e015044.PDF
(
189 KB
)
Pobierz
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
GENERAL
INTEREST
DMX512 Revealed
we tell all…
By B. Bouchez
Since the very beginning,
Elektor Electronics
has tried to help its readers profit
from new technologies by means of original and professional circuits. In that
tradition, we are presenting a series of circuits (DMX to voltage and MIDI to
DMX) that are specifically intended to be used to control lighting systems for
theatres, discos and outdoor events — all based on the DMX512 system.
The DMX512 system, which is already well
known to those who are professionally
involved in theatre lighting, is beginning to
find widespread use. The objective of this
article is to take a look behind the scenes
with regard to the various technical aspects
of this connection protocol, after first briefly
reviewing the history of remote control tech-
niques for lighting systems.
clearly seen that even then, archi-
tects took the ‘technical’ aspects into
consideration. However, it is not pos-
sible to determine with certainty
whether the emplacements for the
ancient light sources served only
practical ends (to allow perfor-
mances to be given after dark) or
Once upon a time…
Starting with the appearance of the
first theatres in ancient times, light-
ing has played an important role. In
Roman and Greek ruins, it can be
44
Elektor Electronics
5/2001
GENERAL
INTEREST
were used to create effects on the
stage. We should not forget, though,
that the Greeks and Romans had
considerable experience with special
effects, which would tend to support
the second supposition.
Once theatre performances started
to take place increasingly often in
enclosed spaces, the need for light-
ing became clearly evident, if only to
fix the attention of the audience on
the stage.
The first remotely controllable light-
ing sources appeared in the middle
of the 19
th
century. The lamps of that
era used natural gas. The term ‘light
organ’ originates from that time and
comes from the appearance of the
control panel, which with its multi-
tude of pipes and valves resembled
the front of a pipe organ.
Gas lighting was a far from ideal
solution, and it was replaced by
electric lighting as soon as this was
made possible by the technology of
incandescent lamps and carbon-arc
lamps.
The first electrical control panels
consisted simply of a set of auto-
transformers connected to a number
of incandescent lamps (it is not pos-
sible to control carbon-arc lamps
electrically). These installations were
large, heavy and rather dangerous,
since the majority of the ‘live’ ele-
ments were exposed to touch, as
was common with control panels at
that time.
This system, which came into being
in the 1920s, was used exclusively
until the 1960s, when the first thyris-
tors appeared.
Thanks to these semiconductor
devices, the weights and dimen-
sions of control panels could be sig-
nificantly reduced. An additional fac-
tor was that all that was needed to
remotely control the thyristors was a
simple voltage, so it was possible to
devise intricate switching circuits
that could be used to create effects
that had previously been impossible,
such as the simultaneous chained
dimming of a multitude of lights (just
imagine trying to operate several
mechanical regulators in unison!).
Before continuing with the history of
remote control, let’s take a closer
look at these thyristor controllers,
which technicians call ‘dimmers’ or
‘power boxes’. They dominated the
scene from the 1960s until the mid-
t
BRK
t
CHR
t
CHR
t
CHR
t
BRK
start code
channel 1
channel N (N
≤
512)
t
MAB
t
IC
t
IF
t
BRK
= break time = 88
µ
s (min.) to 1 s (max.)
t
MAB
= mark time = 4
µ
s (DMX512 - 1990)
t
IC
= inter-character time = 0
µ
s (min.) to 1 s (max.)
t
IF
= inter-frame time = 0
µ
s (min.) to 1 s (max.)
t
CHR
= character time (1 start bit + 8 data bits + 2 stop bits = 11 bits)
(nominal bit time = 4
s (DMX512 - 1986) or 8
s / 3.92
s min. / 4.08
s max.) i.e., 44
s typical
010035 - 11
Figure 1. Structure of a data block.
1980s. Thryistors work on the princi-
ple of phase control, and they create
audible disturbances (the lamps
‘sing’) as well as a large amount of
electromagnetic interference, which
can create disturbances in sound
systems. In order to curb these dis-
turbances, thyristor dimmers are fit-
ted with chokes, which increase the
weight of the circuitry and reduce
the speed with which the lights can
be brought to full power.
Since the middle of the 1980s, tran-
sistors that can work at mains volt-
ages have been available. This made
it possible to produce high-frequency
dimmers. The chokes could thus be
made considerably smaller and thus
lighter (if they were not simply elim-
inated). An additional benefit is that
there is almost no interference with
the sound system (although the
same cannot be said with regard to
electromagnetic interference!).
ted with these semiconductor controllers
could be operated remotely using voltages or
currents generated by a control panel made
up of a number of potentiometers.
This type of remote control gave rise to a
number of concepts that are still used, such
as ‘scene presets’ (a set of lighting com-
mands sent to the stage as a group), ‘fades’
(progressive transitions from one preset to a
different preset), ‘master’ (the simultaneous
control of the levels of a group of presets
using a single potentiometer), ‘blackout’ (fad-
ing a complete preset to the null level) and so
on.
It should be noted that even today, the basic
ideas of stage lighting are still based on the
concepts that were established at the time of
the first remote control panels for lighting
systems. In the catalogues of manufacturers
of stage equipment, you will still find fully
analogue control panels at the bottom end of
the price range.
One of the principal shortcomings of ana-
logue remote control using a voltage (or cur-
rent) is the need for a separate conductor for
each control channel. With a portable instal-
lation, such as is used for tours, re-attaching
all the connectors for a large number of lights
can turn into a real nightmare. In addition,
repeated plugging and unplugging of con-
nectors does not improve their reliability.
Beginning in the early 1990s, many developers
of lighting systems turned their attention to
From remote control to
multiplexing
In addition to the other advantages
that came with the use of power
semiconductors, it was possible to
operate the electronic controllers
remotely. Installations that were fit-
Table 1. Earlier communications protocols
Protocol
Developed by
Remarks
ADB62.5
ADB (Belgium)
Comparable to DMX512, but works at 62.5 kb/s.
No longer used in new commercial products.
AVAB
AVAB (Sweden)
Widely used in Northern Europe.
CMX
Colortran (USA)
Was the basis for DMX512. Still frequently used in the USA.
Micro2
LMI (USA)
Secure protocol, control by means of returned data.
PMX
Clay-Paky (Italy)
Slow protocol (9600 baud); compatible with the PC serial port.
Still used in Clay-Paky and Pulsar products.
5/2001
Elektor Electronics
45
GENERAL
INTEREST
this problem. Several different solutions were
proposed to reduce the number of connec-
tions between the dimmer racks and the con-
trol panel. One manufacturer, Strand Lighting,
developed a protocol based on multiplexing
and gave it the name ‘CD80’. With this sys-
tem, several tens of commands could be
transferred via a standard microphone cable.
Before this protocol could be fully developed,
a standard that was largely inspired by it was
developed by the USITT (United States Insti-
tute for Theater Technology). This standard
defines an analogue multiplex protocol, called
AMX912, that is based on two pairs of leads.
One of these carries a variable voltage ranging
between 0 V and 5 V (which represents val-
ues from 0% to 100%), while the other one car-
ries a digital clock signal (0 V = ‘0’, > 4 V =
‘1’).
The synchronisation of the dimmers with the
transmitter (the control panel) is realised by
sending one high-level clock signal with a
longer duration than the rest. Each receiver
copies the value of the analogue signal while
the clock signal is high, usually be means of
a sample and hold circuit. The high level of
the clock signal must have a duration of at
least 50 µs. AMX (which stands for ‘Analogue
MultipleX) owes its name to the fact that the
maximum number of channels that can be
controlled before the receivers must again be
synchronised is 192.
The USITT recommended the use of a bal-
anced transmission line for the clock signal,
in order to allow large distances to be
bridged. This is a bit strange, since this rec-
ommendation did not apply to the analogue
signal pair, with the result that AMX192 is
sensitive to ground loops and noise. This
standard was only used privately by a few
manufacturers and was never truly success-
ful. On top of this, the analogue protocols of
other manufacturers profited from their com-
mercial strengths and spread more quickly.
Some of these protocols, such as S20 (ADB)
and D54 (Strand Lighting), are still in use.
To next segment
(max. 1,000 m / 32 devices)
Up to 32 devices connected
Repeater
Note: repeater equals one load
Master Console
010035 - 12
Figure 2. The electrical limitations of RS-485.
Sender
(console)
Receiver
(slave)
data line +
data line -
010035 - 13
Figure 3. Principle diagram of an RS-485 connection.
standard version
(5 pins)
simplified version
(3 pins)
3
3
4
2
input,
mal
e
Pin
Function
5
1
5
1
1
2
3
4
5
0V/Ground
Data -
Data +
free (optional data -)
free (optional data + )
3
3
2
4
input,
female
1
5
1
5
010035 - 14
Figure 4. XLR connector pinout.
Digital is coming!
And then came DMX512
sively via the serial link. The value
00h is equivalent to 0% and the value
FFh is equivalent to 100%. You must
take into account that in the DMX
standard, there is no fixed relation-
ship between the received DMX
value and light intensity (or move-
ments, such as the orientation of a
mirror or the rotational speed of a
gobo).
After all the values have been sent
once, the cycle starts again from the
beginning. It must be noted that the
reliability of DMX is simply based on
In the middle of the 1980s, several manufac-
turers of lighting equipment, faced with the
demands of professionals and being con-
vinced of the future prospects of the micro-
controller, decided to integrate this new
device in their equipment. As a result, a num-
ber of communication protocols appeared,
each with its own advantages and disadvan-
tages. You should also realise that these pro-
tocols were completely incompatible with
each other. Nevertheless, some of them are
still used at the present time.
Table 1
pre-
sents a brief summary of these protocols.
At the beginning of 1986, the USITT
and certain manufacturers held a
meeting in order to specify a fully
open digital protocol.
In order to enable existing microcon-
trollers to be used, they settled on an
asynchronous serial connection with
8 data bits, 2 stop bits and no parity,
with a data rate of 250-kbaud.
The control principle is simple: bytes
that contain the values for the asso-
ciated channels are sent succes-
46
Elektor Electronics
5/2001
GENERAL
INTEREST
the continuous repetition of this
cycle, with essentially no check
being made to verify the correct
reception of the values (this is called
a ‘simplex’ link). Due to the absence
of a return signal, DMX is not suit-
able for controlling pyrotechnic
devices or heavy mechanisms (such
as hydraulic lift platforms), for rea-
sons of safety.
In the DMX standard, there is no
value specified for the minimum
number of channels that can be con-
trolled. This number can lie between
1 and 512 (hence the name DMX512).
Clearly, the fewer channels there are
to be controlled, the shorter is the
cycle time (and the faster the reac-
tion time of the controlled equip-
ment). A number of ‘master’ DMX
devices, such as the interface that
will be described later on, also allow
the number of controlled channels to
be changed for each new cycle.
The first byte of each series of bytes
sent via the DMX line is not respon-
sible for controlling the intensity of a
lamp. This byte is called the ‘Start
Code’, and it is intended to be used
for possible extensions. Presently,
the only valid standard value for this
byte is 00h. Normally speaking,
every ‘standard’ DMX device is sup-
posed to ignore a frame in which the
Start Code is not equal to 0. How-
ever, some manufacturers assign a
non-standard value to the Start Code
in order to cause their equipment to
execute special programs (tests,
resets, demos etc.). Take our advice:
if you want to avoid panic situations
during a performance, refrain from
‘playing around’ with the Start Code,
especially when using equipment
from several different manufacturers
(we speak from experience!).
Now you know how data for the var-
ious channels is sent out. The prob-
lems that the receiver is faced with
are how to determine where the
command series starts (and ends),
how to recognise the Start Code and
how to count the number of bytes
that have been sent in order to iden-
tify its associated channel. Synchro-
nisation is achieved by forcing the
serial link into a state that it can
never attain in actual operation,
regardless of what values are sent.
This special state is called Break,
and it is achieved by holding the link
at the ‘0’ level for the duration of at
least two frames (88 µs). Given the
fact that at least two stop bits are
located between every pair of bytes
that are sent, it is impossible for
such a situation to be reproduced
during a normal transmission, even
if two successive frames have the
value 00h (eight bits with the value
of ‘0’).
When the receivers in the communi-
cations link recognise a ‘gap’ gener-
ated by Break, they are synchronised
and thus know that the following
byte is the first byte of a new data
block. According to the DMX stan-
dard, the Break time can have any
value between 88 µs and 1 s. In prin-
ciple, after the expiry of a one-sec-
ond interval the receivers can
assume that the link has been bro-
ken, and they must then switch over
to a predefined state (which is usu-
ally ‘everything off’).
The problem that we now face with
regard to Break relates to the signal
that announces the start of a new
byte to the receiver, and which is
called the start bit. This bit has a
value of ‘0’, since the value of the
line when it is at rest (the ‘idle
state’) is ‘1’. In order to allow the
start bit to be properly recognised,
the serial link must remain ‘high’ for
a certain time following the Break. It
is the job of ‘Mark after Break’ (MaB),
which comes after the Break time, to
ensure that this happens.
In the first version of the standard,
published in 1986, the MaB time was
specified to lie between 4 and 8 µs,
but experience proved that this time
was too short for the electronic cir-
cuitry of that time. In 1990, a revised
version of the standard appeared in
which this interval was increased to
a value of at least 4 µs and at most
1 s. Equipment that supports this
version (which in practice is far from
being the most commonly used type)
is declared to be compliant with
‘DMX512 (1990)’.
If we combine all of the elements
described above, we come up with
the data-block timing diagram
shown in
Figure 1
.
010035 - 15
Figure 5. A homemade DMX terminator.
of the road (in technical terms, this is called
the ‘physical layer’). If we enumerate the fea-
tures of the protocol, we see that we have a
connection that supports communications
over the longest possible distance, a capac-
ity of 250 kb/s and the need to simultaneously
control a certain number of different devices.
It is clear that our old friend, the RS-232 stan-
dard – with a maximum distance of 15 m
between two points, a theoretical maximum
speed of 19,200 baud and the electrical char-
acteristics of a point-to-point link – is totally
unsuitable for this task.
The DMX512 standard employs a different
physical layer, which has been standardised
under the name RS-485. This type of connec-
tion is derived directly from the RS-422 stan-
dard (which is primarily used for the serial
ports of Apple computers), but it allows mul-
tiple receivers to be connected in parallel to
a single transmitter (which is in fact the dif-
ference with RS-422).
The possibilities that RS-485 offers in com-
parison with RS-232 are extraordinary. Judge
for yourself: the maximum speed is 10 Mbaud
(that’s right, 10,000 kbaud) over 100 m, or
1 Mbaud over 1000 m. The 250 kbaud of the
DMX512 standard thus represents nothing to
fear for components that comply with the RS-
485 standard. In addition, the electrical char-
acteristics are such that 32 receivers can be
attached to the same cable. This is referred
to as 32 unit loads (ULs).
Let’s take a closer look at the UL concept.
DMX installations quite often experience
problems for the simple reason that the num-
ber of ULs connected to the transmitter is
more than the upper limit of 32. This occurs
most commonly with equipment in the lower
price range in which optocouplers are directly
connected to the DMX line, so that the
receiver represents more than one unit load.
We know that the RS-485 transmitters are lin-
ear analogue amplifiers that, quite properly,
have current limiting circuitry. If the load
exceeds the expected maximum value, the
current limiting goes into action, which trans-
lates into the output stage of the chip becom-
The transmission medium
Now that we have looked at the
‘cars’ (data blocks) that travel over
the electronic DMX ‘road’, let’s have
a closer look at their characteristics
5/2001
Elektor Electronics
47
GENERAL
INTEREST
ing warmer. The majority of modern ICs have
thermal protection, which completely cuts off
the output stage if the chip temperature rises
excessively. In the case of a DMX connection,
this causes the communications to be inter-
rupted for several minutes (until the chip has
cooled down and communications are
resumed). This type of behaviour generally
provokes nervous breakdowns among users.
Now you know what to expect if you ignore
the limit of 32 connected devices.
What if you want to use more than 32 devices
on the same line? Not to worry, you can buy
repeaters from the makers of lighting equip-
ment. A repeater counts as one unit load, but
it can in turn be loaded with an additional 32
unit loads. (Nowadays, there are even chips
that can handle 256 unit loads.)
You may wonder how RS-485 is able to trans-
mit data over such a long distance. The
answer is that it employs the differential
mode, which means that for the receiver the
important factor is the difference between the
voltages on the two lead of the pair, rather
than the difference between a voltage and a
reference level as in RS-232 (see
Figure 3
.)
One benefit of this approach is that it frees us
from interference problems (the voltages on
both conductors are displaced with respect
to the reference level, but the difference
between the voltages remains the same, so
the receiver is not disturbed). A second ben-
efit is that since no current flows in
the ground conductor (the return
current flows via the second lead of
the pair), there is no common-mode
voltage as with RS-232.
The previous paragraph is actually
just an introduction to a discussion
of the cable that must be used for
the DMX link. The connector that is
recommended by the standard is a 5-
way XLR type, which is called a
‘microphone connector’ in the stan-
dard. This short phrase has seduced
countless technicians into employing
audio cable with a single screened
twisted pair of conductors. This is
just plain wrong! A special data
cable that is suitable for RS-485 must
be used (consult the catalogues of
cable manufacturers), especially if
the total length of the DMX connec-
tion is more than a few tens of
metres.
It is essential that cable that is used
contains a twisted pair of leads and
is screened. The data leads must be
connected to the proper pins of the
XLR connector, as shown in
Figure
4
. The screen, which is usually
grounded, is connected to the shell
of the connector. The data leads
must not make contact with the
screen under any circumstances.
Otherwise, in the best case the
result will be an improperly func-
tioning installation, and in the worst
case the output stages of the RS-485
transmitters will be destroyed.
You may have noticed that there are
two types of XLR connectors: a
three-way type and a 5-way type.
The DMX standard actually refers
exclusively to the 5-way type and
notes that outputs (mixer panels,
master interfaces) must have female
connectors and inputs (dimmers,
lamps) must have male connectors.
In short, most slave devices have
two connectors, male at the input
and female at the output.
So where does the 3-way connector
come in? This is just a question of
money. The 5-way connector is not
used as much as the 3-way connec-
tor in the audio world, so it is more
expensive. According to the stan-
dard, two of the pins re reserved for
an optional link, which is almost
never encountered in practice. Con-
sequently, some manufacturers
decided to equip their equipment
with 3-way connectors, which does
not cause any problems – at least
not in theory, but in fact the well-
Repeater
Check that terminator
is not present in repeater!
Ch
eck that terminator
is not present
in console!
If not present
on repeater,
fit terminator
at the start!
= terminator
Master Console
010035 - 16
Figure 6. Terminator locations in a DMX512 installation.
48
Elektor Electronics
5/2001
Plik z chomika:
gasma
Inne pliki z tego folderu:
bge.jpg
(26 KB)
detail1.htm
(5 KB)
detail2.htm
(4 KB)
detail3.htm
(4 KB)
detail4.htm
(6 KB)
Inne foldery tego chomika:
1974
1975
1976
1977
1978
Zgłoś jeśli
naruszono regulamin