LED PAR Mysteries

[Don Allen]

Came across an interesting situation today. The Quarry Amphitheatre has 12 5in1 IP65 LED PAR's on each side of the stage, in two rows, on truss at 6.5m trim. One of the LED PAR's was not lit, I assumed it had been unplugged as it was causing flickering in the rig. Had to drop the truss this afternoon after one of the local riggers went to replace one faulty LED PAR and lots of odd faults started to appear. When the rig was down, he checked the DMX address of the PAR and it was A000, but should have been A100. With the LED PAR at A000 the rig appeared to work normally, but when he set the address to A100 the rig went haywire with random flickering on all of the fixtures. I asked him to put it back to A000 before we swapped it out , the LED PAR went dark and the rest of the rig went back to normal. He replaced the LED PAR with one set to A100 and everything remained normal.

 

There was not enough cable to bypass any fixture, the way it was dressed, so it he left it in situ with an address of A000 !

 

Another undocumented feature of a Chinese LED PAR.

[ninjadingle]

Seems strange that you can set a DMX address of 000, most (if not all) lights (even the cheap Chinese ones I've used in the past) would only allow a setting of 001 as a minimum. Don't mean to insult but are you sure A refers to the Address? If it's a 5in1 par might A = amber?

[Don Allen]

The fixture had power but with A000 there was no light output. If it was set to menu Amber000 it would also have no light output, but when set to A100 the fixture output white light while corrupting DMX. The mode was not changed, only the up down buttons for address digits.

[bigclive]

I wonder if its settings had become corrupted including any feature that might put data back onto the network like a slave/master mode.

[Don Allen]

I will look at it today to see what the settings are to look for a cause.

[Jivemaster]

My first look would be whether the unit had entered a master/slave mode, my second check would be to look for any (if any) idea of chip and program serial number, they may never have been usable together and easiest sorted by "dumping" one to 000.

 

If you ever have time, then as I once bought a cheap set of four colour changers because they only worked when installed in one physical order, try inserting the problem unit at the end of the chain actually moving them around.

[sunray]

On the units I borrowed, A000 is the slave address. Is address 000 even accessible by the desk?

[timsabre]

There is no address 000 in standard DMX so no desk will allow it, the first data transmitted is 001. Anything actually using 000 must be sending something non-standard.

 

I would tend towards Clive's explanation of corrupted settings on the fixture. Had some LEDJ battery uplighters which would randomly change to 000 and mode 0, neither of which existed. They just sat there and did nothing though.

[sunray]

<br />There is no address 000 in standard DMX so no desk will allow it, the first data transmitted is 001. Anything actually using 000 must be sending something non-standard.

Thanks, I sometimes borrow some LED fittings which are 7 DMX channel, they can be set to master and slave mode, in which case it outputs its selected address data at 000 to 006, and adds 6 to the incoming to the output. In slave mode the only receive data between 000 & 006 Confused???

I certainly was the first time, I had: LED SET TO 13& master, 5 LED's set slave, 3 4ch packs set to 1, 5, & 9. Controller activity at 13 to 19 operated the LED's perfectly, plus the packs channels 1 to 6 and then desk channels 1 to 6 operated pack channels 7 to 12.

Fortunately I was able to wire the packs before the LED's easily after a phone call to the owner.

[dosxuk]

(sorry if you already know all this stuff, just trying to explain for those who don't)

 

The DMX data itself has no concept of channel numbers, it's just a stream of (up to) 512 8-bit values. Addresses are a human construct on top of that stream, by convention 1 - 512. We could just as easily have defined them as being 0 - 511.

 

This is relevant because a notable number of devices where you set the address by dip switch, do in fact address from 0 (allowing a 10 way switch rather than 11).

It sounds like these fixtures are doing the other popular option - using 'all off' (address 0) to change other settings in the firmware which stops them recieving normal DMX and listen for propriety data instead. What it doesn't do is unlock a secret bit of the DMX signal you can't normally get to.

[sunray]

Thanks dosxuk. really I'm a small time noise guy but I've been involved with bits of light since forever.

I did a fair bit of nightclub install for a year in mid 90's and assisted the lighting guy (exclusively Martin) so I became accustomed to hanging fittings, focussing and setting the dip switches.

Recently I've become involved with a local AmDram and suddenly I'm 'the expert' on everything with a plug or battery (which I am certainly not!).  I'm picking bits up along the way, both kit and knowledge but no I don't know all of this stuff. The first surprise with DMX is the number of channels used compared to the old Martin standard of a single address (or occasionally 2). I don't know but I assume Martin do something like use every 1/8th address.

 

Back to the situation I have badly described above. The LED pars are 2 versions of the same thing, some have 12 dip switches, the newer units have LED displays where the first digit is A to F followed by the usual 3 numbers and the same codes can be replicated on the DIP's.

In slave mode all 12 switches are OFF or A000 and copy the action of the preceding Master set to Fxxx. But now it looks like I'm trying to teach to suck eggs.

The next complication is: if every unit on the DMX line is set to F001 they will all be masters and appear at address 001, 007, 013 etc but the problem is they are 7 channel devices so there is a clash of 007 on a unit and 001 on the next.

Basically not thought out very well.

[alistermorton]

What we've mentioned before is that a lot of the lower end devices which have a master/slave configuration, where the master may be operating in automatic sound trigger mode and the slaves follow, it is sometimes the case that they don't even use standard DMX packets amongst themselves, and this will definitely cause problems. So if you have these sort of devices and want to use a desk, don't set them to master and slave, put them in a DMX controlled mode and assign them addresses. If you assign them all the same address (sat address 001) then they'll all do the same thing (whatever the desk tells them to) and will appear all to be slaved. If you give them distinct addresses then they can be individually controlled. If you have a mixture of dimmers and LEds, start your LED addressing beyond the highest address of the dimmer rack. So if you have, for example, just a pair of Act6/Betapacks at addresses 001 and 007 then the first available address is 013

 

Mixing DMX control with master/slave control has, in the past, been found to be a recipe for problems.

[sunray]

What we've mentioned before is that a lot of the lower end devices which have a master/slave configuration, where the master may be operating in automatic sound trigger mode and the slaves follow, it is sometimes the case that they don't even use standard DMX packets amongst themselves, and this will definitely cause problems. So if you have these sort of devices and want to use a desk, don't set them to master and slave, put them in a DMX controlled mode and assign them addresses. If you assign them all the same address (sat address 001) then they'll all do the same thing (whatever the desk tells them to) and will <I>appear</I> all to be slaved.

I have to do this with my Eurolites as I can't control them with DMX in master/slave mode.

If you give them distinct addresses then they can be individually controlled.  Mixing DMX control with master/slave control has, in the past, been found to be a recipe for problems.

This all makes absolute sense to me when I'm sitting at a desk but as a relative newbie (in experience as opposed to time in the environment) standing on a ladder trying to get it working is a steep learning curve.

The other problem with borrowing kit is the request to not change the settings followed by getting it wrong when resetting to original for return. These units I can borrow are plentiful but I am most likely to get a case containing 'a set' preset as master/5 slaves and the slaves are likely to have covers over the controls.  I know my Eurolites do not use DMX for master/slave, however these appear to be running as DMX and appear to work ok in a DMX universe as long as they are physically wired at the end. As I have found, if I plug the slaves directly into a desk without a desk I can happily run them with the exception of red between 001-006. The master will quite happily control subsequent dimmer packs at 000 to 006 and it also adds 6 to incoming DMX data.

What I tend to do with them now is run them on an exclusive output of a splitter to avoid confusion.

If you have a mixture of dimmers and LEds, start your LED addressing beyond the highest address of the dimmer rack. So if you have, for example, just  a pair of Act6/Betapacks at addresses 001 and 007 then the first available address is 013.

This is what I did the very first time I used them with 4 off 4 channel packs at 001, 005 & 2 packs at 009 FOH SL&SR (top row of a 24 channel desk) with the all the packs wired before the master PAR at 013. The second show, at a different venue, I had 3 packs after the LED's and one (009) before, it did take a lot of effort and replugging before I realised the action of the master as pack 009 worked as expected but the others seemed to be all over the place.

ie DMX universe: Desk -- pack 009 -- master LED 013 -- slaves A000 -- packs 001,005,009.

 

I'm still amazed at the operation of this forum where so much useful advise is freely offered without all the argy bargy I see elsewhere.

[timsabre]

It should not make any difference where in the physical chain of cabling each light is connected. If it does then you have a wiring fault somewhere. Even if there is a "Master" fixture generating the DMX it can be anywhere in the chain.

The only way to isolate one fixture from others is to put a DMX splitter in the cable.

[sunray]

Typo correction in my second section: if I plug the slaves directly into a desk without a MASTER PAR I can happily OPERATE them with the exception of red between 001-006

It should not make any difference where in the physical chain of cabling each light is connected. If it does then you have a wiring fault somewhere.

Not in this instance Tim, The master LED PAR operates it's slaves using DMX 000 to 006 (slaves set to A000 or all dips off) and the master adds 6 to the DMX data for any subsequent units, ie DMX 001 data from the desk is changed by the master PAR unit to DMX 007.

The only way to isolate one fixture from others is to put a DMX splitter in the cable

I do this now to reduce the confusion and even use different coloured leads.