percussion robot

version 2.0


This instrument is a computer controlled assembly of five rototoms. Each drum is equiped with a set of beaters and the pitch of each rototom can be controlled. The lowest couple of rototoms have 3 beaters each, whereas the upper three suffice with two beaters each. The beaters are velocity sensitive and have a very wide range dynamic control. Heavy duty stepping motors are used to achieve pitch control of each individual drum. Instead of rotating the drums on their fixed axis, we fixed the frames of the drums and rotate the threaded axis through a geared construction using dented belts. This also contributes to more silent operation of the mechanics.

The instrument can listen to midi commands directly, using midi channel 8. However, good control of the pitches requires a lot of midi controllers to be send to the robot. The picture above gives an idea of the construction of the beater side of <Rotomoton>. The construction of two of the five stepping motors can be seen on the picture below. The picture was taken with the position sensors rermoved.

<Rotomoton> uses a lot of dedicated hardware, designed for musical automats such as player pianos, percussion instruments, organs, brass instruments, and even bowed instruments. Details can be found in our course on experimental music on this same website as well as on the detailed pages treating our other musical robots composing the Logos M&M robot orchestra.

The original version dating from 2000 used hardware based on a parallel bus controlling Intel 16 bit timer chips with a resolution of a single microsecond, next to motor controllers of our own design. It was not directly controlable using midi but required a separate laptop computer to translate midi (or midi via UDP/IP) commands to the parallel bus commands required. In 2005/2007 we redesigned the hardware such that we could get rid of the laptop. No less than 7 PIC microcontrollers now take care of the control of every detail of the mechanics. The timing resolution suffered a bit under this change and is now limited to 19.2 microseconds. An 8th PIC microcontroller takes care of the support for Midi via UDP/IP now.

The stepping motors use special circuitry, and make use of special motor controller hybrid power modules. Since in this application, very high force but no holding torque is required from the steppers, we could use many power saving features available from these motor controllers. The hybrid stepper motors used in this design are Cetronic or MAE type HY200 3424 170 A8, specified for Umax = 90V and maximum phase current 1.7A. We use them with the motor windings in parallel.


The solid aluminium dented wheels are selected as follows:

  motor shaft number of teeth drum axis wheel number of teeth belt type maximum stepfrequency nr. of motor steps required for one revolution of the drum axis
smallest drum (Sop) 12 42 220XL 500Hz 700
second drum (Mez) 12 48 180XL 450Hz 800
third drum (Alt) 12 48 200XL 450Hz 800
fourth drum (Tenor) 12 60 220XL 350Hz 1000
fifth drum (Bass) 12 72 210XL 350Hz 1200




Two bits on each microcontroller for the steppers are used to read the start and end position of the tending mechanism, two bits for each drum. Miniature microswitches were first used as sensors. These may be replaced by optical or inductive proximity sensors with 0.01mm resolution later on, since our microswitches suffer from a larger than necessary hysteresis. In 2007 we experimented with inductive proximity sensors from Pepperl+Fuchs. Herewith we could obtain a hysteris of ca. 30µm. The sensors are now connected to two analog inputs on the PIC. For precize positioning, it is mandatory to reset all motors first to the lowest and next to the highest end position prior to running music compositions. This calibrates the pitch range for each drum. Software to handle this automatically has been integrated into our <GMT> language (also part of our GMT midi-file player for the M&M robot orchestra). Technical exploded drawing showing the construction of the tuning mechanism driven by the stepping motors:

5. The power supply for this instrument is rated for 1130Watts. Three transformers are used: 230V/12V - 630VA and 230V/2x35V 200VA. and 230V/24V - 300VA. During normal operation however, power consumption is reduced to ca. 150VA.

As other instruments belonging to this <Slag-Werk> project (a range of robotic percussion instruments), this one also is designed for mobile use and thus mounted on sturdy steerable wheels, the back of which can be seen on the picture.

Midi implementation and music:

If you are using <GMT> under Power Basic, you can use all specific hardware control functions and procedures provided in our libraries. However, direct midi control is also possible from any decent sequencer. (We use Cakewalk or Sonar, written by Twelve Tone Systems). The midi and hardware mapping of the different components of <Rotomoton> is:

  • Solenoid beaters ( from low to high rototom):
    • Drum 1: notes 48, 49, 50 (beaters from center to rim), velocity implemented.
    • Drum 2: notes 51, 52, 53, with velocity
    • Drum 3: notes 54, 55, with velocity
    • Drum 4: notes 56, 57, with velocity
    • Drum 5: notes 58, 59, with velocity
    • Lights:Bright white carlights: notes 114, 115 (the velo-byte controls the light strength)
    • Light: Blue LED sportlight, frontal forward orientation: Note 116 (on/off only)
    • Light: Blue LED spotlight, frontal oriented towards drums: note117. (on/off only)
    • Lights: Notes 118-119: reserved for future expansion. (on/off only)
    • in music staff notation:
  • for drum rolls: [only implemented under GMT and thru our midi file player]
    Note on followed by note pressure. The repeat frequency (in Hz) will be: (0.1 + 30 * aftertouchvalue / 127 ) for the highest three drums; (0.1 + 25 * value / 127) for the lowest two. The velocity can be changed during the drum roll by sending a note on with a different velocity. The drum roll continues until a note off for that note is received.
    • Drum 1: note 48 + pressure
    • Drum 2: note 51+ pressure
    • Drum 3: note 54 + pressure
    • Drum 4: note 56 + pressure
    • Drum 5: note 58 + pressure
  • Stepping motors:
    You can tune the drums individually by sending controllers 101 to 105 (from low to high). Note that the absolute pitch is very difficult to control, since it depends highly on the initial and manual tuning of the drums.
    • Drum 1: controller 101, values 1 - 127 mapped over a fourth (range set by calibration commands)
    • Drum 2: controller 102, values 1 - 127 mapped over a fourth
    • Drum 3: controller 103, values 1 - 127 mapped over a minor sixth
    • Drum 4: controller 104, values 1 - 127 mapped over a major sixth
    • Drum 5: controller 105, values 1- 127 mapped over an octave + minor third.
  • Callibration commands (these command cause the range to be recalculated internally):
  • Reset Drum 1 to lowest position: controller 81: value > 0 - value = %False = cancel command
  • Reset Drum 2 to lowest position: controller 82: value > 0
  • Reset Drum 3 to lowest position: controller 83: id.
  • Reset Drum 4 to lowest position: controller 84: id.
  • Reset Drum 5 to lowest position: controller 85: id.
  • Set Drum 1 to Highest position: controller 91: id.
  • Set Drum 2 to highest position: controller 92: id
  • Set Drum 3 to highest position: controller 93: id
  • Set Drum 4 to highest position: controller 94: id
  • Set Drum 5 to highest position: controller 95: id
  • Calibrate range drum 1: controller 111: id. (after calibration the drum will be at the mid-position)
  • Calibrate range drum 2: controller 112: id.
  • Calibrate range drum 3: controller 113: id
  • Calibrate range drum 4: controller 114: id
  • Calibrate range drum 5: controller 115: id.
  • Beater power on/off: Controller 66 (on/off)
  • Motor power on/off: Controller 65 (on/off). This command also initiates a calibration command on all five drums, at the end of which all drums will be tune to the midi midrange position (64). So you do not have to send any of the above calibration commands.
  • Lights:
  • notes 114, 115 (large white carlites), velo controls the dimming,
  • notes 116, 117 (blue LED spotlites in the front) (on/off)

Composers that wrote pieces for <Rotomoton> sofar, include: Godfried-Willem Raes ('Rotstuk' voor Rotomoton), Michael Manion, Moniek Darge, Kristof Lauwers, Rene Mogenson ('The swarm breaks through the net'), Claude Coppens. Rotomoton is a fixed member of the M&M ensemble and as such also performs in orchestra pieces such as 'Technofaustus', 'Gestrobo', 'Picrada', 'STOvSQE4MM' , "Quadrada", "g_tech611" .. by Godfried-Willem Raes and many other pieces.


  • Filip Switters (TIG welding)
  • Xavier Verhelst
  • Moniek Darge (painting)
  • Kristof Lauwers (GMT player and software simulator implementation)
  • Johannes Taelman (PIC revision 2007)


  • width: 1500mm
  • height: 2100mm
  • depth: 600mm
  • weigth: 220kg
  • power consumption: 1130Watts / 230V AC
  • data input: midi input. (direct network UDP/IP will be added soon)

Insurance value: 19.000 €.


Construction of this automated instrument started august 2000 and was completed in its first working version around easter 2001. Rotomoton played its first scales on april 26th of 2001. It is now a permanent robot member of the Logos <M&M> ensemble. Its hardware was completely revised and rebuild in 2007.




The picture on the left shows the first step in the construction of <Rotomoton>: the assembly of the five rototoms in a TIG-welded frame.

The final instrument is shown on this picture:



The <Rotomoton> automat can be heard on the Logos Public Domain CD <Automaton> (LPD007).

Revisions 2005-2007:

aim:1. improving pitch change speed by increasing torque from stepping motors.

In the original design we used unipolar steering of the stepping motors. By using bipolar drives, the torque can be increased with a factor 4 compared to the first version, where considerable power got lost in the series resistors. We decided to use Intelligent Motion Systems (IMS) stepping motor drivers, type IB106, each capable of delivering currents up to 6A and able to cope with voltages from 24 to 80 Volts. (Cost: ca. 350 Euro/piece) . For the motors we use in this robot, we could easily increase the power supply voltage to 48V and thus achieve high rotational speeds with high torque. So for each stepper, we used a single IB106 module, controlled from a PIC (18F2525) directly. The timing requirements for the motor driver are:

In the first design we had five microswitches to sense the lowest possible position for the stepping motors. No sensors were foreseen for the highest possible position. This was dictated by the fact that on a standard PC printer port, there are only 5 bits free for input. Since we do not have this limitation anymore using a PIC design, we now fitted rotomoton with end position sensing microswitches as well.

aim 2: changing control such as to get rid of yet another Windows laptop, by controlling rotomoton directly via midi commands and optionally midi via UDP/IP.

aim 3: Increasing the dynamic range of the instrument. This involved merely increasing the power supply voltage for the solenoid drivers to ca. 75V. The pulse generation can be confined to a separate PIC controller.

By also using PIC microprocessors for the steppers, we could at the same time make the robot listen to midi commands directly.

The five boards for motor control we now use were build after the following schematic:

Some sections of the circuitry do not apply to Rotomoton, since the board was designed such as to replace similar functions in our <Flex> robot as well. The power supply section had to be completely revised also. We kept the hefty 12V/630VA transformer since it was required for the brigth lights on <Rotomoton>. In the new version dimming becomes possible, since they use PWM on their supply voltage. Note the absense of a smoothing capacitor on its power supply. As a whole, the power supply is clearly overdimensioned now, since the bulbs will draw at the most 200W of power. The solenoids for the beaters get their power from a separate 70V ungerulated power supply using a toroidal transformer. For the stepping motors, an extra 24V /12A transformer is connected in series (after rectification) with the already present 12V supply. The unstabilised voltage available for the motors thus becomes 50V. With all motor activated this voltage may collapse to 44V.

Board layout for the five motor control PIC-micro boards:

Circuit drawing for the beater pulse-generating boards.

On the left picture below, we see the new five motor control boards and next to it the new midi-hub board.

Last updated: 2017-08-05 by Godfried-Willem Raes

Maintenance, research and repair logbook:

  • 01.09.2000: Start construction rotomoton.
    20.09.2006: Rewiring of power supply block
    21.09.2006: Tests for the new power supply sections. Beater PIC programmed. Given the higher voltage we use now to operate the beaters, we have to remove all surge arresters over the solenoids.
  • 23.09.2006: Velo scaling tests and PIC1 debugging session.
  • 25.09.2006: Return springs on beaters replaced. PIC beaters, code version 3.0
  • 27.09.2006: Start construction frontal lights assembly.
  • 28.09.2006: Front lights construction and wiring finished. Mapping on notes 116, 117. Blue LED spotlights running on 12V.
  • 09.10.2006: Discussion meeting with Johannes Taelman with regard to code implementation for the steppers and coding for the midi-hub board.
  • 10.10.2006: Midi hub board programmed: relais ctrl.66, 67. PWM for the strong lights implemented. Blue frontal lights working,
  • 11.10.2006: New test code written in GMT.
  • 05.12.2006: New motor control PIC boards designed.
  • 11.12.2006: New PIC-micro board design send to Europrint for production.
  • 04.01.2007: New boards arrived. First test board assembled and soldered.
  • 05.01.2007: Construction stainless steel chassis parts for the motor control boards.
  • 06.01.2007: further assembly and soldering of PIC-micro boards.
  • 08.01.2007: wiring of microswitch sensors to the controller boards.
  • 10.01.2007: finalisation of wiring: power lines and current sensing resistors. (330 Ohm). Without these the maximum current is 6A. With 330 Ohms, we bring it down to 3.45A
  • 12.01.2007: PIC-specs upgraded. Jumpers for H/F placed parallel to 6p weidmuller, to Vcc to set full step mode.
  • 13.01.2007: construction detail pictures added on this webpage.
  • 22.02.2007: Johannes Taelman contacted again for PIC coding. Pic-specs upgraded.
  • 26.02.2007: Still no news from Johannes...
  • 13.03.2007: Johannes shows up. Discussion on PIC implementation of motors.
  • 18.05.2007: Programming works for the motor control PIC's.
  • 19.05.2007: wiring bug in motor Cetronic HY200 3424 170 A8 hybrid stepping motor repaired.
  • 28.06.2007: PIC programming session with Johannes Taelman. Sensor reading code and end position commands implemented.
  • 30.08.2007: PIC programming works continued. Midi-out added on boards for debugging. Reading end switches seems to work well now.
  • 31.08.2007: PICS 1 -5 for motor control reprogrammed. Code is autocalibrating now. Command request for low pos and highpos on init is essential now.
  • 01.09.2007: GMT test code rewritten for the new functionality. Pitch controller mapping must be in error somewhere...
  • 02.09.2007: Loose wiring on highest drum motor repaired. Further debug of test and evaluation code. Search into the origins of the collapse and heavy load on the logic power supply. (5V). Problem with the too high hysteresis of the microswitches.
  • 03.09.2007: Due to its large size and heavy weight, Rotomoton cannot be transported with our regular Dockx car rental trucks. A large truck is required. Rotomoton is higher than <Krum>. All springs on the beater solenoids replaced with pull-type springs with eyelets 0.7 x 6.7 x 30. (Fabory assortment box). Some beaters replaced with bakelite M5 bullet-balls.
  • 04.09.2007: 12V car relais for the beater voltage burned out. Replaced with new 30A type. Current drawn: ca. 150mA. Voltage regulator on midi hub board (7805 type) equiped with cooling profile for small TO220.
  • 06.09.2007: New programming session for the PIC firmware. Note order bug... Sysex implemented.
  • 08.09.2007: Programming sessions with Johannes Taelman PIC's. Electrolytic exploded on midi hub board... Component replaced with 25V type and board cleaned thouroughly.
  • 09.09.2007: Hardware debug session: the italian made car relais apparently cannot survive long periods in the activated state. Not only the coils get very hot and eventuially burn out, but worse even, at doing so they cause a short between the coil driving voltage and the relais contacts themselves. Such designs ought to be forbidden since they contradict the reason of existence of the relais-device itself! So we did throw the relais out altogether. They caused a new explosion of the 1mF/25V electrolytic on the hub board. The dissipation of the 7805 regulator seemed on the high side. It got very hot, so we decided to mount a 7810 preregulator on the midi-hub board, spreading the dissipation equaly over two devices. Relais replaced with small Zettler types, 12V coil, 16A contacts. Electrolytic capacitor on power input of hub board replaced again with a new 1mF/25V specimen.
  • 10.09.2007: New Rotomoton demonstrated for Kristof Lauwers and Sebastian Bradt. PIC finalisation session with Johannes Taelman does not happen: Johannes is sick.
  • 12.09.2007: Pepperl & Fuchs proximity sensors delivered. First evaluations and tests. The possibilties examined are drawn in the following circuit drawing: The principle of operation of these inductive proximity sensors is that they contain an LC oscillator circuit (operating frequency ca. 714kHz) that as soon as a metalic object comes close enough, the oscillation damps and eventually stops alltogether. The amplitude output can be used as an analogue signal to measure the distance in a traject-range of about 1mm corresponding to a voltage range of 1 to 2V, depending on the resistor values taken for R11,R12. The remainders of the oscillation signal can be greatly reduced by placing a 47nF capacitor across the inputs to the PIC. This goes at the detriment of response speed however. The measured hysteresis of the sensor device is ca. 3% of the range (1mm), so ca.30 microns. Although the sensors are specified for a normal operating voltage of 8V DC, they work fine down to ca. 3V.
  • 14.09.2007: Robotic tests with sonar interface (Gestrobo Study for Rotomoton), Picradar (Picradar Study for Rotomoton) and 2.4GHz radar (Quadrada study 'Rotom').
  • 17.09.2007: PIC programming session. Now we autocalibrate only on the edge of the switches. Controller 111-115 also implemented now as well as autocalibrate on controller 65. Rotomoton code in GMT adapted according to new version.
  • 25.09.2007-02.10.2007: Rotomoton tests under GMT control.
  • 17.09.2007: Rotomoton part added in 'De weg,der Weg, the way' with HY1 sensor.
  • 26.10.2007: Some extra 24V solenoid valves ordered from A.Laukhuff to add beaters on the drums.
  • 31.08.2008: Upgrade of the GMT tescode for Rotomoton
  • 07.06.2011: 10A fuses on the beater power supply went off. The connectors on the pulse board need replacement as the contacts became shaky. Solenoid 3 on the lowest drum also shows shaky behaviour... repair session.
  • 08.06.2011: Rotomoton plays in the M&M orchestra in a piece by Rene Morgenson.