Weird Scroller problem

[Bryson]

So, I have a show in that's using 39 scrollers, and 33 are behaving weirdly. 6 of them are behaving as expected.

 

They work fine if you move the DMX value slowly, or in increments of less than 20% - but a snap cue of 20% or more makes them reset before going to the correct value after the reset.

 

None of the PSUs are overloaded (they're PSU12s and they each have 6 or 9 units plugged in)

 

The 6 that work fine are not the last or the first in the DMX chain. There's also some Chroma-Q Colour block PSUs and a DMX dowser in the chain, but removing them doesn't seem to make much difference.

 

I'm wondering if it could be due to excessive chain length (these are all on a single chain so probably 300'+ of DMX in the chain with 11 devices in the chain) so I'm rushing in a DMX splitter to hopefully resolve the issue, but has anyone else had the same problem?

 

EDIT: Just a thought: when counting "devices" do you count PSUs, or Scrollers? If the second, then I'm way over 32..... Would splitting help, or do I need a whole new universe?

 

EDIT 2: Yes, it's terminated, DMX pedants. :P

[madorangepanda]

What scrollers are they? I'm guessing something Chroma-Q as you have PSU12s.

 

Does the 300' of DMX cable include the scroller cable?

 

Each PSU only likes about 60-70m of scroller cable running from it normally.

 

Edit: You count scrollers not PSUs, so you have way over 32. But it seems an odd problem to be having.

[gareth]

If you're talking in terms of devices on a DMX line, then you count PSUs, but not scrollers. The data portion of the PSU is effectively a DMX buffer, so you get a 'freshly-generated' DMX output onto the scroller loop.

 

It could be worth checking what the maximum permissible 'head-feet' is on the particular scroller PSUs that you're using. Scroller systems often define maximum cable length in this way as opposed to simple linear feet - it's basically a way of expressing the idea of the sum total of the amount of cable between each scroller and the PSU that's feeding it.

 

Say, for example, you had PSU -> 30' cable -> Scroller -> 30' cable -> Scroller -> 10' cable -> Scroller .... in this system, although the overall cable length from PSU to the last device in the chain is only 70', the total effective head-feet of cable would be 30+60+70=160. This takes into account the amount of voltage drop that you're going to experience at each device based around how far it is from the PSU and how many other devices are before it in the chain.

 

(There's probably some sort of mathematical way of expressing how this changes once you add a return from the last scroller to the PSU into the equation - but I don't know what it is!)

[Bryson]

300' plus scroller cable, but I'm way under the 60-70m per PSU. they are Chroma Qs (*spits*).

 

So I suspect the 32 device limit is the problem here*. Will stick the splitter in and that should fix it. (Crosses fingers...)

 

 

*=Although, isn't each PSU a little splitter, really? Hmmm.....

 

 

EDIT: Having seen Gareths post, maybe not. Hmmm. None of the scroller-PSU runs are unusually long. In fact the PSUs are remarkably close to the scrollers. I'm just flummoxed by how they behave fine with small changes, but big or fast changes freak them out.

[Bryson]

Tracked it down: turns out that whoever plugged the system up didn't spread the scrollers equally across the PSUs - so one had 18 scrollers into it! Clearly overloaded. Redistributed and all is well. Weirdly the problem seemed to spread downstream of the overloaded PSU.

 

Next week at the Roundhouse: Crew Training about why it's called a PSU12 http://www.blue-room.org.uk/public/style_emoticons/default/dry.gif

[pritch]

one had 18 scrollers into it!

 

Crew Training about why it's called a PSU12

 

Counting in hexadecimal? :unsure: