When starting the design and construction of this musical robot, we could build further on the experiences gained through the realization of five automated brass instruments build since 1999: <So>, a sousaphone, <Korn>, a cornet, <Heli>, a helicon, <Bono>, a valve trombone and <Horny>, a French horn. This new robot, an automated flugelhorn, was designed and constructed to the order of LOD, for a theater production 'Helden' with Josse De Pauw. Musically speaking the cornet, the trumpet and the flugelhorn are very similar instruments, differing only in the timbre they produce. The flugelhorn is said to sound more mellow than the trumpet, the cornet finding a place in between. The ambitus, for Bb instruments, is the same. The name given to this robot was derived from the Dutch naming of this instrument 'bugel', although the word bugle is also common in English, referring to a valveless instrument however. The term valve bugle is often used in English and refers to exactly the same instrument, the flugelhorn. Organologically speaking it's a saxhorn, its invention going back to the first half of the 19th century. The bore is much more conical as compared to the trumpet. Here is a picture of the instrument:

Thus it made sense to start the design of this robot by examining the <Korn> robot and improving over it where ever possible. Main improvements we envisaged were the sound volume, the attack characteristics and the speed of response. For the <Bug> design, we did not start with a mechanical design for an artificial embouchure with mouth, lips and mouthpiece coupled to and in acoustic interaction with the tubing of the instrument, as we did in <So> and the first version of <Bono>, but rather used a membrane compressor directly coupled to the flugelhorn via a capillary comprised in the original mouthpiece.The motor driver causes resonance in the flugelhorn tubing, but in this case there is no unidirectional windflow through the instrument. When a note is requested from the bugle, the firmware will calculate the optimum valve combination -including non orthodox fingerings- for the requested pitch. Thus a resonant standing wave in the instrument can be produced. Microtonal pitches are implemented such that the instrument is capable of performing quartertone music, as well as a wide range of different tunings and temperaments with great perfection. The relatively low Q-factor of the horn (compared to strings...) as an acoustic resonator renders this very well possible. The signal generated in the motor was shaped after a physical model of the air pressure waveform in the mouth cavity of a player. (Beauchamp, 1975,1980). Since there is no loop coupling from the resonator to the generator, the sound generation mechanism is a hybrid somewhere between synthetic/electronic and natural/acoustic. The advantage being that the reliability of the robot becomes very high, but this is obtained a bit at the detriment of realism, in particular with regard to the onset of the sounds, A dynamic formant filter was added using an electronic circuitry, in this case a pair of germanium diodes in series with a resonant LC circuit Thus the sound color will be to a certain extend a nonlinear function of loudness, conform to acoustic reality.

The valves are used in this instrument to tune the fundamental frequency of the instrument. The valves can be controlled independently from the mouth driver frequency. They are mechanically driven by unipolar pull solenoids (Emessem types) and use the return spring provided in the pistons. The notes that are normally produced using different valve combination, following theory are: Note that this is untransposed, so in reality everything will sound a full tone lower on a Bb instrument. However, we found out that using the valve combinations entailed by this system, does not lead to optimum resonance in the instrument. Thus we used the optimal valve combinations based on empirical acoustic measurement. A deficiency we encountered in our <Korn> robot was that the buildup of a sound pressure wave in the instrument was anticipating the valve movement. The valves take about 10 ms to take position, thus here we delayed the driver signal with the same amount in all cases where changes of fingering are involved. This introduces some latency but makes the sound quite a bit more realistic, in particular for the attack portion of the envelope.

High brass instruments in their normal human and musical biotopes tend to move quite a bit in space. The highly directional characteristic of these instruments make this also an expressive valuable parameter. Thus we tried to implement movement in two degrees of freedom in this robot: the flugelhorn can be tilted in the vertical plane over an angle of about 90 degrees and in the horizontal plane, it can rotate over nearly 320 degrees. This conforms pretty well to what human players do in terms of movement on stage. The movements cannot be very fast however, at least not much faster than what a real flugelhorn player could do whilst playing. Horizontal movement is faster than the vertical movement, the forces involved in horizontal movement being much smaller than in the case of vertical movement. However, the intention never was to render Doppler effects possible... Note that on a cold boot, the robot will always perform a calibration of its movement. During this calibration phase the red light on the hub board (the board with the MIDI connectors) will be on and the robot will not be listening to any midi commands sent to it. Before switching the power on, make sure the robot is in a more or less central position, both vertically (pointing straight forward) and central horizontally. As long as power is switched off, the positions can be adjusted manually. Never ever do this when the power in switched on, as this will confuse the firmware and cause forcing on the motors excerting braking force against movement.

The electronic circuitry -in overview- consists of four PC-boards:

Overview:

1. Midi-hub board: This board, using a Microchip 18F2525 controller, takes care of the Midi I/O handling and communication as well as the control of all movements: horizontal stepping motor with two end sensors (Pepperl+Fuchs) , and vertical stepper motor with tilt sensor (Penny & Giles). Circuit details, including the PCB design, can be found at the very end of this webpage.

2. Vertical stepping motor driver board using a Nanotec SMC42 compact microstep constant current driver. The noisy fan was removed as well as the DIN-rail mount. The motor is a Nanotec ST5709L-1108-B. The operating current is set to 2 A.

This motor is designed for 200 steps for a complete rotation in full step mode. In this robot, the motor is driven in microstepping mode at 8 clocks for a single step. For the vertical position control we used a Penny+Giles analog tilt sensor connected to the an0 analog input of the PIC microprocessor.

3. Horizontal movement motor: MAE HY200 type . The motor controller here is a Geckodrive G203V.

3. Pulse & Hold board: This board steers the three solenoids for the pistons. The light bulbs and LED's mapped on the midi notes 120 to 127 are also controlled by this microprocessor board. The board is mounted on the upper part, below the membrane compressor. The circuit for the pulse/hold circuit applied for the valves is:

4. Sound generator board: This board steers the 75 Watt motor compressor horn driver via an audio amplifier module. A 16-bit processor is used to generate the required waveform and two analog multipliers are used for envelope shaping and amplitude modulation. Thus we could maintain audio resolution even at the lowest soundlevels. Handling this in the full digital domain would have required a 32-bit processor. The analog dynamic formant filter also found a place on this board.

All these components are on the upperside of the bottom plate.

Power supply voltages and currents:

• +24 V dc (6.4 A) for the valve solenoids, motors and lights (SMPS)
• -48 V dc (2.3 A) for the valve solenoids, pulse (SMPS)
• +5 V/ 2 A for processor boards (SMPS)
• +27V-0- -27V for the power amplifier (linear)
• +15V - 0 - -15V for the operational amplifiers and the analog circuitry (SMPS)

Midi Mapping and implementation:

Midi channel: 12 (fixed in the firmware)
Midi note range: 52 to 94. (Optimum sound in the range 66-89) Note on, velocity is implemented and steers the level of the sustain phase in the adsr.

Note Off commands are required, but can be dropped for pure legato playing. Note off with release is implemented and can steer the release phase of the envelope. Note Off commands for notes that were not sounding are disregarded.

Controllers:

Controller 1: Noisiness of the sound [default = 20]

Controller 3: FM modulation depth (vibrato depth). Large values can cause audible artifacts, due to the modulation of the sampling frequency.

Controller 4: FM modulation speed. (vibrato speed)

Controller 5: AM modulation depth (tremolo depth)

Controller 6: AM modulation speed (tremolo speed)

Controller 7: used as a general volume controller. Note that timbre will change as the volume is changed. On high settings, the formant frequency becomes more dominant. [default = 100]

Controller 10: (Panning) Horizontal movement controller. Value 64: center, 127->0= move left (CCW), 0->127=move right(CW). The firmware will recalibrate on each of the extreme positions (0 or 127). The full semicircle takes about 1 second in time., depending of course on the setting for controller #31 (motor speed). The firmware assumes that after a cold start, the flugelhorn is in a more or less central position. It will recalibrate whenever an end position is encountered.

Controller 11: Tilt or inclination controller. Value 64 = Horizontal, values < 64 = downwards, >64 upwards. After reset and on cold boot, it will default to 64.

Controller 13: allows changes of valve fingering during sounding notes. Bit 0 corresponds to the 1/2 tone valve, bit 1 to the 1 tone valves, and bit 2 to the 3 semitone valve. Using this controller it is also possible to change the fingering for a sounding note whilst it is sounding, thus rendering some sound coloration possible without changing the actual pitch.

The table below gives all details:

Ctrl 13 Value	-1/2t (valve 2)	-1t (valve 1)	-1 1/2t (valve 3)
0	off	off	off
1	on	off	off
2	off	on	off
3	on	on	off
4	off	off	on
5	on	off	on
6	off	on	on
7	on	on	on

Controller 15: ADSR-period [default 40]
Controller 16: used to control the duration of the attack phase in the ADSR cycle. [default 16]
Controller 17: set the level of the attack of the sound defined as the extra boost above the sustain level set with the velo byte.[default 127]
Controller 18 is used to control the duration of the decay after the attack, the time required to reach the sustain level of the sound. [default 18]
Controller 19 steers the duration of the release decay (from sustain level to zero) starting after reception of a note off command. Release will we canceled or interrupted with a new note on command if such a command comes within this time. [default 64]

The interdependencies of these controllers together with the velo byte is shown in the graph below:

Controller 20: Tuning. To be implemented.

Controller 20 - tuning for the flugelhorn. By default equal temperament and A = 440 Hz for value 64. Acceptable values for this controller are limited to:

• 48-59: sets <Bug> to just intonation mode based on fundamentals 48 to 59. The just intonation scale implemented here is: 1:1, 16:15, 9:8, 6:5, 5:4, 4:3, 10:7, 3:2, 8:5, 5:3, 16:9, 15:8, 2:1
• 61: sets <Bug> to equal temperament but transposes all notes a quartertone down as compared to A = 440 Hz.
  62: sets <Bug> diapason to A = 435 Hz
  63: sets <Bug> diapason to A = 438 Hz
  64: sets <Bug> to the standard diapason A = 440 Hz. This is the default on cold boot and reset.
  65: sets <Bug> diapason to A = 442 Hz
  66: sets <Bug> diapason to A = 445 Hz
  67: transposes all notes on <Bug> a quartertone up relative to A = 440 Hz.
• 72-83: sets <Bug> to just intonation mode based on fundamentals 48 to 59. The just intonation scale implemented here is: 1:1, 21:20, 9:8, 7:6, 5:4, 4:3, 7:5, 3:2, 14:9, 5:3, 7:4, 15:8, 2:1

Controller 25: valve movement force controller. With value 18, the movement is smooth and a bit sluggish, whereas with 127 it may get noisy but very fast. With values below 18, valve movement may become a bit unpredictable since the movement will depend on wear, temperature, return spring force variations and greasing of the pistons. After a cold start, this controller will be in a default 64 position.

Controller 26: This controller steers the time the valves stay in their position after reception of a note-off command. This implements some resonance in the instrument after the excitation from the mouthpiece has stopped. It also avoids unnecessary mechanical noise from the valves, in particular on repeated notes or notes that can be produced with the same valve combination.

Controller 31: Motor speed for the horizontal motor (left-right movements). The speed can be varied between a fixed minimum value and a (safe) and fixed maximum value. Value 0 sets the horizontal motor to move at the slowest speed.[default 80]

Controller 32: Horizontal motor acceleration/deceleration time. On cold boot this controller is always set to 100. Larger values lead to a longer acceleration time on horizontal motor movement starts as well as to longer slowdown times at approaching the destination. Acceleration and deceleration are always symmetric. As a consequence, small movements will be performed slower than longer trajects. The larger the value for this controller, the smoother the movements will be performed.

Controller 33: Vertical motor speed. (Up-down movement). At high speed settings, oscillations can occur.

Controller 66: Power on/off switch (0 = off, any other value is on). Power off also resets all controllers to their default startup values. Also resets the program change to the default startup setting.

Controller 69: Enable or disable automation of the left and right eye lights on the bugle. Default value : > 0, ON. To switch this off, send controller 69 with value = 0.

Controller 80: Dynamic range mapping. The default is 64, resulting in a 40dB dynamic range.

Controllers 100, 101, 102, 103: parameters for waveshape. Program change must have been set to Parametric1 in order for this to work. It is mandatory that Ctrl100 < Ctrl101 < Ctrl102 < Ctrl103. After reception of controller 103, the wave table will be recalculated. Following graph describes the waveform and its parameters:

Controller 104: parameter for setting the symmetry when Program Change has been set to Smoothsquare prior to setting the controller value. Value 64 makes a symmetric square. Valid values are between 1 and 126.

Controllers 105 and 106: parameters for the minimum and maximum points of a triangle wave. Program Change must have been set to sawtooth prior to sending these controller values. The parameters will be clear from following waveform graph: The wave tables will be recalculated after reception of the second parameter.

Controller 107: parameter for the pulsewidth: range 2 - 126. The graph illustrates the parameter:

Controller 108: parameter for setting the symmetry when Program Change has been set to Assin prior to setting the controller value. Value 64 makes a symmetric sinewave. Valid values are between 1 and 126. The graph illustrates the parameter:

Controller 109: parameter for setting the symmetry when program change has been set to DirtAssin prior to setting the values of the two controllers.

Controller 110: parameter to set the level of noisiness in the waveform. The wave table for the DirtAssin program will only be recalculated after reception of this controller.

Controller 123: switches the sounding note off, unpowers the steppers, dims all the lights.

Program change: implemented to change the wavetable in use for the sound excitation. Following waves are implemented, using the lowest 4 bits of the value.:

• 0 Sin(x) SineWave
• 1 Sin(x) / x Tophat function
• 2 Sin2(x) Squared sinewave
• 3 Sin3(3) Sinewave to the power 3
• 4 Beauchamp (acoustic model of pressure wave in the mouthpiece)
• 5 Parametric1, as Beauchamp but generalized with 4 different parameters (controllers 100 - 103)
• 6 Smoothed Squarewave. The symmetry can be set with controller 104.
• 7 Smoothed Triangle. Parameters can be set with controllers 105 and 106.
• 8 Sawtooth wave (power on default)
• 9 Pulse. Pulsewidth can be set with controller 107
• 10 Asymetric sine. The symmetry can be set with controller 108
• 11:Dirty asymmetric sine: Symmetry can be set with controller 109, noisiness in the waveform with controller 110.
• 12-15: not yet implemented

Values > 15 (using the 3 highest bits) are used to change the lookup tables in use for the valve combinations. So far following programs are implemented:

• 16: Fingerings after the textbooks on trumpets, cornets and flugelhorns
• 32: Fingerings with a better acoustic result on <Bug> (Default setting)
• 48 not yet implemented
• 64: not yet implemented
• 80: not yet implemented
• 96: not yet implemented
• 112: not yet implemented

Program change commands take some time as the microcontroller recalculates them on the fly. Therefore they should not be issued whilst the robot is playing notes as audible glitches in the sound may occur.

Lights: The lights are mapped on very high midi-notes as follows:

• note 120: Right eye tungsten bulb. The velo byte sets the flashing speed. 0= off, 127= fully on.
• note 121: Left eye tungsten bulb. The velo byte sets the flashing speed. 0= off, 127= fully on.
• note 122: Tubular tungsten light 1. The velo byte sets the flashing speed. 0= off, 127= fully on.
• note 123: Tubular tungsten light 2. The velo byte sets the flashing speed. 0= off, 127= fully on.
• note 124: Tubular tungsten light 3. The velo byte sets the flashing speed. 0= off, 127= fully on.
• note 125: RED led strip mounted on the underside, front, of the horn assembly. The velo byte sets the flashing speed. 0= off, 127= fully on.
• note 126: implemented but not yet mounted. (on/off only)
• note 127: bright red power LED mounted on the hub and motor control board. This light will be on during calibration after cold boot and will dim when the calibration is performed.

Pitch bend: The <Bug> robot can be used in any tuning system. In the drawing below we give the coding example for a quartertone scale:

Most good sequencer software (such as Cakewalk or Sonar) uses the signed 14 bit format. Note that one unit of the msb corresponds exactly to a 0.78 cent interval. To convert fractional midi to the msb only pitchbend to apply follow following procedure: if the fractional part is <= 0.5 then msb= 63 + (FRAC(note) * 128), if the fractional part is larger than 0.5, we should switch on the note + 1 and lower the pitch with msb= (1-FRAC(note)) * 128. Note off does reset the pitch bend for the playing note! The resolution implemented on <Bug> for pitch bend (and vibrato) is limited to 1/10th of a semitone.

Technical specifications:

• size: w: 700 x d: 600 x h: 1070 (in mm)
• Flightcase size: 610 x 710 x 1130
• weight: 33 kg, with the flightcase: 63 kg.
• transportation: needs its flightcase. Cannot be taken as luggage in airplanes, since the weight with case is far over 35kg.
• Power: 230 V ac / 150 W. At rest, power consumption is ca. 5 W.
• Note: on cold boot is is mandatory that the flugelhorn is positioned in a more or less central position, without touching any proximity sensor. A cold boot will always cause a calibration to be performed.
• Tuning: based on A = 440 Hz (within 1 cent). Programmable.The tuning can be adjusted.
• Ambitus: 52-94 (optimal range: 56 - 87)
• control: MIDI-input, 4 MIDI-Thru (differential).
• Insurance value: 24.000 Euro. (Final production cost)

Design, research and construction: dr.Godfried-Willem Raes (2016-2020)

Collaborators on the construction of this robot:

• Kristof Lauwers (application and demo code)
• Xavier Verhelst (Flugelhorn purchase research)
• Mattias Parent (Flightcase construction)
• Lara Van Wynsberghe (orchestrations)

Music composed for <Bug>:

• Dominique Pauwels wrote the music for <Bug> used in the 'Helden' production with Josse De Pauw, produced by LOD. This production toured and had over 50 performances.
• Yvan Vander Sanden: 'Korn en snar op reis', adapted to serve as a demo for <Bug> solo by Kristof Lauwers. Here is the MIDI-file and the audio file in MP3 format as well.

Back to composers guide to the M&M robot orchestra.

Back to Main Logos page:index.html

To Godfried-Willem Raes personal homepage...

To Instrument catalogue

Construction & Research Diary:

• 01.11.2016: Start discussing the building project with Dominique Pauwels and Josse De Pauw (LOD). Demonstration of the capabilities of the robot orchestra.
• 16.11.2016: Preliminary cost calculation of the building project. Inventarising required parts, starting from the design of the <Korn> robot. Main issues to be improved over <Korn>.: loudness level should be much higher to match real sound pressure from a human player. Movement should be faster and more smooth.
• 17.11.2016: Construction of the vertical pole and the horizontal platform in TIG welded 25 mm tubing. Nylon bearings will be used for the movement. . Both vertical and horizontal movement will be implemented: up-down and left-right.
• 18.11.2016: Start circuit design. Inventarizing required components...A stepping motor and a membrane compressor driver. The flugelhorn -brand new, 600 Euro- came in... the first design problems rise up: the line through the valves crosses the main instrument pipe on the low side valves, so in-line activation of the valves from underneath requires some redesign and cannot be optimal. Friction will be unavoidable.
• 19.11.2016: Start construction of the holding structure for the bugle. The shapes of the bugle are too complicated to make precise drawings before the works possible... The base plate will be 3 mm thick stainless steel, width: 160 mm, the holding posts for the flugelhorn 25 x 10 mm stainless profile.
• 20.11.2016: Construction of the tractures for the solenoids and the valves. Here is a picture of the pull solenoids we will use:
• 21.11.2016: A whole day of mechanical work on the holding structure. Test mount of the tractures after welding the vertical posts: We will use M4 threaded rods made from nylon, to dampen solenoid and mechanical noise. Start construction of the holder for the membrane compressor. This has to be minimum 5 mm thick stainless steel (50 x 5). Should lights be implemented? Two blue eyes would be kind of nice... Mouthpiece to fit the compression driver turned on the lathe. Precision work this is. Six working days far we are with it...
• 22.11.2016: Construction of the holding structure for the membrane compressor. We add provisions for two 'eye' lights here as well. For mounting the special nut on the compression driver, a special tool, a hook wrench, is required: This is a detail of the worked mounthpiece and the fit to the driver: . Tubular lites (tungsten filament) mounted on the bottom plate of the horn assembly. On the underside, we mounted a red LED strip light. All lites operate on the 24V power supply. Balancing point for the assembled structure determined and two nylon bearings for 15 mm tubing mounted on the determined exact spot. Mounting with four M8 x 25 cylinder bolts.
• 23.11.2016: Construction of the tractures for the operation of the valves. We glued Helicoil M4x6 thread repair set springs with cyanoacrylate glue in the hole on the underside of the valves. In these such formed threads we screwed nylon M4 threaded rods. Measurements done on the mechanism: the velo-pulse voltage must be minimum 48 V. The hold voltage can be as low as 12 V. The trajecory required from the solenoids is 15 mm. In principle, we can start the measurements and experiments required for the excitation signal for the compressor driver. With no valves depressed, the pitch ought to be 233 Hz. (Bb, 58). Oddly enough, the instrument resonates on B (midi 59) with no valves depressed... As yet we have no explanation for this phenomenon. Should the velocity of the wind flow through the instrument be added to or subtracted from the velocity of sound??? Waveform experiments performed using our Tektronix Arbitrary function generator (Type AFG2021). Sine wave excitation with noise modulation seems to give the most convincing results.
• 24.11.2016: Component selection for the motors and the drives. Order placed at Farnell. For the vertical movement we can use a Nanotec ST5709L1108-B stepping motor (7.8 V - 1.1 A) and a motor controller from the same brand, type SMC42-2,0-1. Construction of a mechanical mount for the motor on the vertical pole. The motor controller will find a place near to the motor, on the pole. Mounting on the pole is with a motorbike titanum tubular hold. A 12 mm axle mounted with two bronze bearings (12 mm x 18 mm x 25 mm) inside the horizontal piece of the carrying T-structure. The drive wheel (Aluminum dented belt wheel 72XL037) with 72 dents turns on this hardened 12 mm axle. The dented belt wheel on the motor (axle size 6 mm) has 14 dents. Thus the speed reduction is 1: 5.14. Shame over Nanotec: the axle of the motor is 1/4" instead of metric 6 mm... and that for a German factory! Motor drive wheel hole drilled to 6.35 mm on the lathe.
• 25.11.2016: Continued mechanical work on the vertical motor mount. Trouble is that the angle of the welded T-piece is a bit off... The dented belt required for the tilt movement should be a 250XL (Gates, Powergrip). Motor controller fits on the stainless steel plate with two M3 bolts. The heatsink has M3 threaded holes. Horizontal movement motor: HY200 3424 470A8 MAE. Dented wheel for the 9.5 axle of this motor: CTB18XL037. Adapted to fit on the lathe. Start design of a board for the valve control.
• 26.11.2016: Technical drawing and construction sketch for the horizontal movement worked out and calculated: All sizes in mm. The distance from the motor axle to the rotating rod is to be determined, as it will depend on the dented belt to be used. Start sawing of the required pieces under the correct angles from stainless steel 30 x 10 and 100 x 10. Suitable wheels with brakes bought: building height 100 mm.
• 27.11.2016: Further work on sawing the 10 mm pieces. Start welding, using special stainless steel electrodes. As the bars are 10 mm thick, this technique is more appropriate than TIG welding. The structure ought to be really strong and vibration free. Here you see it clamped on the welding table: Here is the finished assembly for the horizontal movement: Base plate sawn and sizing determined. Decided to mount it on three wheels, not four. Tentative placement of components. It is clear that the vertical movement will have to be assymetric. The trajectory for downwards will be much smaller than for upwards movement.
• 28.11.2016: Waiting for the ordered dented wheels to be delivered from MEA. Further work on the horn assembly and the bottom plate. Wheels tentatively mounted with M12 bolts. Mains switch holder constructed. Attempting to move as many mains voltage carrying components to the underside of the bottom plate. Power supply for the power amp recalculated: 50 VA at 2 x 18 V should give more than enough margins and overload protection. Precision drilling of all holes in the bottom plate. For some bolts, countersunk types with the head welded on the plate will be required. Mounting order is becoming critical. Construction of the heatsink for the horizontal motor controller. Verical structure welded to the bottom plate.
• 29.11.2016: The proximity sensors we used thus far in robots such as <Flex>, <Korn> etc..., Namur types NJ-2-V3-N (Pepperl & Fuchs) appear to be no longer in production. As we had only one in stock, but needing two, we had to order an alternative type. So, we selected the NBN4-12-GM40-Z0 (Farnell order code 3725741) also produced by Pepperl & Fuchs. Here is the datasheet. We had two Omron sensors in stock, but we decided not to use them as they require a 12- 24V power supply. Continuing mounting works on the base plate assembly: mounting of the horizontal motor. Construction of a holder for the two proximity sensors. With this design, the rotation circle will be limited to 340 degrees. Midi-hub board designed: this board has 3 analog input channels, connecting to the sensors on the robot and 3 x 2 motor control outputs. One output is not dedicated and one output serves a high power LED mounted on the board. Welding of the mounting 2 mm thick plates for the 24V power supply on the base plate. Welding of the T-shaped switch holder to the base plate as well as the L-mounts for the power amp on the upper side of the base plate. Wheels mounted. Start mounting of components: 24V power supply, 100 W amplifier, toroidal, electrolytic capacitors, rectifier bridge, 48V power supply.
• 30.11.2016: Start wiring of the mains power lines. <Bug> implemented in our GMT software such that testing becomes possible. This code is greatly simular to the we code developed for <Korn>. The ordered Pepperl & Fuchs proximity sensors came in from Farnell, so wiring works can continue. Power supply wiring finished and tested. The voltages on the power supply for the amp without load are +28.4 V and -28.4. SMPS voltages checked and found o.k. Vertical stop designed using a nylon M20 bolt with two nuts. This is how far we are today: Considering to swap front and backside on the design...
• 01.12.2016: Mounting of the valve control board on the upper part of the robot. Start design of the firmware for the valve board.
• 02.12.2016: Work on <Bug> halted due to academic obligations...
• 03.12.2016: Firmware coding for the valve board. Mounting of the Penny & Giles tilt-sensor on the underside of the upper mounting plate. To make this possible we TIG welded a T-piece in 2 mm stainless steel. Wiring of the valves and the lights finished, using 2-pole Weidmueller connnectors to the pulse/hold board.
• 04.12.2016: Extensive work session on the firmware for the valve control board. Controllers 25 (valve attack force) and 26 (release timeout for valves) implemented. Controller 13 implemented for user valve combinations during playing notes.
• 05.12.2016: Start coding for the compressor driver, using a 16 bit motor-control PIC-micriprocessor type 24EP128M0202. We can profit from the experience gained during the design of our <HybrLo> robot last summer. Here we first develop the firmware and test it using our testboard and the Tektronix TDS2-24C oscilloscope, since the design of the final circuitry and the PCB highly depends on the feasability of this firmware. A first circuit design worked out: Filter component values still have to be calculated. Implemented in the firmware so far: pitch generator, ADSR control, Pitch bend, global volume control. Still to be done: noisyness or frequency-jitter, lookup tables for tunings, vibrato (FM) and tremolo (AM). These need separate timers. This is going to be a very nice model for a monophonic synth based on acoustical modelling. The use of analog circuitry (multipliers) for all amplitude related parameters is a great benefit for the sound quality. If we would have implemented it digitally, on a 16-bit processor, this would go at the detriment of resolution and unwanted noise/artifacts.
• 06.12.2016: Design and production of a prototype for the generator PCB. This is what it looks like:
• 07.12.2016: Hardware debug session on the synth board. Sallen & Key filters recalculated. The values are very critical here! This is the transfer function for this type of filter: The component topology is: Firmware rewritten with a better implementation for the ADSR. Function of the velocity byte redefined. The use of a variable sampling rate at 256 times the required audio frequency with the PWM frequency set to 1.8MHz seems to work quite well. Pitch bend implementation, tentative, added to firmware. Left to be done here: basic tuning, vibrato and tremolo. This is the soldered board:
• 08.12.2016: No further work on <Bug>, too many other obligations...
• 09.12.2016: Further develpment of the firmware for the generator: synthesis up and working as well as amplitude modulation. This uses controllers 5 and 6: tremolo. The implementation for the envelope is now as shown here:
• 10.12.2016: Frequency modulation added (vibrato, using controllers 3 and 4) as well as pitchbend with a quartertone range. Left to be implemented: noise/jitter and tuning. Test code in GMT adapted to make testing of the firmware possible. Circuit drawing and component selection worked out for the midi hub board, also serving the sensors and the motor controllers.
• 11.12.2016: Start coding for the gestural possibilities of the robot. All motor control is handled by the processor on the hub board. This board also handles midi parsing and filtering. The TTL midi outputs, output only filtered data required by the valve and the synth boards. The coding promisses to become an awfull lot more difficult then initially thought... All components soldered in place on the hub board.
• 12.12.2016: Two soldering bugs found and removed on the hub board. Midi-parser is working fine now. Sampling for the sensors checked and adjusted to 64S/s. First sketches for the motor behavior worked out in the firmware.
• 13.12.2016: Checking polarity requirements on the motor controllers. Polycarbonate carrier plates for the hub board and the synth board cut out and drilled with all mounting holes. The PCB's are mounted with M3 x 20 cylinder bolts and mounting to the base plate uses M5 x 20 cylinder bolts. In place upgrades of the firmware will be possible without removing the PCB's.
• 14.12.2016: Mounting and start wiring of the hub and synth boards on the base plate. Design and construction of the mounting parts for a 30-pins connector to make user-friendly disassembly of the upper part from the lower part possible. This is the wiring:
• 15.12.2016: Wiring works on the large connector, according to drawn schematic and wire color coding. . Contacts A8, A7, A6, A5, A4 are left unconnected and may be used for later modifications and features. Contact C8 is wired but has no function as yet. Contact C7 is implemented on the hub board as a switch. Cable bundle constructed such that 360 degrees rotation is possible. Motor wiring done for both motors. An entire day of wiring this was... Almost done. The only thing left to be done now is the wiring to the power supplies on the underside of the robot and the sensors for the horizontal movement. This we can only do if the ordered dented wheel flows in from MEA. Chances are <Bug> can undergo its first tests tomorrow...
• 16.12.2016: Wiring on the underside finished: On firing it all up... nothing happening. Power supplies remain dead. Apparently we miswired the 48V supply, thus blowing up the 24V supply... The 48V supply survived the shock. The 24V supply appears unrepairable. Replacement part (XP Power LCL150PS24) ordered. After fixing some minor wiring bugs on the amplifier, we got this part working. The tuning is about a quarter tone off however. This can be fixed in the firmware. The sound quality in the low register is quite good. Needs more sharpness in the trebble though. The valve board appeared to have a short on the +24V line. Short found and removed. Now waiting for a new power supply.
• 17.12.2016: Back to the sound driver firmware: bugs killed in the PWM subsystem. Wave lookups improved and extended. Program change command implemented to change the lookuptables and thus the timbre of the excitation wave.
• 18.12.2016: Formant filter soldered on the synth board. Circuit drawing updated. More wavetable calculations added to the firmware. Attempts to add noise to the wave so far unsuccesfull. Modulation of the sampling rate in the timer1 interrupt crashes the PIC. Further comments can be found in the source code for the synth firmware.
• 19.12.2016: Test code in GMT added such that we can test the optimum values for realistic flugelhorn sound in controllers 100 to 108. Assymmetrical sine waves implemented in the firmware. Replacement 24V power supply came in. We replaced it right away. The aluminium dented wheels also came in from MEA. These have to be drilled out on the lathe and threaded (M4) for fixation on the shaft.
• 20.12.2016: Continued testing of the valve firmware and mechanics. Work on the horizontal and vertical movement mechanics. Vertical movement dented wheel extended with a mover to transmit movement to the flugelhorn mounting plate. This piece mounts on the wheel with a single M6 x 16 bolt. Vertical movement now works fine, but the motor is not powerfull enough to perform the full mechanical traject that would be possible. Horizontal motor still not operational. We still have to mount the sensors and make up our mind as to frontside or backside...
• 21.12.2016: Further refinements on the coding for the vertical motor in the hub firmware. Vertical motor speed implemented with controller 33. Start work on the horizontal movement. Sensor holder made to the 'new' front side. This sensor assembly mounts on the base plate with two M5 x 10 cilinder bolts. The head of the left bolt is grinded off sideways in order to fit in the tight space next to the 48V power supply module. Sensor output voltages checked: 409mV at rest, 1.5V when touched with metal. Code in the midihub firmware adapted to these values. Decided not to provide an automatic calibration procedure on cold boot. Calibration will be performed in use when both of the sensors are touched once. We do assume the robot to be in a more or less central position on startup. Considering to add and construct a 'hard' blocking stop as well for safety. With some minor modifications, it's still possible to swap front and backside. However, we will not leave this option open for normal users as it involves mounting the sensors on the other side and precise readjustment of the positions.
• 22.12.2016: Setscrews honed such as to fit the 25 mm vertical tube. Construction of the bearing on the underside.
• 23.12.2016: Mounting of all components for the horizontal movement. Column shortened with some 140 mm, so the total height of the robot (with the flugelhorn in a horizontal position) becomes 1070 mm.
• 24.12.2016: Testing horizontal motor movements. We are running past the sensors... so there must be a bug in <Bug>...
• 25.12.2016: Continued work on H-motor coding. All sorts of problems rise up. The sensors do work now. but there is interference between vertical and horizontal motor and the horizontal motor seems to refuse change of direction commands.
• 26.12.2016: Further coding for hub board.
• 27.12.2016: Serious problems with the implementation of the dual motor controlls on the hub board. Vertical movement works o.k. now, but horizontal remains very shaky. Autocalibrate procedure for horizontal movement added. During calibration, the robot will not receive any midi commands!
• 28.12.2016: After yet another day of working, we finally got it working o.k. We suffered from spikes on the analog sensor inputs, so we added buffer caps and bypasses. Now we also implemented acceleration and decelleration to the horizontal movement. Interference between both motor movements is gone now.
• 29.12.2016: A snippet of video on the working <Bug> robot posted on Facebook by Moniek Darge. Further firmware improvements. Design and construction of a blade spring mechanism to limit up-down movement elegantly. The up down traject can now be enlarged a bit. Definitive allignment of the valve mechanism and the tractures. M4 bolts secured with red Loctite lacquer. Considering to make a protective cap over the circuitry, taking into account the envisaged use of the robot.
• 30.12.2016: Construction of the polycarbonate cover. This is fixed on the base plate with one long M12 x 120 bolt through a 98 mm long distance holder tube (30 mm diameter) and screws in the long hexagonal nut holding the wheel on the backside. On the side of the power connectors, it fits the base plate with an M6 X 120 inbus bolt and a nut through a 12 mm diameter distance holder tube. At the other side, it leans on a piece of L-profile mounted on the horizontal crossbar and holds with a single M6 x 25 bolt. Also, we made a protection over the advantec motor controller using a piece of polycarbonate mounted with an M4 x 35 inbus bolt. <Bug> demonstration for Laura Maes. Airtighting the connection between mouthpiece and membrane compressor and adding felt to the bugle supports.
• 31.12.2016: First midi testfiles made and tried. Running from Sonar sequencer. Mute circuit designed. The felts we inserted to fix the flugelhorn seem to cause more friction on the valve mechanism. We have to readadjust this.
• 01.01.2017: Firmware for valve board revised and improved now version 1.1.
• 02.01.2017: <Bug> demonstrations for Dominique Pauwels and Johannes Taelman.
• 03.01.2017: Start design of a flightcase. Work on some demo files in midi.
• 04.01.2017: Welding of a frame for the <Bug> flightcase. This frame TIG-welded with normal construction steel L-profiles. Bottom panel sawn. In need of more flightcase wood...
• 05.01.2017: Demonstration and some tuition for Brecht from LOD. Further construction of the flightcase. All panels sawn and mounted. Hinges and front door left to be done. First photoshoot with <Bug>
• 06.01.2017: Continued work on the flightcase, with assistence from Mattias Parent.
• 07.01.2017: Construction of the hinges and the locking bolts (M12). Holding studs sawn to hold the robot in place inside its case. These blocks are mounted with wood screws and siikicon glue. they hold the robots in place in the flightcase during transportation., Red felt glued on the right side of the case to prevent scratches on the flugelhorn.
• 08.01.2017: Finalizing the flightcase. Mounting of the blocking block for the upperside inside the case. Glueing of padding foam on the case inside. Mounting of the carrying handles. Instructions on how to get <Bug> in its flightcase added at the end of this page.
• 09.01.2017: Tightening profile mounted on the upperside of the flightcase door. An 10 mm inbus hexkey is required for opening and closing the case. Weatherstrips added on case lid.
• 10.01.2017: <Bug> delivered to LOD
• 19.01.2017: <Bug> played it's demo piece in the Logos Robotorchestra concert tonite. Tomorrow it leaves again for LOD.
• 27.01.2017: Odd behaviour reported by Dominique Pauwels, when setting controller 17 to zero.
• 15.06.2017: Bug back at Logos. Checked and found o.k.
• 01.08.2017: <Bug> on the road again in the 'Helden' production by LOD.
• 13.07.2018: <Bug> played on the opening of the Ghent Feasts at Sint Jacobs, in tandem with <Korn>.
• 29.07.2019: On its return from its appearance at Tomorrowland, the blade spring limiting vertical movement was found broken. Repair required. The flight case as well needs some repair, as it was left in the rain...
• 30.07.2019: New blade spring made and fitted on the robot.
• 01.01.2020- 12.01.2020: <Bug> on tour to France in the 'De Helden' production with LOD.
• 28.01.2020: <Bug> is back in the robot orchestra.
• 04.02.2020: Considering to improve the firmware in <Bug>, taking advantage of our experiences with the newly build <Flut> robot.
• 07.12.2020: PIC24 reprogrammed with corrected firmware: we had a timer 3 reading bug causing amplitude glitches every so often. Some sluggishness noticed in one of the valves. Mechanical wear we presume...
• 08.12.2020: Complete revision of the 24EP128MC202 processor firmware. Newly implemented controllers: CC1 (wind-noise), CC15 (ADSR-period), CC20 (tuning). Documentation upgraded. Controller 80 is still left to be implemented.
• 09.12.2020: Controller 80 implemented. Test code in GMT updated, including all default controller settings. Hub board and Valve board firmware upgraded to comply with the upgraded Proton compiler as well as with the newly implemented controllers. All interrupt handlers improved. Eye lites now work also automated in sync with note-on/off commands. <Bug> is fully up-to-date again.
• 19.01.2021: Some bugs discovered in the implementation of vibrato. Firmware upgraded. No changes for users required.
• 01.08.2021: Apparently, under some conditions, the valve-control board can crash. On cold start, it reboots normally, so it's difficult to analyse where it may fail.
• 27.09.2021: After playing uninterrupted for 25 days at the SMAK museum, <Bug> returned to the robot orchestra in perfect condition.
• 11.03.2023: Taking inspiration from the <Bug> design, to further develop <Kazumi>.
• 14.09.2023-02.10.2023: <Bug> joined the Zeroth Law production for the Deutsche Oper in Berlin. It returned safely to Ghent.

To do:

• - removing the noisy glitches on cold boot. This is a hardware problem.
• - Construction of a 'hard' stop for the movements, such that the robot does not ruin itself on fault conditions.
• - Corrections and improvements based on user feedback.
• - Implement a delay between fingering and membrane compressor drive.
• - Apply delayed vibrato, as done on Trumpeter, if possible here.

(Terug) naar logos-projekten: projects.html

Terug naar Logos' index-pagina: index.html

Naar Godfried-Willem Raes personal homepage...

Naar katalogus instrumenten gebouwd door Godfried-Willem Raes

Last update: 2023-10-07 by Godfried-Willem Raes

The following information is not intended for the general public, but is essential for maintenance and servicing of the robot.

Technical drawings, specs and data sheets:

Proximity sensors: Pepperl+Fuchs, NBN4-12GM40-Z0 (Horizontal movement sensors)

Tilt sensor: Penny+Giles STT280/60/P2 (Datasheet: STT 280/60/P2. ) (Vertical movement sensor)

Horizontal motor controller (G103V)

Horizontal stepping motor: MAE HY200 3424 170 ( 1,7A - 2 Ohm with windings in series, 2.38A - 1 Ohm, with windings in parallel) or MAE HY200 3424 470 ( 4.7A - 1 Ohm with windings in series). Maximum voltage on these motors is 90 V.

Vertical stepping motor: Nanotech (SMC42)

Valve lookup tables (according to acoustic theory)

Power supply:

The power supply for the power amp determines maximum power. With the values shown, we have a maximum output voltage from the power amp of 17 Vrms. Our load -the membrane compressor- has 16 Ohm impedance and thus maximum deliverable power will be 17 W rms.

The 24V/150W power supply is an enclosed unit made by XP-Power, part nr. LCL 150PS24 with medical specifications. Farnell order nr.:1738313. Datasheet.

Wiring & circuit details midihub and motor-control board:

Circuit drawing:

PCB design for this circuit:

PCB:

Soldered board:

Firmware for the hub board:

• source code Bug_Hub.bas
• hex dump: Bug_Hub.hex

Pulse-hold board for the valves and the lights:

PCB:

Populated board:

Firmware for this board: Bug_Valves.bas Hex dump for this board: Bug_Valves.hex

Compressor driver board (monophonic synth with acoustical modelling)::

Circuit drawing:

PCB for this circuit:

Soldered board: (Formant filter not yet mounted on this picture).

Firmware for this board: PIC24_Bug.bas (source code for Proton24 compiler) and hex dump: PIC24_Bug.hex

Wiring of the connector between moving upper part and lower part:

Flugelhorn: Selmer 'Vincent Bach' horn.

Mouthpiece: Bach Flugelhorn 3427C, adapted on the lathe to fit the membrane compressor.

The <Bug> robot has a special flightcase. Following pictures are made to clarify the way <Bug> should be placed in its flightcase:

References:

Beauchamp, J.W. "Analysis and Synthesis of Cornet Tones Using Nonlinear Interharmonic Relationships". In: j-aes, volume 23, number 10, pages 778--795, 1975.

Beauchamp, J.W., "Analysis of Simultaneous Mouthpiece and Output Waveforms of Wind Instruments" . In: j-aes, 1980, Preprint No. 1626,

Benade, Arthur .H., "Fundamentals of Musical Acoustics". Ed.: Oxford University Press, 1976.

Fletcher, N.H. & Tarnopolsky, A. "Blowing pressure power and spectrum in trumpet playing" In: J. Acoust. Soc. Am., volume 105, number 2, part 1, 1999.

Martin, Daniel W., "Lip vibrations in a Cornet Mouthpiece", In: J.Acoust.Soc.Am. vol13 . 1942

National Semiconductor, LM18298 Dual Full-Bridge Driver, datasheet. April 1992

Raes, Godfried-Willem, "Kursus Akoestiek", Ghent University College 1982/2014, Internet: http://www.logosfoundation.org/kursus/4023.html

Raes, Godfried-Willem, "Logos @ 50, het kloppend hart van de avant-gardemuziek in Vlaanderen", ed. Stichting Kunstboek, Oostkamp, 2018

Raes, Godfried-Willem, "Expression control in musical automates", 1977/2021,

Raes, Godfried-Willem, "Trumpeter", 2021

Rose,Nicholas and Holloway, Damien, "Finite element modeling of brass musical instruments', in: Proceedings of Acoustics, Fremantle, Australia 2012.

Smith, Bob H., "An Investigation of the Air Chamber of Horn Type Loudspeakers", in: The Journal of the Acoustical Society of America 25, 305-312 (1953); https://doi.org/10.1121/1.1907038

Cost calculation for <Bug>:

1. Parts and components

2. Consumables

3. Depreciations on tools and equipment used (calculated over 2 months)

4. Labor and design (45,-/hour)

07-09.12.2020: 3 days of work on firmware upgrade.

[EOF]