In our quite large collection of musical instruments at Logos Foundation, we had since very long a bunch of different Zithers of different kinds: German made ones with some 48 steel strings, a Han-Koto as well as a few zithers stemming from 19th century musical automatons. The latter I digged up from the cellar of my grandmothers (Helena Strunz), 1899-1996) house. They were completely rusted... For many years we had thoughts about finding a way to turn at least one of these zithers into a musical robot. The problems, as soon as we started experimenting and designing it on the drawing table seemed very insurmountable. The main reason being the too close spacing of the strings. No matter what kind of plucking mechanism we imagined, it either took too much physical space (using bi-directional solenoids) or it would be way too slow and monophonic (using a sledge mechanism with a single plectrum) to allow automation of all of them. So the idea was dropped for many years.

In 2013 we were asked by Osama Abdulrassol to consider the automation of an Arabic Qanun. The same problems we had analyzed already reappearing and some new ones in top: the Qanun uses microtonal pitchchanges using a mechanically pretty simple system (mandalar) , but again due to size/force constraints, very difficult to automate well. Although in may 2013 we decided to have a throw at it, and started making a first prototype for a plucking mechanism. The perspective being to also make the zither itself rather than trying to automate an existing instrument. For the first time in our career as a robot designer, we decided to construct the automation mechanism prior to and fully independently of the actual sounding instrument. This entails that we designed the actual instrument only after the mechanism for the plucking was fully up and running. Thus, first a prototype plucker was made using a solenoid assembly from Syndyne. Detailed information about this design and its final failures can be found on the legacy <Zi> webpage. The mechanism never had enough force to really pluck the strings. Hence we recycled the mechanism and thus the <Tinti> robot was eventually born...

A second attempt to build a plucking mechanism was made, making use of bi-directional solenoids with permanent magnets. These solenoids are stable in either of their end positions and they only require a pulse of changing polarity to make them change position. As this type of solenoid could not be obtained with an anti-rotation shaft, we decided to design round plectra with a 2 mm central hole for plucking the strings. However, once more, it was a complete failure. The solenoids only develop some force in either one of their end-positions, whereas for or a decent plucking mechanism, most force is required in the middle of the trajectory. Experiments we carried out prove that presumably the best pluckler mechanisms can be made using stepping motors. As a consequence, spacing between the strings has to be quite large, say at least some 50 mm.

Third time, good time as a Dutch proverb claims, we forsake the idea of plucking and turned back to an old German made zither in mint condition. Instead of plucking, now we used a hammering mechanism. So, organologically speaking, we turned the project into some kind of cymbalon, a hammered citer after all. The hammering mechanism was derived from the design of the 'poltergeist' we designed for our <Ubu> robot. It uses small Binder solenoids to load a spring and on release of the power, the anchor bounces back against the string. Velocity control is very well possible by varying the pulse duration's. Very short pulses and quite high voltages ( 48 V) are mandatory here. To understand the limits of velocities versus repetition speed, it might be good to understand this graph: The solenoids, rated for 12V at 100% duty cycle work here with pulses of varying duration (proportional to velocity) and a voltage of 48 V. So in order to avoid overheating of the solenoids we have to make sure every pulse (tv) is followed with a period of inactivity (ts) at least three times the duration of the pulse.

Only when integrated in the context of our robot orchestra with its wealth of varied sensor systems allowing full interactivity with gesture as well as audio, this automate will become a true robot. That's after all were its destination is to be sought.

This is the ambitus for the instrument when using the original steel strings. This also is the ambitus as implemented in MIDI.

Lights mapped on notes as:

• note 120: white frontal LED strip (automated when CC69 is 1)
• note 121: red LED lite (3 modules) between instrument and chassis plate (automated when CC69 is 1)
• note 122: red LED spotlights to the ground. Velo steers PWM.
• note 123: two 5W tubular tungsten lights left and right, front-sides of the robot. ON/OFF only.

Midi channel: fixed to 4 (counting 0-15).

Note Off: Not at all required, unless automated repeats are used. In that case, note off will stop the repetitions for that note..

Note On: Implemented for all notes in the range. Velo-byte is used for the striking force. When repeats are active, note that all subsequent strokes will be performed with the velocity as set with the first note-on command. The lights are also mapped on notes, but make use of a range outside the normal range of the zither.They are mapped on notes 120 to 123.

Channel aftertouch: can be used to let notes repeat automatically. The parameter value sets the repeat frequency. Range: 4 Hz to 20 Hz. The command can be sent even prior to note-on commands. The value sent will be preserved until reset with either a new channel aftertouch command or an all notes off command.

Controller #66: Robot on/off switch. Sending a power off command (Ctrl 66 set to 0) will cause a reset of all controllers to their default start up value. Also settings for note repetition will be reset.

Controller #69: sets automation of the lights to on or off. By default this is set to ON.

Controller #123: all notes off. Stops note repetitions and dims the lights.

Technical specifications:

• size: width: 450 mm, depth 540 mm, height 500 mm.
• weight: 12 kg. (estimated)
• transportation: needs a flightcase.
• power: 230 V ac / 200 W (peak, not playing: 5 W, normal playing < 40 W)
• Ambitus: 2 octaves. [midi 60 - 84] + 6 chords: C, G, A, D, E, F mapped on midi notes 40 to 50
• Needs tuning before use!
• control: MIDI-input, 4 MIDI-Thru (differential)
• Insurance value: 4.500 Euro.

Design and construction: dr.Godfried-Willem Raes

Collaborators on the construction of this robot:

• Hans Roels
• Bert Vandekerckhove (workshop assistant)

Music composed for <Zibalo>:

Godfried-Willem Raes: 'Perpetuum Mobile' (6') [2025]

Traditionals: 'Silent Night', 'O Tannenbaum'

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

Back to Main Logos page:index.html

To Godfried-Willem Raes personal home page...

To Instrument catalogue

Construction & Research Diary:

• 03.01.1969: We got a Zither as a birthday present from an aunt. It was purchased at the 'Cnudde' music store, in those years housed in the Volderstraat, at the left side of the University aula. The zither, branded 'Jubeltoene' was made in the DDR in the late sixtees..For many of the following years it was used as a sound source for the 'Logos werkgroep voor geengageerde avant-garde muziek'.
• 03.03.2009: We were donated no less than 1500 telephone relays... Examining weather or not these could be used in the automation we could make of our Zither.
• 09.03.2009: Prototypes assembled using some 36 relays, with welded-on beaters. Operation is sluggish and mechanically unreliable.
• 03.04.2010: Experiments with bi-directional solenoid driven pluckers. They take up too much space, but do indeed work quite well.
• 19.12.2013: Experiments with Maxon DC motors, diameter 10mm and 13mm. This works fine, although the mechanics for making this work reliably on an existing Zither may become pretty complex and difficult in alignment.
• 13.05.2014: Bi-directional solenoid prototype prepared as plucking mechanism. This is a modification of an August Laukhuff part used for registration knobs on pipe-organs (Catalogue nr. 3 002 00). The original manufacturer appeared to be Syndyne, and after checking their catalogue it became apparent that we could also get these parts without an angled anchor. We ought to proceed a bit faster with this design as we are urged to do so by 'De Centrale', who commissioned it for a collaborative project with Abdullah Abdulrasol.
• 14.05.2014: Syndyne contacted for custom made bi-directional solenoids. As we plan to use pulse/hold boards for this instrument, we estimate that the hold voltage should be no higher than 6 Volts, whereas the pulse voltage may be very well be risen to 60 Volt. This ought to give a good range for the velocity control. Plucking tests performed on a balalaika. In fact with just 3 pluckers we could automate the balalaika but there is not enough space on the neck to accommodate the required solenoids for the frets.
• 15.05.2014: Construction drawings for the complete plucker mechanism. Sketches for the tuning mechanism: we consider using mandolin tuning pegs in rows of four.
• 04.07.2014: All wiring done. Start of the first tests. As we didn't even start making the actual stringed instrument, we perform the tests with the above mentioned Zither, placed in an upright position. We use a lab power supply adjusted to 6 V for the hold voltage and a -36 V power supply for the negative pulse voltage. We observe excessive bouncing of the plectrum/anchor combination. The pulse duration's have to be scaled down in the firmware.
• 06.08.2014: Tests with different plectrums and materials that can be used as plectrums. Design of a new laboratory power supply with high current and high voltage range. It should withstand highly inductive loads, this not being the case for commercially available laboratory power supplies...
• 09.08.2014: Here is a link to our app note for this new hefty power supply.
• 06-15.09.2014: Experiments and research with regard to the best plucker shape, material and construction. The alignment promises to become an extremely difficult and tedious task. So far, bone material as used on guitar bridges seems to work best.
• 01.12.2014: Mounting of plectrums confined to Mattias Parent.
• 09.04.2015: Experiments performed with increased mass of the anchors in order to increase plucking force. We can use small cube neodium magnets on the one side, or try to obtain the equivalent through a different coding in the firmware. We wired up a prototype board to perform these experiments. The result were not worth the effort though.
• 23-24.04.2015: Further research and experiments by Mattias Parent.
• 06.08.2015: After many hours and days of experimenting we came to the conclusion that the mechanism for the pluckers is unworkable. So we started a new design, using linear bi-directional solenoids.
• 22.08.2015: Plectra made from nylon staff material on the lathe. Experiments with plucking carried out on an old zither with steel strings. The plucking does work indeed but we notice a click-noise on each position change of the solenoid. Seems to be unavoidable.
• 23.08.2015: Further redesign of the Qanun. Considering to use the prototype 1 assembly for another robot, maybe <Tinti>, using tiny bells...
• 30.09.2015: The missing and ordered bistable solenoids came in from Conrad, so the works on the new plucker mechanism can continue.
• 02.10.2015: Helene Wolf contacted to help out with the design and cutting of a new comb. As a cembalo maker, she ought to have the experience.
• 03.10.2015: Prototype board for the new solenoids etched, drilled and soldered. We can test the first 8 solenoids if we have also the firmware ready. Work for tomorrow...
• The PCB here is at 200% and should be reduced to 50% before printing and exposing.
• 12-28.12.2015: Construction of a new bridge by Helene Wolf.
• 11.01.2016: Work on Zi taken up again.
• 26.01.2016: New bridges made by Helene Wolf presented and found to be excellent. Start restringing the instrument.
• 19.02.2016: Sliding feet of the soundboard cut open again as the mechanism was too stiff.
• 24.02.2016: Adjustment studs drilled and mounted to allow precise positioning of the soundboard versus the plucking mechanism. For now using M6 threaded rods, but this may become a larger size later on.
• 15.10.2016: All experiments with suitable plectrum material failed. The problem that plagued our first and rejected design is reappearing: the solenoids do not have enough power in the mid position. Furthermore, it appears unfeasible to adjust the plucking distance reliably. As in top of all this bad luck, we also lost all funding due to a corrupt, bad and big-ego oboist, Piet Van Bockstal that issued a negative and canceling advise for funding against the Logos Foundation. So, the project has to be placed on hold.
• 09.10.2024: <Zi> project taken up again. Considering to use linear motors instead of solenoids. These for sure have enough power but they are pretty slow... In principle we could use the same electronic circuits to drive the motors. The firmware needs a complete revision and note repetitions seems to be impossible, unless we allow only very slow repetitions...
• 10.10.2024: Two new PCB's etched and drilled.
• 11.10.2024: One board soldered. PCB layout improved.
• 12.10.2024: Second board soldered. Tests of the new firmware. Abandoning all existing Zi hardware...
• ---------------------------------------------------------------------------------------------------------------------------------------------
• 12.11.2025: Work on the zither taken up again. The very successful poltergeist mechanism we made for <Ubu>, gave us the idea to use some of the hundredths of Binder 12V solenoids we have in stock as rebounce hammers. To be successful we have to replace the springs with a much stronger type. The original springs have following sizes:The thread for mounting the solenoids is M10x1 (metric extra fine). We made a paper drawing (scale 1:1, or real size) of the string layout, useful for the drilling works and precise placement of the solenoids:
  
• 13.11.2025: It becomes mandatory to equip the Binder solenoids with much stronger springs. Experiment performed: using a pressure spring (Fabory stock item) 0.5 x 6.5 x 20, shortened to 7.5 mm (4 turns), works well with a supply voltage of 48 V. Duty cycle should be kept smaller than 5%. Current is 240mA, the coil resistance is 200 Ohms. Peak power follows as 11.5 Watt. For the control we can use some older pulse-only PCB's. These were originally designed for robots such as <Troms>, <Puff>, <Autosax> etc. As there are 25 chromatic strings, we can even save some pins for the ICD. Each board is designed for 16 pulse-only outputs. Here is the schematic:
• 14.11.2025: Work on the 15-solenoid assembly. Not an easy undertaking under narrow size constraints. This is the rudimentary construction, before trimming and adjusting to available space. The solenoids, although new, are dated 1967 and 1968. This explains the corrosion on the bodies. Nevertheless, they work perfectly well. There are no traces of rust on the anchor's of the solenoids. A plastic cap has protected the anchors as well as the spring against moisture.
• 15.11.2025: Cut out through the Zither finished. Note that we did this without even removing the strings. A risky undertaking... As for now, the solenoids assembly mounts from underneath the zither and is fixed to the upperplate (the soundboard) with three M3 bolts. Later on we may mount a few more bolts as the way it is now, feels a bit shaky. After mounting we noticed that all the strings went pretty much out of tune. Probably we undermined some of the structural strength of the instrument by removing too much wood at the right hand side. We had no choice though, as we need some space to mount the solenoids for the top four strings. To give the instrument some new and alternative structural strength, we made a curved side plate in stainless steel, glued and mounted with screws on the right side of the zither:
  16.11.2025: Curved plates (35mm wide, 2mm thick) made and mounted for the left side as well as for the top of the instrument. All three plates have a welded on M6 high nut for mounting of the instrument on the chassis (to be designed still). We foresaw two M6 high nuts on the left side reinforcing plate, as this could come in handy in case we decide to automate the chords as well. As today was crafts-day, we opened up and welcomed quite some visitors to our workshop.
  17.11.2025: Tests and designs for a hammering mechanism for the six chords. We can use tubular solenoids, provided we equip them with an anti-rotation shaft or another mechanism to avoid rotation of the anchors. We can use piano hammers, provided we give them a more or less flat face on the felt hammertop.
  18.11.2025: Construction of the main chassis plate in stainless steel. Thickness 2 mm. Design of a suitable wheelbase.
  19.11.2025: Frontal double wheel assembly prepared.
  20.11.2025: Welding works. Chassis and wheelbase welded together. Mains power entry mounted and wired. This IEC entry has a fuse holder. We should not forget to insert a fuse in due time... As the pretty flat structure had a tendency to bend, we TIG-welded a border, 10 mm high, on one of the sides. This ought to make it very ridged now.
  21.11.2025: Development of the firmware's for the hub and the pulse boards. This is the mapping for the pulse boards:
  22.11.2025: Full implementation of <Zibalo> in GMT, including test and torture code.
  23.11.2025: Extensive testing of the hub board using GMT. This is the wiring detail: The pulse range for the chord-hammers is now 3ms to 9.5ms. Loop speed for this microprocessor is minimum 98kHz under maximum load conditions and maximum 128kHz at no load.
  24.11.2025: Recycled 16-pulse output board patched for use in <Zibalo>. This board, servicing note 60 to 72, was firstly used for <Autosax> in 2005. We mounted an ICD programming connector, glued to the PCB with epoxy rosin. The wiring and connections now look like this: We used MUR620C double diodes here. The mosfets are IRL640. The microprocessor is an old 18F2525. First version of the firmware uploaded in the microprocessor. On board2, the mosfets used are PHP79NQ08LT types (Nexperia) . This type is now obsolete but we had enough stock for this project. The short specs are 73 A - 75 V , Ug = 5 V, Ron = 0.014 Ohm. On board 2, the diodes are the same as on board1 : MUR620C.
  25.11.2025: Finishing soldering work for pulse board 2. The microprocessor is an 18F2525, the last one of that type we had in stock.
  26.11.2025: Start mounting of the PCB's on the chassis. Placement of the lights mapped on 120, 121, 122. No idea yet for lite 123... Could be tungsten ( 2 x 24V with E14 sockets) Mounting of the 12V - 2A SMPS (encapsulated Vigortronix unit) on the chassis.
  27.11.2025: Wiring of all components on the chassis. We decided to mount two tungsten bulbs with Festoon sockets: 2 x 24 V in series, connected to pulseboard on the last output after note 84. This tungsten light is mapped on midi note 123, on-off only.
  28.11.2025: M6 shock absorbers ordered from MEA. Finalizing work on the distance holders between zither and mounting plate, such that we can reach the connectors with our hands. Should we make a new frontplate?
  29.11.2025: all wiring finished. Boot test passed, as yet without midi. Anchors equipped with newly cut springs and placed back into the solenoids. If we are careful this can be done without removing the solenoid assembly.
  30.11.2025: Complete tuning session. Midi connected and testing with the test-code in GMT. Scale test passed, although we have some force irregularities due to unequal springs. At first sight, it seems we can even reduce the scaling of the velo-pulses. Very fast note repetitions give no problems at all. Firmware changed: channel aftertouch fully implemented now for note repetitions. Minimum repetition speed will be ca. 4Hz, maximum ca. 20Hz.
  01.12.2025: Operating conditions as implemented in the firmware documented in a graph: We finished implementing fast repetitions, using the channel aftertouch command. Repetition rates faster than 20Hz seem to be very well possible.
  02.12.2025: Further work on the firmware. A mysterious bug puzzles us and we lost the full day on debugging... in vain.
  03.12.2025: Two bugs found: the first one is actually a compiler bug: if the code between IF ... THEN... ELSE get's over an undocumented limit, the entire code stops working... The second bug was our own mistake: we wrote in a single instance idx = SortTimers, where it had to be idx = SortTimers (). When we omit the empty brackets, the function always returns &HFF in idx. All chips programmed with the new firmware. Further work on the hammers for the chords. The solenoid anchors have to get an M2.5 thread such that we can fine-adjust their vertical position. The wood on the hammers get an internal M2.5 thread as well. PTFE-spacers made to work as anti rotation shafts for the hammers. Once more we have a job for a metal milling machine but unfortunately, we sadly cannot afford one... It would make the perfect tool for constructing an adjustable hammer assembly.
  04.12.2025: Further work on the hammer assembly. Standoff plates welded on to make sure the assembly can be mounted parallel to the soundboard. L-shaped solenoid holder welded on. Solenoids bolted on. Hammers roughly cut to size. Wiring of the solenoids. We use color codes for the wires: black. brown, pink, orange. yellow, green for C, G, F, D, A, E chords or resp. 48, 43, 41, 50, 45, 40. The red wire is 48V. Very first test with 1 solenoid passed.
  05.12.2025: Finalising the hammer assembly. We labelled the hamers, as we had to file the felt different for every chord. Note that all four strings in each chord have a different thickness. Complete tuning session. Measuring the velocity ranges for all the solenoids such that we can optimize the firmware. First real 'portrait' pictures made of the newborn robot:
  06.12.2025: Bizarre failure... We checked the velocity ranges for the Binder solenoids and noticed that we need abour 4 times larger pulse durations as compared to yesterday... Also, the hammers refuse to pulse properly... Checking the 48V XP-power SMPS, we measured a quite shaky 22 V. Even if we remove all load, the output voltage stayed that low. The 'power-good' LED lights up! Clearly the (new!) power supply - XP-Power type VCS100US48- failed. This obviously explains the range change for the Binder solenoids. However, it does not explain the refusal of the hammers steered by the hub board. On the testboard, the firmware seems to do exactly what it is supposed to do: giving pulses in the range3.2ms to 9.5 ms. We ordered new power supplies from Farnell, avoiding XP-Power devices this time, we choose Meanwell Medical RPS-200-48C. Maybe we should build in a small analog voltmeter to monitor the power supply...
  07.12.2025: As in absense of a 48V power supply we cannot proceed in finalizing the project, we decided to build a 'classical' analog power supply from components we had on the shelves in the lab. Once more it confronts us with an observation that puzzles us since our childhood, when we started our carreer in electronics: according to the textbooks, the rectified voltage measured over the capacitor is 52.9V whereas we would expect (35V - 0.4V) * SQR(2) = 48.93 V, or if we ommit the voltage drop over the diode, 49.49 V. How should we explain the ca. 5% difference? Anyhow, we wired the new analog supply and everything works again as before. A mechanical problem has risen up now: the transformer with its weight of ca. 2.5 kg tends to bring the three-wheel construction out of equilitrium. We can only bring it back in balance by adding a 2 kg weight on the extreme right side...
  08.12.2025: Balancing weight in red copper made: 50 mm massive staf, cut to 12 cm gives 2 kg. For pure copper we have 8.9 kg / liter. Hub firmware improved and uploaded. Examining the problem with note 82 on the second pulse-board. Carefull analysis revealed some of the pins if the 18F2525 chips are not functioning. The chip is to be disposed off. We replaced it with a new 18F2620 chip and everything seems to work as it should now. We decided to mount a panelmeter on the front, such that we van monitor the 48 V power. We used an old Kyuritsu A-meter, removed the internal shunt and calculated a new series resistor to make a full scale 50 V DC meter. As the metering components provide a leak and discharge path for the 10mF capacitance, the voltage will slowly decay to zero after a power shut down. Further testing has led to the decision to change the velocity scaling for the pulse boards. The low side is o.k. now, but the range 70 to 127 is useless.
  09.12.2025: The bug with the hammers is persistent but difficult to provoke. The velo-pulse mechanism fails every so often. Timer refresh instruction added in the check-timers routine. This seems to work at first sight. Apparently the midi-parser causes time errors when a lot of midi messages are to be sent out. Two christmess files prepared for the concert on thursday 18th of december: 'Silent Night' and 'O Tannenbaum'.
  10.12.2025: Glueing a piece of thick felt to dampen the sound of the hammers on their return after striking. Demo piece for <Zibalo> composed: 'Perpetuum Mobile', duration is six minutes.
  11.12.2025: All lookup tables for the velocities changed to quadratic instead of linear: this gives a much improved scaling. Frontal light on the scale of the voltmeter added. This LED assembly has a 2500 Ohms wirewound series resistor, such that the current is limited to ca. 13 mA. This light follows the presence of the 48 V power and as such reflects controller 66. <Zibalo> demonstrated for Mattias Parent. We found a quite elegant solution for the required balancing weight: instead of mounting a massive copper cylinder, we can place a heavy permanent magnet (taken from an old loudspeaker) that can be usefull for storing the ferromagnetic tuning key inside the robot.
  12-15.2025: work on orchestrations with a part for zibalo in preparation for our end-of-the-year concert.

   

• TO DO:.

  • cutting and placing the best possible springs in the Binder solenoids.
  • retuning
  • Final scaling of the velocities and repeat frequencies to realistic values.
    

  ---

  Terug naar Logos' index-pagina: index.html

  Naar Godfried-Willem Raes personal homepage...

  Naar katalogus instrumenten gebouwd door Godfried-Willem Raes

  M&M orkest

  ---

  Last update: 2025-12-15 by Godfried-Willem Raes

  The following information is not intended for the general public nor is it required for composers wanting to make use of our <Zibalo> robot, but is essential for maintenance and servicing of the robot by our collaborators. It also might be usefull for people that want to undertake similar projects. Feedback is mostly welcomed.

  Technical drawings, specs and data sheets:

  Power supplies:

  • +5 V DC - 2A (Logic and microcontrollers), XP-Power ECE10US05
  • +48V DC - 4.5A (solenoids), analog power supply. [also in stock as a spare part: Meanwell Medical RPS-200-48C. ]
  • +12V DC - 2A (lights) Vigortronix unit.

   

  Wiring & circuit details midihub board, servicing the six chord hammers and the white frontal light.:

  

• 

   

  Circuit details solenoid driver boards:

  

• The three LED-strips connected on the GND-PWM connector are OMC LTD01209 LMMR1, 12V, RED.
• Pulse board 2: The red spotlights connected on the GND-PWM connector are Paulmann GU5.3 2.5W and 3W. On the output marked NC we connected two 24 V -5 W bulbs with festoon sockets, connected in series. The sockets are BS242 (JKL Components, part# 2131124.
• 
• Analog 48 V power supply:
• LED-scale light: LFBML-24V-6S167, 1 flexible 6-led unit. www.ledxon.com, 60mA / unit. With the 2k5 resistor in series, current drawn is limited to less than 15 mA.

• ---

  
  Criticism:
• Hammering the zither strings indeed does work, but the sound volume is pretty low.

• ---

  References:

  Meanwell Medical RPS-200-48C. (datasheet)

• Microchip PIC 18F2525 manual

  RAES, Godfried-Willem, "Expression control in musical automates", 1977/2025

  ROSSING, Thomas.D (editor), "The Science of String Instruments" , ed: Springer NY, Stanford CCRMA, 2010 ISBN 978-1-4419-7109-8

  SMIT, Thorsten a.o., 'A highly accurate plucking mechanism for acoustical measurements of stringed instruments', in: Journal of the Acoustical Society of America, EL223, may 2010.
  

  Syndyne catalogue

   Wikipedia entry on the Qanun.