based Synth

After finishing the Drumsynth Prototype i started another little synth project. It consisted in turning an atmega16 microcontroller in a simple 4-voice polyphonic synthesizer.
I've been working a lot with microcontrollers so i was curious to see if i could turn an atmega into a polyphonic synth. Another reason for this project was a lot of spare time at my dayjob since this project is much about programming which can be done ver easily in the office ;-)

By simple i mean that it only contains one oscillator per voice, no filter and a ADSR envelope per voice. Of course there are some other avr-based synth projects (http://www.jarek-synth.strona.pl/,http://www.avrx.se/), in contrary to these projects i tried to focus on polyphony, with the aim of including it in a modular synthesizer.
What came out until now is a very cheap, cheap-sounding and cheap-sounding box. Technical data:
wave sampling rate: 31,25 kHz
resolution: 16bit
envelope (and possibly lfo) sampling rate: 244 Hz
midi: note on/note off
Extensions: basic wavetable capability, lfo


here is the video of this synth(sorry about the sound)

PCB layouting

lots of thing have been going recently. I've started a new project: a digital polyphonic synth based on an atmega16 ic (pics and vids soon). I was looking for a good schematic and pcb designer software. After a long search and much hairpulling i finally decided to use an open-source package called gEDA. Although it has some small bugs it's much easier and less complex than many commercial options.
I've now (nearly) finished the layout of the sequencer master part. Here it is. I've noticed that (clever) layouting pcb's is a quite demanding task..
Please don't etch boards with that layout yet (that's why i didn't put the resistor values..) i have'nt checked and tested the layout yet

Synthesizer Treffen

One of the most important personal goals regarding my home-built instruments is to make music with them, so i did recently at the swiss synthesizer meeting in zurich. It was whole an deeply satisfying new experience making electronic music in a large group (8 musicians). The largest part of the session was ambient stuff with very minimal beats. Coming originally from drone/stoner rock i really liked this slow and spheric, but not so hard-n-heavy part of the session, simply magic!!
Here a video of the session, i was there with the soundlab, electric guitar and, of course the drumsynth, in the video i'm playing guitar


picture and audio files can be found here , if you want to hear the sound of my drumsynth check out track 'Jam teil 8'

Schematics (finally!)

Here they are, i put them on a real website as pdf's, currently without explanation. I hope they understandable and helpful. Since many people were interested in building the entire or part of the project i'll definitely work on pcb's and make them available on the net as well. Once working flawlessly i'll publish the source code for the sequencer part too. At the present state it still crashes from time to time but i'm working on it..

Beta Version finished

Hi there

The beta Version of my drum synth is finished!! Since the case is fully packed with electronics i will add midi-support and some of the optional stuff externally. It's extremely fun to play around, and, as i wrote in the first post in this blog, very intuitive and easy to use since i created this instrument exactly as i wanted is (withing the limits of my mechanical building capabilities of course).
Thats certainly not the end of this blog and my era of building synth's, i feel more it's just the beginning .. there are too many ideas spinning around in my mind. Anyway, currently i'm working on clean documentation of the drum synth and on PCB's. On idea of pusblishing the building process on the web was also to make it accessible for other people. For this i got some webspace where i will put the sequencer source code, schematics etc. Having it as pdf's is probably more usable than plain pictures... Please let me know if you're interested in re-building the project of parts of the project, if someone is interested i will start designing pcb's. I'm also willing to sell pcb's and programmed avr's ..

Here is the video of a first demo, unfortunately i turned up the reverb a little bit too much..

Hi-Hat's Module Demo Video

This video shows the sound of the hi-hats module. This is certainly the most 'musical' one of all videos since i'm trying to play rhythmic patterns with the 'open' and 'closed' hi-hat sound, habe fun!

Another Demo Video: The Tom Module

Here's the second video of a serie of three video's i recorded recently, my spoken english is improving slowly but steadily i guess.. Anyway here's the demo of the tom emulating synth.. enjoy!

Crash Ride Module Demo Video

Finally i managed to get time to turn some demonstration video. This one's about the not so peacefully sounding crash ride emulation. I worked on the sound of the demo videos a bit, mainly because i got my mixing table back from rehearsal room i once used. The sound of the modules heard is comes directly from the line input of the audio interface. My explanation are recorded through the camera, thats why the synth sound are crystal-clear and the narrating isn't. Have Fun!

Tom Module Concept

Oops, i forgot to put the concept of the tom module on this site.. by the time it's already finished, anyway... The sound source is a 'true' sine wave oscillator.
This means that the sinewave is generates using a 18dB/oct filter made to resonate permanently. On other synth vco's sine wave are sometimes generated by 'bending' a triangle-wave into a sine wave. The difference between a true sinewave oscillator and a 'bent' sine-oscillator is only noticeable in fast frequency modulation. Due to it's nature the sinewave gets asymmetric generating overtones. I wanted a very clean sounding tom (as a counterweight to the more rough sounding bassdrum) so i realized a true sinewave oscillator. Frequency and volume can be/is modulated with two decay-only envelopes..
p.s. for sake of simplicity and layout esthetics (see post below) i omitted the distortion unit, that anyway contradicts the paradigm of a clean sounding tom

Some layout changes

The drumsynth project is progressing well, currently basic versions of all the sound modules are built (except for the 'sampler', this is gonna be either a circuit-bent mp3-player controlled by an atmega8, or an atmega16 which plays 'computed' samples). Also the mixer part, the device which sums up all the individual sound sources is finished and partly tested. For testing i attached only the bass drum to the sequencer and played around.. great fun.. can wait for the thing to be finished!!

For the sake of order i removed the 'distortion' knob on the tom module so that all the modules expect the snare have the same height. To make the layout a little 'sexier' in added some decoration (paying tribute to my patron and mascot LB ;-) ).
Here is how the new layout looks like:

Crash Module Design Changed

Currently the last sound module is being finished. While testing the sound i made some changes in the design. In the new design the phaser 'position' can be changed with the amplitude envelope as well, the amount of envelope in the phaser 'position' can be controlled with a knob. On the other hand the high-pass filter has been omitted. With this configuration some very cool sound ranging from starting jets up to something similar to gunshots can be realized, not really what you expect from a crash cymbal, but definitely unique!

Trial and Error vs. Abstract Math

Before putting any new info about the other sound modules i'd like to talk a little bit about the electronic design 'strategies' i used to create the schematics. A lot of this was done base on a trial and error approach without really calculating exact values for the resistors and capacitors. Usually this worked quite well, since a lot of design ideas are taken from other schematics that already have shown to work. For some things however it's good to force yourself to go through the entire math to see whats happening in the circuit.

For the last module (the crash) i'm gonna use a phase shifter to shape the sound. For this had no idea which values to use so i was forced to go through the calculation. Also almost no (understandable) info was on the web so i decided to uncover the working mechanism of a phase-shifter or allpass myself. I put the results together in a short article (if the link doesn't work copy http://moads.100webspace.net/the_phaser.pdf in the adress bar of your browser. I hope that the document is helpful for others everything is there is correct.

Sequencer Finished

The sequencer part is finished! I finally found all the major bugs in the firmware for the two atmega8 microcontrollers. I'll publish the source code under gnu public licence on this site as soon as it's cleaned up a little. I've made video showing the functionality of the sequencer, enjoy!.

Hi-Hats Concept

The hi-hats sounds are based on 'true' white noise coming from a transistor based noise source - a little tip about transistor based noise source: use PNP transistors as a noise source (i use standard 2N3906 types) instead of the normally used NPN transistors, i compared the noise level of some 2N3904 (NPN) and 2N3906's and found out that the noise level fluctuates a lot for the 2N3904 but for the 2N3906 the average noise level is much higher and it fluctuates less - the noise is then fed through a decay-only controlled VCA. The envelope has two decay times, one for the 'closed' case and one for the 'open' one. For each triggering the envelope is restarted. The noise is then filtered with a 12dB high-pass filter. Thats it!

Snare Drum Concept

Lately i was working heavily on the sequencer software, fortunately a basic version of it now works! for me debugging is a painfully boring yet demading task which is don't like at all. Especially because digital circuits are not likely (at least this one was not) to show any funnly behaviour when working incorrectly. Debugging analog sound circuits can be very inspiring and funny since in general malfunctioning circuits tend to make some sound which is of course is not what you expect, sometimes even better than that.
Lets come to the snare dum concept. The sound source is six oddly detuned square wave oscillators which should make a 'grainy' noise sound. This sound is then fed through a normal decay only evelope controlled VCA. An additional 6dB high pass filter is added to be able to shape the sound a little bit.

Writing Software and Burning Atmega8

Currently i've finished the controller hardware part and the software. At the moment i'm trying to test the software. Problem is that i've managed to burn four Atmega8 microcontrollers while trying to program them. For those of you not familiar with Atmel's microcontroller see Wikipedia or AVR freaks. Unfortunately things get expensive quickly once you burned some of them. So i spent hours testing the boards where the atmega's were mounted but couldn't find any errors. Since i've done all my programming with a cheap home-built programmer in the end i suspected the programmer to damage the atmega's. So i forced myself to spend 80 euros for a 'real' programmer (the avr dragon) with this i hope to prevent damaging more micros and even bringing to life the 'dead' ones. The software will be published (und GPL) once working.

Some photos of the case

The case in the beginning (unmodified). The plastic separators have to be taken out.
Some plastic separators melted out... was a long and tedious work
LED and rotary switches mounted, staring the wiring of the user interface components
The switches are wired the leds not (yet)
front view (note that the paper layout gets glued over the switches again once finished), probably the layout for the 'upper' part isnt finished yet.

Bass Drum Module Demo Video

Hello everyone

My first video demonstrating the bass drum module of my drumsynth is finished, have fun ;-)

Bass Drum Module - Concept and Layout

Lately i started working on the sound-modules themselves rather than the 'computer'-part controlling the entire thing. I started with the bass drum module because it is one of the most inportant modules. When working with software drum synth or synthesizers in general i always found it difficult to create a bass-drum sound which is punchy and pumping at the same time. Another important aspect for this project is that the electronic should be as simple as possible and the electronics should be buildable with cheap and easy to find parts.

The core of bass drum modules is a 40106 hex inverted-based oscillator. Normally this oscillator is used to create rectangle-wave signals. However one can also get something like a triangle-wave signal when using the voltage at the input rather than the output of the hex inverter. The frequency of this oscillator is controlled with a decay envelope and a pot.
The output the goes into a voltage controller amplifier. The volumes is controlled with a simple envelope which consists of a decay envelope signal and a square pulse signal summed up. The idea is to have a punchy attack, done with the decay-envelope and some kind of a compression effect, done with the square pulse.
The schematic of this layout is shown below. At the moment the module is built and being tested, sound clips and/or vids following soon, stay tuned!!

The layout

Here is how the frontpanel will look like. I've never been a big graphic designer so the design is kept quite simple. Each blue field contains the knobs for one sound module. The sound modules are intentionally not labeled since the modules are (or should be) designed flexible enough to to create sounds which are beyond the scope each separate module was designed for. Still the intended use is (from left to right): bass drum, snare drum, high-hats (open and closed), tom and crash. The bottom part is the step sequencer with red leds showing the pattern of a single instrument and green leds which show the position in the sequence. Two rotary knobs select the intrument which should be edited. Apart from the six internal modules an external channel (there's an accent and trigger out for this channel) and a digital sample player (a future project) can be programmed. Four patterns can be stored, the pattern to be edited is selected using the second rotary switch. The rest is self-explanatory i think.
Oops, i forgot to say: the case will be an plastic toolbox, so the size is adapted to this box.

Digital Part: Concept and schematics

For the digital part i chose to use atmega8 microcontroller to do all the tasks related to the sequencer control. These are mainly handling the buttons (start, stop, running mode, load/save pattern etc.), controlling the led's and generating the trigger pulses and accent switches to control the analog sound modules. When starting i thought that one atmega8 will be sufficient to do all the task but i soon realized that i needed a second one.
I decided to split the two controllers into a 'master' controller and a slave controller. The idea behind this is that the master processes all the incoming information (buttons and sequencer clock) and the slave controller is used to control the output (the leds and interface to soundmodules). Communication between master and slave is done via i2c.
Here is the schematic (not tested yet) for the master partHere is the one for the slave part

How it all started...

The first time i came in touch with music electronics or electronics in general was roghly six years ago. I was playing the electric guitar for a while and wanted to buy some cool distortion pedals, so that i'm able to play heavy rock and metal. I realized that i couldnt really afford them. While surfing on he internet i found out that it's not that difficult to make them myself so i gave myself a try on building some effect. I started with building some designs found on net, later i tried to design some pedals myself. But since i didn't know how the things inside work it wasn't that successful. Also because of financial issues i never used good quality switches and boxes. It ended up by putting the ~7 effects is had in a huge box. So in the end i had a huge, humming box with 7 strange sounding distortion effects in it. Because of several reasons i didn't toucha soldering iron for a while.

Renewed Interest
About a year ago i was looking for a midi controller keyboard to used with pc-based software instruments. By chance i found one at a local flea market in berne, switzerland. This flea market is organized by a alternative culture center 'Reitschule' (b.t.w. its a great place to visit), so a lot dubious sellers are around there. Maybe thats the case for all flea market but anyway... So i found this controller keyboard which looked totally new. After some negociations i bought is for 80 Swiss francs which is ~50 eur0s. Back at home i quickly found out the keyboard wasnt working at all. After nearly throwing it from the balcony i rememembered that i once wanted to start building a synthesizer. After some days of investigation i came to the decision that i'm gonna try to put the keyboard back to life by constructing an analog synthesizer around it.
Since i haven't really worked on electronics for a long time i started with very simple projects. By that time i discovered ray wilson's excellent music from outer space site. Without the invaluable information on this site i probably wouldn't have been able get into analog synthesizer technology. Thanks to ray wilson for building and supporting this excellent site!.
The first 'real' project built is the weird sound generator from MFOS. Due to a lack of an appropriate enclosure i mounted it into a a folder.
The folder gained some popularity amolng my bandmates and friend. So i offered them to build the same circuit in an specific enclosure which they could choose on their own. By the time i already biult three copies of WSG, one is mounted in a plastic case (see the myspace site dedicated to it) another one in a wooden box and the last one in a former model airplane remote control. By the time i planned my next project: the MFOS mini synth. After building three WSG's i felt sure enough to modify the mini synth slightly to fit my needs. In order to use it with the keyboard i added the scanning matrix encoder (also from the mfos site, your guessed it, right?). I also added an external (guitar) input which routes the external signal to the filter and vca section, beside that it has an envelope follower and a signal to gate converter. Another major change compared to the original sound lab is that it can be patched beside the switches that do the signal/cv routing. Below are some photos of the synth and the keyboard.
There are already enough not-so-musical soundlab demos on the net so i decided not to make a 'raw' soundlab demos yet, once i have time i'll add a video. You can hear my soundlab in a musical context (well that always a matter of taste i guess) on my myspace site. The sound lab can be heard in the songs 'the funky chicken', 'a word' and 'drifting'. In 'the funky chicked' the bass is done with the soundlab, the horns in the band (not in the intro, these are samples), and in the keyboard solo in the middle. The playing of this is lousy but the sound is quite decent i think. In 'a word' the bass and all the synths are done with the soundlab (except drums). The wah-wah guitar well heard during the first verse is the acoustic guitar processed by the soundlab. In 'drifting all the synths except drums are made with the soundlab.
After the soundlab was finished is built three guitar effect boxes, a little guitar practice amp and a midi to control voltage converter. More on these, especially the midi-to-cv converter later.

The Beginning

Today I'm starting the blog which is dedicated to the development of an analog drum synthesizer. I was asking myself for a long time if I really should start this since there are already too many blogs out-there (in my opinion) with doubtful content sometimes. Anyway, this is about building a analog drum computer. A drum computer is a device on which rhythms can be programmed, it is one of the basic ingredients for all sorts of electronic music. There are some famous examples of drum computers such as the tr-808 from roland which is unfortunately not produced anymore. My own design will be heavily influenced by the tr-808 although i never had an example in my hands. I'm mainly doing this because the tr-808 is considered to be a classic and it looks cool too. And i've player around with the virtual version of it (Propellerheads rebirth). So whats my motivation for building a drum computer? If you just want to create beats then you probably use a pc and a software drum machine. There are some good ones out there for free (for instance propellereads rebirth. Thats not a i want to do with is. I'm rooted in 'live played' music and i'm a guitar and piano player for a long time so i don't really get along well with creating music by clicking on virtual buttons on a computer screen. I'm happier with a hardware instrument which is quickly started and easy to use (one button - one function, no menus). Well in this case a hardware groovebox could do the job so why still build a drum computer myself?
First: building an instrument is fun, and, once finished, a deeply satisfying experience.
Second: There are almost no drum computer diy projects on the web. With this project i hope some give something back to the synth diy community of which i benefitted a lot during my first synth diy projects. I will post some information on these later on.
Third: A self-built instrument is very easy to use is the since you built it the way you want to have it.

Basic Concept

Before starting the first sketch of a schematic i put up a list of general features the drum synth should have. The basic concept is a step sequencer based drum computer with analog sound generator circuitry. The pattern control and programming is done with atmega8 microcontrollers. These are chosen because they are relatively cheap and they can be programmed with very little extra hardware required. The following sound modules will be built:
->Bass Drum
->Snare Drum
->Open and Closed Hi-Hat
->Tom
->Crash
The sequencer should have 16 steps, it should support 4 running modes:
->forward (the standard one)
->backward
-> forward followed by backward
-> random
Beside a clock is supports also a manual advance mode. A 'swing'-mode which enlarges every second note period is not planned currently. The sequencer will be able to control eight instruments whereof six are integrated in the drum synth. The sequencer will have a trigger output to control external equipment. One instrument control line is reserved for a digital sample player which might be added later on. Each trigger pulse will have a normal and an accentuated mode, the amount of accent will be tunable for each sound module separately.

There will be the possibility to store 4 patterns in the internal eeprom of the atmega8's. It won't feature a 'song-mode', i.e. the ability to store a sequence of patterns since it is designed as simple instrument which will be played as 'life' as possible. But we'll see, maybe i add that later on. However it supports playing one pattern while programming another one, as many things else this is shamelessly copied from the tr-808. Thats it so far. Next i will some sound clips of the prototypes of the snare and bass drum modules since i already built as well as some information on my finished synth projects, schematics etc. Beside electronics i play music solo and in a band.