Jump to content

moving lights, how do they work?


mr techie

Recommended Posts

why is there "gobo 1 rotate function " then "gobo 1 rotate low"?

16-bit control of gobo rotation. A standard DMX channel can send values from 0 to 255 decimal (00 to FF hexadecimal) - it's 8-bit. For particularly fine control of things like pan, tilt and gobo rotation, you can assign two control channels to a function - a 'coarse' channel and a 'fine' channel. the coarse channel will deal with the most significant 8 bits out of the 16, and the fine channel the least significant. So instead of being able to set a function of a moving head with an accuracy of 1/255 of its total range of movement, you can work to an accuracy of 1/65,000 (or thereabouts).

Link to comment
Share on other sites

Before just posting I did my research on the X-spot complete with 38 channels. *gasp* is it possible to get more on a fixture and what could you have? Would this lead to bigger universes or desks having to cope with more? 

 

Back to the X-spot http://www.highend.com/pdfs/products/xspot.../XtremeTech.pdf has a list of all the functions but why is there "gobo 1 rotate function " then "gobo 1 rotate low"?

 

You will also notice a "gobo 1 rotate high". What is going on here, and as is also being done for pan and tilt, is that two channels are being used to run that function as a 16 bit parameter (65536 possible values, instead of the 255 on a single channel) The net result is higher resolution. The function channel typically controls HOW the parameter channels act - forward rotate, reverse, snap to gobo, constant movement, etc. etc. etc - the full manual will tell you. The High/Low value split channels make running this type of fixtures on a "traditional" desk a real bear - the "High" channel will give coarse movement, and the "Low" fine movement. On a moving light desk which knows about this, the low channel cycles through 0-255 for every step of the high channel, giving smooth fades, etc. On a traditional desk, both channels will move X percent on a cross fade, without high resolution of the low channel, and moves can be rough or bumpy . . .

 

Also, a moving light desk will do things like snap changing colors at the start of a cue, and having the grand masters and submasters only control intensity. On a "traditional" desk, pull down the master and not only does the fixture dim, but it moves, the colors change, the gobos change . . . .etc. etc. etc. so what you have to do is create a LOT of extra cues to do blackouts, and forget that you have manual faders . .. . .

 

- Tim

Link to comment
Share on other sites

Would this lead to bigger universes or desks having to cope with more? 

 

Reaches for PLASA Recommended practice for DMX512

 

Ok you can have more channels in a RS485 (what DMX is) transmition but then it can’t be classified as DMX and will probably not be very reliable. It is a rather complex subject and I can’t really explain fully why you can’t have more than 512 channels as I don’t know how to express properly it but buy the book and it will all make sense :(, it is to do with timings and refresh rates and data packets ect…

 

Sorry for going a little :unsure:…

Link to comment
Share on other sites

A little :unsure: again is this not one of the uses for the Misterious 5th pin?

 

Do I smell a topic split?

 

 

16-bit control of gobo rotation. A standard DMX channel can send values from 0 to 255 decimal (00 to FF hexadecimal) - it's 8-bit. For particularly fine control of things like pan, tilt and gobo rotation, you can assign two control channels to a function - a 'coarse' channel and a 'fine' channel. the coarse channel will deal with the most significant 8 bits out of the 16, and the fine channel the least significant. So instead of being able to set a function of a moving head with an accuracy of 1/255 of its total range of movement, you can work to an accuracy of 1/65,000 (or thereabouts).

 

So if I understand the function is coarse (of BIG moves ) the high is fine and low is very fine?

 

Tim im guessing that by "traditional" you meen analogue running a ML?

 

 

Do I smell a topic split?

Link to comment
Share on other sites

A little :unsure: again is this not one of the uses for the Misterious    5th pin?

 

What for another DMX universe? Well for DMX you need two hot and cold lines and they need be twisted together so that any interference happens to both lines so you would need pins 4 + 5 for another universe down one cable... and personally I don't think that two universes would work very well down one cable... they would interfere with each other... They would have to bee fully screened from each other... (pins 2 + 3 from 4 + 5)

 

Anyway this is :( so I will pipe down...

Link to comment
Share on other sites

No, DMX expansion without a new universe.

 

What do you mean... how would you expand DMX without making a new universe... a new universe is making another pair of hot and cold wires to send 512... So yes down one 5 core/pin cable you could send two universes but they would interfere with each other... there is no other way of expanding DMX channels down one cable and it is not recommended… The only way to have more than 512 DMX channels is to create another universe and this is defined as a pair of transmission lines not a physical cable… you could technically but I wouldn’t recommend it have 11 DMX universes down one 24 core cable with a few spare lines for grounding but you would have very bad interference and it would not really work…

 

What I am trying to say is one cable doesn’t equal a universe rather one pair of twisted lines (hot and cold line) transmitting 512 channels make a universe…

Maybe I have misunderstood you but how would you use the 5th pin to expand the DMX without creating another universe?

 

Maybe I have misunderstood you but how would you use the 5th pin to expand the DMX without creating another universe?

Link to comment
Share on other sites

So if I understand the function is coarse (of BIG moves ) the high is fine and low is very fine?

Almost, but not quite. In this particular example, the Function channel determines what mode the gobo rotation is operating in (indexed or continuous. basically), while the Coarse and Fine channels respectively control the most significant bits and least significant bits of the values which tell it where to index the gobo or how fast it should spin.

Link to comment
Share on other sites

A little :unsure: again is this not one of the uses for the Misterious    5th pin?

 

Do I smell a topic split?

 

 

16-bit control of gobo rotation. A standard DMX channel can send values from 0 to 255 decimal (00 to FF hexadecimal) - it's 8-bit. For particularly fine control of things like pan, tilt and gobo rotation, you can assign two control channels to a function - a 'coarse' channel and a 'fine' channel. the coarse channel will deal with the most significant 8 bits out of the 16, and the fine channel the least significant. So instead of being able to set a function of a moving head with an accuracy of 1/255 of its total range of movement, you can work to an accuracy of 1/65,000 (or thereabouts).

 

So if I understand the function is coarse (of BIG moves ) the high is fine and low is very fine?

 

Tim im guessing that by "traditional" you meen analogue running a ML?

 

 

Do I smell a topic split?

 

Not really correct. On the high byte, there are 255 addresses across the entire range of the fixtures motion. As such, ANY move will only be able to happen with this resolution at it's finest - for example, a head that can move 360 degree, will have a coarse resolution of 360/255 degrees, or about 1.4 degrees per step, no matter how far the individual queue caused the fixture to move. The fine channel give an additional 255 points BETWEEN each of these, or about .005 degrees. 1.4 degrees at say 75 foot throw will amount to about 1.5 feet of motion, wheras the fine of .005 degrees will be about .08 inches. Coarse and fine are correct - fine and very fine is being way too generous!

 

Secondly, when I say "traditional" console, I refer to any desk which does not specifically have software that knows movers - just dimmers. Still DMX . . . .

 

On a traditional console, typically all channels move to their new values over a fixed amount of time, with the same curve. So, let's say you have two pan values, and for grins, I will use a high and a low byte for each position:

 

Pan High Pan Low

Q1: 40 5

Q1: 200 50

 

Where the values of 40 +5 and 200+50 correspond to the two points of aim desired. On a console that

understands that this is a two byte value, the overall value will move together:

 

The Low will go from 5 to 255, the high will increment to 41 while the low goes back to 0. The low will then move to 255, the high will increment to 42 and so on. As such, the move can be in steps as fine as the .005 degrees I detailed in my example above.

 

Now on a traditional console, both the low and high will move together, the low from 5 to 50 and the high from 40 to 200 together. As such, the fine values between the steps of the high value will not be handled, and the fines step the fixture can move in is the 1.4 degrees from the example above. If this is a fairly long throw, that can bit a fair distance at the intended area. And on slow moves, this will look like *jerk* - stop *jerk* - stop etc. Not necessarily good. Now, a lot of fixtures have a speed parameter (on the HES stuff it is called MSpeed) on which you can apply some timings internal to the fixture, which can smooth this out. It's a pain, but doable.

 

The next issue is one of color and gobo changes - let's say that you have a couple of queue as such (and I am going to use single values here so I don't confuse everyone to hell and back . . . . . )

 

PAN TILT COLOR GOBO DIM

Q1 20 20 RED open full

Q2 20 20 RED open 0

Q3 120 120 YELLOW cone 80

 

The goal here is to light an area in red, blackout, and open in a gobo in another color on another area. On a moving light console, you can let this happen two ways. If you want to see the lights move, run as shown - the color and gobo will snap to thier values the instant that Q3 starts, and then the fadein and move will happen according to the Queue time. If you want to fade in in place on Q3, then (At least on a Hog) put a Mark-Q in front of Q3, which causes all parameters OTHER than intensity to snap to the values in Q3 as soon as Q2 hits black. Queue three then fades in already in position.

 

Run these three queues on a "traditional" desk, and you will see the backout happen OK, but then as Q3 runs, you will see the color slowly roll from RED to YELLOW over the queue time, as will the gobo, and the movement. To do this on the traditional desk, you would need to insert anotherr queue which would be a copy of Q3 but at DIM 0 to preset the positions and colors. And you would not be able to snap change colors while making intensity changes at all, unless you can run two queue stacks, and separate the channels with separate times.

 

Not impossible to run movers on a traditional desk (I did it a fair amout when I only had 1 or 2 movers, and mostly used as specials) but if you have a lot of standard dimmers, and it's not a tracking console, then the need to copy a lot of queues to set moves, etc. can run you out of memory in a hurry if it is not a pretty simple show.

Link to comment
Share on other sites

What for another DMX universe? Well for DMX you need two hot and cold lines and they need  be twisted together so that any interference happens to both lines so you would need pins 4 + 5 for another universe down one cable... and personally I don't think that two universes would work very well down one cable... they would interfere with each other...  They would have to bee fully screened from each other... (pins 2 + 3 from 4 + 5)

 

My attempt to explain a couple of things. DMX uses the protocol RS485 -- this is similar as the way RS232 is for serial ports on a PC. It basically wiggles the voltage on a wire up and down to send a series (hence its a serial port) of 1s and 0s. The 1s and 0s are grouped together into 8 bits -- a byte (8 ones or zeros gives you 256 combinations, hence the values 0 to 255) (though get extra one's tacked on so you know where a byte starts and finishes). Greatly simplified, but in order for the signal to be recognised in time, (shorter than this and things may miss the pulse) each bit must be 4 microsecs long -- it can't be shorter, so in theory you can get 250,000 bits per second. So far so good hopefully....

 

So, a moving fixture needs to know which channel you mean and the value for it. The way DMX handles sending these two parts of data is to only send the channels' values in sequence. So it sends a byte for channel 1, then 2, then 3 and so on. However obviously it needs to know when its starting to send the value from channel 1, and then to count up. It does this by sending a break -- basically it stops wiggling the voltage for a fairly long period of time, so all the receivers know the next packet will be the first. With 4 microsecond pulses, and the breaks, and the start/stop bits etc, it means that only 512 channels can be sent per DMX run.

 

Other point, about interference. RS485 is a differential protocol, this means that the signal is based on the potential difference between the two wires (rather than just connecting one wire to ground and putting a signal on the other). Because of this, DMX is more imune to interference -- the idea, is that any stray voltage (by inductance, capacitance, floating earths, pixies etc, whatever) equally affects both wires, and because the signal is based on the difference between them, it should be ignored. Similar to balanced lines for sound I believe. There should be no problem sending 2 DMX signals down 4 wires in the same cable. Ethernet (and I mean the physical cabling not sending DMX packets over it) does a very similar thing and works fine.

 

coming back to fine and coarse, think of it as tens and units. The coarse is given by the tens and the fine by the units. 10 units make up a coarse unit. In DMX its the same, its just that its really units and 256s!

 

In terms of using more than 38 channels, fixtures like the pixel batterns can eat channels. Some were designed to use 60 DMX channels each! Even with 10 cells and RGB for each, there's 30 channels gone in a block!

 

hope it helps

Link to comment
Share on other sites

Before just posting I did my research on the X-spot complete with 38 channels. *gasp* is it possible to get more on a fixture and what could you have? Would this lead to bigger universes or desks having to cope with more?

 

As others mentioned already, one universe has always 512 channels. That's the definition of a universe (DMX of course, a real universe is a little bit bigger).

 

So if you need more then 512 channels then you need more universes. 2 universes gives you 1024, 3 universes 1536, and 4 universes 2048 channels.

 

As you found out, yes there are fixtures which take 38 channels or more. That means that having 12 of those would already almost take up one universe alone.

 

Or take e.g. a Catalyst media server. It has 8 layers and each layer takes up 40 channels. That alone also would take up almost one entire universe.

 

So make sure you choose a controller which can acoomondate what you need and that you don't run out of channels (universes).

 

Dan

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.