<Ror>
|
an automated barrelorgan module

Godfried-Willem
RAES
2007-2021
|
A mobile barrelorgan module using recycled pipework by Gerard Pels.
Description of the organ register:
Register names: Roerpijp 8' , (6 octaves), Salicional 4', Blockfloete 8'
The pipes sounding the notes 36 to 50 are made of zinc, from note 51 on, the
are Sn/Pb in the traditional organ pipe alloy. All these pipes are closed. The
lowest octave pipes have roll beards for intonation and tuning. Here is a picture
of some of these zinc made pipes: 
Here is a view on the pipes for notes 48,49,50 :
These pipes are 'roergedeckt'. The pipes for the 'blockflute' part of the register
(notes 84 to 93) are shaped anticonical:
Higher up, the pipes are open and ciclindric.
Since the instrument is designed for transportation, the pipes are inserted
deeper into the upperplate of the windchest than usual in traditional organ
building.
The wind to the pipes is turned on and off with vertical solenoid valves. Experiments
revealed that it was worth the effort to implement velocity sensitivity on these
valves. The circuit to achieve this is a mere repetition of what we did for
many more of our musical robots:
The circuit for a complete board, serving 14 solenoids/ notes looks like this:
The soldered board
with components: 
A view on the inside of the windchest for the notes 48 - 108: Left side:
Center part:
Left
side: 
Circuit Overview:
- Hold velo boards: see <HarmO> ,
<pos> etc...
- Midi hub board
: see bottom of page
Mapping

Midi implementation:
The midi channel for <Ror> is 3 (0-15) or 4 (1-16).
Midi note range: 36 - 108, velocity implemented (steers the speed wherewith
the valves do open and hence the note attack). Individual
note aftertouch (polyphonic) under development.
Note Off commands are required.
Controller 66 is used to switch the motor on or off.
Controller 11 controls the speed of the tremulant.e tremulant
speed is the midi value divided by 10.
Modulation depth can be controlled with controller #12.
Controller 7 is used for the wind pressure (motor speed). The normal setting
should be 72. Default startup value in the PIC firmware is 0. It cannot be used
for fast wind pressure modulation but is perfectly suitable for slow crescendo
and decrescendo. Note however that the pitch as well as the intonation may be
affected when the robot is operated on nonstandard wind-pressures.
The robot responds to the midi all-notes-off command. This command also switches
off the lights and the wind valve, but not the motor. To switch off the motor
controller 66 should be used.
Program change is implemented to select different pipe sets for the highest
register.
Technical specifications:
- size: 1200 x 500 x 1800 ( first estimate)
- weight: to be determined (? ca. 120 kg)
- power: 230V ac
- Tuning: based on A = 440 Hz at 21 degrees Celsius
- static windpressure in the windchest for normal tuning and intonation: 80
mmH2O or 8 mbar
- Sound pressure level:
- Ambitus: 36 - 108
- Insurance value: 15.000,- (first estimate)
Design and construction: dr.Godfried-Willem
Raes (2007-2021)
Collaborators on the construction of this robot:
- Mattias Parent
- Yvan Vander Sanden
- Johannes Taelman
- Kristof Lauwers
Music composed for this robot:
Nederlands:
<Ror>
De uitgangspunten bij het ontwerp, de planning en de bouw van
deze muzikale robot zijn drieerlei:
1.- Heel regelmatig bereikt ons de vraag naar muziekautomaten die kunnen worden
ingezet op openbare plaatsen, straten en pleinen. Omwille van de gebruikte materialen
en ontwerptechnische beperkingen was dit met de grote meerderheid van de robots
reeds beschikbaar bij Stichting Logos praktisch onmogelijk. Vandaar de aanvankelijke
overweging van het bouwen van een 'draaiorgel'-achtig instrument dat geschikt
zou zijn voor bedrijf in openlucht: op straat dus. Anders gesteld, voor het
ontwerp vereist dit een zekere mate van regenbestendigheid, een eigenschap die
in geen van de muzikale robotten die we bij Stichting Logos ontwikkelden in
voldoende mate aanwezig is. Deze vereiste dikteert het gebruik van kunststoffen
en metalen eerder dan de traditionele materialen uit de orgelbouw, met name
hout en leder. Ook voor de elektronika vergt dit uitgangspunt enkele bijzondere
maatregelen. Voor de pijpen kunnen we uitgaan van een register gebouwd door
Gerard Pels (1955-2014) voor een klein kistorgel en waarvoor zink werd gebruikt.
Pels heeft het orgeltje nooit helemaal afgewerkt, en we konden dus vertrekken
van dit halffabrikaat. Er was een houten windlade, voorzien van elektrische
ventielen, voor de pijpen vanaf noot 49. Helemaal geen windlade echter voor
het grondoktaaf (noten 36-48). Voor die windlade zelf, kwam dik massief PVC
in aanmerking, zoals we dat met groot sukses al hadden toegepast in onze Hybr
reeks: <Hybr>, <HybrHi> en <HybrLo>. Maar ook zonder echte
windlade, mits gebruikmaking van losse 1/2" magneetventielen en heel wat
slangen moest het mogelijk zijn. De draagplank kon dan ook in traditioneel hout
worden gemaakt.
2.- Het bouwen van een op zichzelf staande module die bovendien ook interaktief
kan werken in funktie van publiek, bespelers en/of omgeving. Daarmee is bedoeld,
dat de automaat zelfstandig moet kunnen werken en dus geen externe apparatuur
zoals interfaces, een laptop, sensoren... nodig mag hebben. Om die interaktiviteit
mogelijk te maken voorzien we minstens twee Doppler-radar sensoren, waardoor
interaktiviteit via expressief relevante beweging mogelijk wordt gemaakt. Anderzijds
vergt deze vereiste de aanwezigheid van heel wat geheugen in de processoren
die voor de robot worden gebruikt. Alle repertoire moet immers in die chips
kunnen worden opgeslagen.
3.- Los van vorige uitgangspunten, wilden we met deze module een
proefprojekt opzetten ter evaluatie van het koncept van zwaartekracht-kleppen.
Traditionele orgelautomaten maken gebruik van een windlade waarop de pijpen
worden geplaatst. Binnenin die windlade bevinden zich elektrisch bestuurde verntielen
die via een veer in rust gesloten worden gehouden. De veer is nodig, omdat de
ventielen 'ondersteboven' dienen te werken. Principieel zijn tegen dit traditioneel
ontwerp nogat wat bezwaren aan te voeren: de veer introduceert inherent een
risico op resonanties bij bepaalde bekrachtigingsfrekwenties. Het lekvrij monteren
van de ventielen vereist een engelengeduld en lekken ontstaan ook na montage
vrij makkelijk na verplaatsingen of transporten van de windlade. Konische ventielen
zijn wat dit betreft nog vele malen lastiger dan de meestal gebruikte vlakke
exemplaren. (cfr. bouwdagboek voor onze <Bomi>
robot en het paper dat
we in dat verband publiceerden) Het herstellen van een lek is bijzonder
tijdrovend, omdat alle pijpen dienen te worden verwijderd (... en achteraf herstemd...)
teneinde de bovenplaat van de windlade te kunnen bewerken. Vandaar ons idee
om de ventielen bovenop de windlade te monteren, zo dat ze in rust door de zwaartekracht
gesloten worden gehouden. Aangezien de pijpen -op grond van hun konstruktie-
niet goed ondersteboven kunnen worden gemonteerd, wilden we bij dit ontwerp
uitgaan van de toepassing van 'omkeer-cancellen', een soort U-konstrukties tussen
orgelpijp en ventiel. Gebruik van elastische slangen is natuurlijk ook mogelijk,
al moet bij langere trajekten rekening worden gehouden met de vertragingstijden
en drukverliezen.
Mensuur tabel voor de pijpen:
noot = midi note
length = measured from flue to pipe end. Between brackets: extra
length roerpijp (Rohrpfeife)
Noot |
diameter (in mm) |
length |
foot |
orifice |
valve seat |
material |
36 |
110 - 113 |
1110 |
250 |
1/2" |
10 mm |
Zn |
37 |
107 - 108 |
|
|
1/2" |
10 mm |
Zn |
38 |
104 - 105 |
|
|
1/2" |
10 mm |
Zn |
39 |
100 - 100 |
|
|
1/2" |
10 mm |
Zn |
40 |
97 - 96 |
|
|
1/2" |
10 mm |
Zn |
41 |
94 - 94 |
|
|
1/2" |
10 mm |
Zn |
42 |
91 - 92 |
|
|
1/2" |
10 mm |
Zn |
43 |
87 - 90 |
|
|
1/2" |
10 mm |
Zn |
44 |
85 - 87 |
|
|
1/2" |
10 mm |
Zn |
45 |
81 - 84 |
|
|
1/2" |
10 mm |
Zn |
46 |
79 - 80 |
|
|
1/2" |
10 mm |
Zn |
47 |
77 - 78 |
563 |
200 |
1/2" |
10 mm |
Zn |
48 |
79 |
574 [125] |
200 |
|
|
Zn |
49 |
76 |
|
|
|
|
Zn |
50 |
71 |
|
|
|
|
Zn |
51 |
66 |
|
|
|
|
Sn-Pb |
52 |
|
|
|
|
|
Sn-Pb |
53 |
|
|
|
|
|
Sn-Pb |
54 |
|
|
|
|
|
|
55 |
|
|
|
|
|
|
56 |
|
|
|
|
|
|
57 |
|
|
|
|
|
|
58 |
|
|
|
|
|
|
59 |
|
|
|
|
|
|
60 |
|
|
|
|
|
|
61 |
|
|
|
|
|
|
62 |
|
|
|
|
|
|
63 |
|
|
|
|
|
|
64 |
|
|
|
|
|
|
65 |
|
|
|
|
|
|
66 |
|
|
|
|
|
|
67 |
|
|
|
|
|
|
68 |
|
|
|
|
|
|
69 |
|
|
|
|
|
|
70 |
|
|
|
|
|
|
71 |
|
|
|
|
|
|
72 |
|
|
|
|
|
|
73 |
|
|
|
|
|
|
74 |
|
|
|
|
|
|
75 |
|
|
|
|
|
|
76 |
|
|
|
|
|
|
77 |
|
|
|
|
|
|
78 |
|
|
|
|
|
|
79 |
|
|
|
|
|
|
80 |
|
|
|
|
|
|
81 |
|
|
|
|
|
|
82 |
|
|
|
|
|
|
83 |
|
|
|
|
|
|
84 |
|
|
|
|
|
|
Voor de windvoorziening maakten we gebruik van een kleine Ventus
orgelblazer van de firma Laukhuff, met een regelbare winddruk van maximaal 80
mm waterkolom, of 8 mBar = 785 Pa, in eenheden uit de fysika. De aansturing
van de 80 Watt motor gebeurt met een motorcontroller. Zoals voorspelbaar en
volkomen normaal bij orgelpijpen, is ook hier de stemming enigszins afhankelijk
van de winddruk. Alleen bij een motor AC frekwentie van 50Hz is de stemming
korrekt. Winddruk 80mm H2O. Om een eenvoudige afregeling,
stemming en intonering mogelijk te maken, monteerden we een precieze manometer
aan de buitenkant van de windlade. Het maximale debiet van de kompressor is
3 kubieke meter, wat dus brede klusters ruimschoots mogelijk maakt.
De radiale kompressor is voorzien van een geluidsdemper aan de
aanzuigkant evenals van een regelklep op de inlaat.
Tessituur:

Building logbook / Bouwdagboek:
Omdat ons vaak wordt gevraagd hoeveel werk en tijd kruipt in,
en nodig is voor, het bouwen van onze muzikale robotten, houden we ook voor
<Ror> een beknopt en geilllustreerd bouwdagboek bij:
- 01.08.2007: Het onafgewerkte kistorgel komt via Yvan Vander Sanden in het
Logos atelier terecht. Twaalf losse baspijpen, een windlade, een doos losse
ventielen...
08.10.2017: Eerste ontwerpideen en bouwvereisten opgesteld.
- 09.10.2017: First estimate of building costs, including labor, calculated:
ca. 64.700 Euro.
- 10.10.2017: Sketches for the gravity operated valves worked out. Test conical
valve with seat from massive PVC fabricated.
- 11.10.2017: The conducts between valves and pipes will introduce a certain
latency.
wind pressure |
wind speed |
50 mm H2O |
29 m/s |
80 mm H2O |
38 m/s |
100 mm H2O |
42 m/s |
120 mm H2O |
47 m/s |
230 mm H2O |
60 m/s |
These data derived from the A.Laukhuff Catalog p.6.44. If we want the latency
to stay below 1 cs, the lengths of the conducts should be limited to 290 mm
(50 mm H2O).
- Calculated results for the conical valves in Bomi:
-
cone diameter |
top angle |
traject |
diameter of equivalent orifice |
35mm / 15mm |
110° |
5.2mm |
10 mm |
25mm / 12mm |
100° |
5.0mm |
7 mm |
20mm / 11mm |
85° |
6.0mm |
5 mm |
16.5mm /10.2mm |
81° |
6.0mm |
4.3 mm |
13mm/ 8.7mm |
72° |
6.0mm |
3 mm |
- 20.11.2018: Solenoid valves ordered from Banggood, China, came in, as we
were curious about the quality. We took them to the lab and found that they
start opening with a voltage of 5.6 Volt. At the nominal 12 V they draw 480
mA of current. On decreasing voltage they stayed opened down to 4.3 Volts.
Between this voltage and ca. 7V they can be operated as flow regulators. Operation
is reasonably silent, certainly if we compare them to the M&M valves we
used for <Vox Humanola> and <Piperola>. Another difference is
the orifice size: 1/2", way more then what we had on the earlier mentioned
organ robots. Using these valves for the lowest two octaves seems possible
and also would give us a lot of design freedom, as we are no longer forced
to place them on a regular windchest.
- 29.11.2018: ABB drive (type ACS355, ordered from Farnell) tested with a
small Laukhuff ventola blower... Failure, as we cannot bring the output voltage
down to 132 V... It overheats the motor. We need to figure out a star connection...

- 30.11.2018: Siemens Sinamics G110 drive, 0.12kW ordered from Conrad (ord.
nr 19879), with a BOP programming panel. Cost 201.49 Euro.
- 01.12.2018: Further tests conducted with the ABB motor controller and the
mini Ventola blower, rewired in star configuration. This now works fine, however
wind pressure will be too low to drive this robot. Found an interesting report
on organ wind supply by a dutch electro-engineer.
Here is a link. BOP-panel for the ABB drive ordered from Farnell.
- 05.12.2018: Design of a double sided PCB for the midi-hub board. Basically
it's the same board as designed for <HybrLo> and <Balsi>, but
double sided such as to avoid wire bridges and guarantee a better screening.
Here is the PCB design, component side and copper side:
After assembling, we wonder whether using a double sided PCB is worth all
the effort. Particularly because we cannot make metalized holes. After all,
placing a few wire bridges on a single sided PCB is much less work then soldering
all the via's on a double sided board...
- 12.12.2018: Here is a picture of the finished double sided hub board:
The single sided version was used for our <Balsi> robot.
- 24.05.2019: Further experiments performed with the Chinese PVC solenoid
valves (1/2"). We could either use 1/2" flanges for mounting, or
cut 1/2" threads in the windchest material, if we use PVC for it. Sketches
for the pipe layout drawn.
- 26.05.2019: Continued analysis of the existing pipework. Looks like the
minimum width of the organ should become 1200 mm.

- 28.05.2019: Compressor ordered from August Laukhuff: Ventus 600380 (3 m3/min,
80mm H2O 0r 800Pa). Motor power: 150W. Weight: 15 kg.
This is the same type as what we used for our <Piperola> robot.
- 29.05.2019: Inventary of the available pipes (99 in total, covering 72 notes!):
The highest
pipes for sure, will need some restauration work.
- 02.06.2019: We got a message from Laukhuff saying that the compressor we
ordered can only be delivered by the end of September 2019...
- 01.09.2019: Two organ blowers found on the flea market this morning: a Swiss
made Meidinger MF0601 type NTKX21 as well as an EMI K640, made in Utrecht,
Holland. Both brands do not exist anymore...
Both blowers can be
made to work again...
- 30.09.2019: The ordered Lauhuff blower came in. So we now can choose between
three blowers...
- 25.01.2021: After a year plagued by both health (cancer) and financial problems
(Logos subsidy) we reconsidered the project. Anticipating that we probably
do not have a long life in front of us anymore, we may very well simplify
this project such that finishing it and bringing it to a usefull state for
musicians becomes realistic. This would entail that we would keep as much
as possible from the original construction, including the original windchest.
As this is made in wood, we need to foresake the original idea of making it
a weatherproof instrument. Minimum sizes of the instrument would become: 1200
mm x 450 mm x 1400 mm. The original windchest measures 1100 x 270 mm, but
does not include the twelve lowest octave pipes. Here are some pictures of
the internals of the windchest:
The wiring is done
with wires of a gauge smaller than what we would normally use, but as its
in a perfect shape we decided not to replace it. The wiring style, using these
typical curled wires, goes back on a tradition that started at the end of
the 19th century!
- 26.01.2021: Two 270 mm square stainless steel tubes cut to hold the sides
of the windchest. Solenoid valves 1/2" selected for use on the bass octave.
Accordion bellows digged up for potential use on <Ror>. Can be used
to stabilize windpressure and to make the tremulant. 1/2" nuts ordered,
for mounting the valves on a windchest or holding plate.
- 27.01.2021: Accordions work on windpressures between 5 mm H2O
and 100 mm H20 ( 50 Pa to 1000 Pa), so we are on the
safe side by using such bellows, taking into account that the static pressure
produced by the Laukhuff blower we anticipate to use here, is 80 mm H20
(800 Pa). Bellow and windpressure regulation construction started. Sizes will
be 422 x 224, h=235, in mm. Considering the possibility to mount a large solenoid
inside the bellows to be used as a tremulant. Here is a drawing with the sizing
used:
These bellows
connect to the Ventus blower with four M5 x 40 bolts. The heads of these bolts
are glued inside the bellow construction. We cannot finish this construction
before we have the angled flanges to be used for feeding both windchests available.
- 28.01.2021: 1/2" nuts digged up... we had 100 pieces in stock: Viega,
made in Germany nr. 102 241 - 1911 1/2, in bags with 10 pieces each. Windvalve
made for the inlet of the bellow construction. Two outlets welded using 90
degree 50 mm diameter knees and a flange. Bottom plate finished, but we still
have to make a gasket to make it airtight. We can use a piece of silicon membrane
here.
- 29.01.2021: Some pictures of the newly made bellows:
. At the same time we also
made some documentation pictures of the Chinese Banggood valves to be used
on the low windchest:
And this is what the valves look like after placing the solenoids upside down:
The left valve is the original, the right one the reversed one.
As can be read from the label, these valves can be used for pressures up to
10 kPa, equivalent to 1000 mm H2O. The cold resistance
of the coil is 24.8 Ohm, so at 12 V the current will be 500 mA. Welding work
on the windchest for the lowest pipes: length 1200 mm, width 130 mm. Test
mount of four valves with 1/2" nuts. Drilling of the holes: 20 mm size,
after drilling, feed a 1/2" BSW tap through the hole.
- 30.01.2021: Milling of a first series of holders for the pipe feet in the
lowest register:
With the
pipes standing on these holders and mounted on the windchest it looks now
like this:
As we run out
of 1/2" couplers in brass, we ordered a bunch of new ones:
These
are a bit smaller in diameter, but still large enough to do the milling and
match them to the somewhat higher pipes. Possible designs for the wheelbase
sketched. A four-wheeler with two large wheels in the middle seems to be the
most compact construction. However, we do not have the tyres for 600 mm diameter
wheels in stock... We do have a couple of very sturdy 150 mm wheels, mountable
in a pivoting wheel holder. This would be the same type as used for <Bello>.
- 31.01.2021: Further design work. We are extremely limited by not having
any funds and thus by having to work with whatever we still have available
and in stock. We digged up a bunch of PCB's designed for our player piano
in 2006. These can be used for <Ror> in any case.
- 01.02.2021: Scary to discover that our favorite Microchip PIC's 18F4620
aren't any longer available at Farnell and neither at RS-Components... ASAP
we ordered everything (16 pieces) they still had in stock at Conrad.
- 02.02.2021: Start soldering of the required pulse-hold boards. This is the
old recipy, checked once more on the oscilloscope:
P-channel
mosfets used here are VP0109, as the BS250 types we used in older robots are
no longer on the market. The VP0109's are specified for -90V, 500mA, RsON<
8 Ohm. They are made by Microchip. Here
is the datasheet. For the velo-pulse drivers we used IRF540, as these
withstand Ug +/-20V easily. The specs. for the IRL640 are a bit ambiguous
in this respect: data sheets give max.values between +/-10V and +/-20V for
this type.
- 03.02.2021: Three pulse-hold PCB's soldered. Leadfree for a change. Looks
like we are running out of IRL640 Mosfets, VP0109 P-channel mosfets, diodes,
red LED's, 4k7 resistors... A fresh load of Microchip 18F4620's came flowing
in from Conrad. None in stock anymore there now. Further milling of the 1/2"
brass pipeholders. Since we have note overlaps in the pipes, we may just as
well implement program change to select pipe sets here.
- Board mapping:
- Board 1: 36 - 47 (12 pipes)
- Board 2: 48 - 61 (14 pipes)
- Board 3: 62- 75 (14 pipes)
- Board 4:76 -89 (14 pipes)
- Board 5: 90 - 93 ( 4 pipes)
- Board 6: 76 - 89 ( 14 pipes) - reg 2.
- Board 7: 90 - 103 (14 pipes) - reg 2.
- Board 8: 104 - 115 (12 pipes) - reg 2.
- 04.02.2021:Solving some design problems... Four wheels or six wheels, that's
the question...
- 05.02.2021: TIG-welding of the base chassis: 30 x 50 x 2 profiles, 1100
mm long, welded together in a rectangle as a holder for the compressor and
the bellows. Vibration dampers with M6 threads digged up, so we can use them
for the mounting of the compressor.
- 06-07.02.2021: Some quite extensive soldering work at hand...:
And
of course, we are running out of components.
- 08.02.2021: Sizing recalculated and redrawn:
Vertical drawing to be measured and calculated such as to stay with a height
smaller than 2 meters.
- 09-11.02.2021: More soldering work on the boards.
- 12.02.2021: Measurement of the characteristics of the solenoid valves inside
the existing windchest: DC resistance 130 to 150 Ohms, so there is quite some
variation here... Nominal current at 12V is 85 mA. When we bring the voltage
down to 5 V, the valves closes. It re-opens at 9 V. So the power supply requirements
can be calculated at 12V - 10A, for the hold voltage, and 6 to 10 V - 6 A
for the pulse voltage. Some flow regulation is possible with voltages between
5 and 9 V.
- 13.02.2021: Construction of a power supply. As this PCB is very simple,
we hand-milled it on a suitable piece of PCB material. Here is the circuit:
Using this circuit
in combination with the pulse-hold boards, will set the hold voltage over
the solenoids to 10.8 Volts.
- 14.02.2021: Soldering the midi-hub board. This board seems suitable to also
carry the motor control functions. Circuit redrawn:
Some more welding on the base structure for holding the radial compressor.
For mounting the motor we use 2 pieces of 30 x 30 x 2, length 200 mm, stainless
steel. Shock absorbers (M6 threads) are used to mount the motor. There will
be place for the Siemens motor controller adjacent to the motor., thus minimizing
EMC.
- 17.02.2021: Power supply PCB finalized and tested.
- 18.02.2021: Construction of a mount for the Siemens motor controller.

- 19.02.2021: Further welding work on the base structure:
Construction of the mains power inlet and the bipolar on/off switch:
Tentative mount of the wheelbase using wheels recycled from a wheelchair:
Overview power circuitry:
- 20.02.2021: Further work on the bellows and the mounting components for
the dual 12V power supply and the hub board.
To be done:
Robodies Pictures with <Ror>:
to be done as soon as <Ror> gets ready, or, likely never, if we die before
<Ror> gets finished.
Last update: 2021-02-19
by Godfried-Willem Raes
Technical data sheet and maintenance instructions:
- Motor specs: Ventus 6 003 80, 150 W, 80 mm H2O
- Motor Controller: Siemens Sinamics
- PIC controllers: 18F4620 and 18F2620
- Power supply:
- Hold voltage: 12 V/ 10 A supply.
- Velo pulse voltage: -12 V SMPS module, 10A
- Logic 5 V supply: from separate SMPS on the midihub board.
- Front light:
Wiring for the Laukhuff blower:
Valve solenoids:
Banggood, 1/2" solenoid valves used for the lowest pipes.
Pressure sensor inside the bellows: MPXV7002, pressure range -2000 to +2000
Pa (= -200 to +200 mm H2O)
Powersupply modules: 4 pieces Meanwell IRM-60-12. Input: 100-240Vac, 1.8A ;
output 12V 5A. Made in Taiwan. Manual: http://www.meanwell.com/manual.html Here
is the relevant datasheet.
Literature and documentation:
Meidinger catalogue
RAES, Godfried-Willem "
Logos @ 50, het kloppend hart van de avant-gardemuziek in Vlaanderen",
Ed. Stichting Kunstboek, Oostkamp, 2018