Another odd DMX issue - data over Pathport?

[Sayen]

Here's the setup:
- Controller is an ETC Element into a Pathport, splits out to several locations including several two port nodes. All Pathports and the Element are running latest firmware/software.
- One Pathport splits out to two electrics. One pipe has 5 Altman Spectra Cyc 200s, the other is loaded with Elation OptiTripars.

If the TriPar line is connected, the Spectra Cycs refuse to work, and one turns a dull red. Disconnect data from the TriPars, and the Cycs work fine. Power off but still connect the TriPars, and the Cycs die. Unplug the data cable and they work. Change Pathports, same result. Terminate, don't terminate, try different cables, try every TriPar individually, same results.

Originally we sent the Spectra Cyc that turns red to Altman, and they were unable to find any problems. Disconnected from the TriPars and they work just fine. They even work with the TriPars as long as they don't share the same output Pathport node - even though the Pathports are two port units, and each set of lights is on it's own run. And they seem to work with other DMX devises - tried a Source4 LED and a hazer with no issues.

I'm stumped, but this causes significant issues in our setup to lose an entire DMX output location. Any thoughts?

 

[GaryDouglas]

Hi Sayen,

It sounds like you may have a classic DMX grounding issue. The DMX ports on the Pathport are not isolated from each other, but they're isolated from mains power, so the ground on those units is going to 'float' to wherever the DMX common is. When there are 2 VERY different grounds on each of the Pathport's DMX ports however, you can see problems like this. It sounds to me like the AC power on each pipe has radically different grounding, causing one pipe's DMX to float outside of the detectable range (as the other pipe's ground may be 'stronger').

It would be a good exercise to take a volt meter and look for AC and DC voltage between DMX pin 1 and true ground, and between pin 1's at the end of each pipe's run of DMX. If you see voltage, You can try temporarily lifting the ground pin ( pin 1 ) on the DMX running to 1 pipe and seeing if the problem goes away. You should also look at how pin 1 on the DMX cables is handled ( are there any broken wires?). It's best to look at how one of those pipes is grounded ( I'd look for a bad AC cable!) and make sure it truly is.

Additionally, if you have an isolated DMX splitter (one where all of the ports are isolated from each other) you could put that inline with one of the pipes' DMX lines. This is not ideal, but should fix the problem.

NOTE: Be VERY careful with your 'hot' DMX cables. There may be abnormally high voltages on pin 1 if there's indeed an issue with grounding on 1 pipe!

Hope this helps,

Gary Douglas
Director, Engineering and Design
Pathway Connectivity

 

[Codeguy]

Hi Sayen,

It sounds like you may have a classic DMX grounding issue. The DMX ports on the Pathport are not isolated from each other, but they're isolated from mains power, so the ground on those units is going to 'float' to wherever the DMX common is. When there are 2 VERY different grounds on each of the Pathport's DMX ports however, you can see problems like this. It sounds to me like the AC power on each pipe has radically different grounding, causing one pipe's DMX to float outside of the detectable range (as the other pipe's ground may be 'stronger').


Gary Douglas
Director, Engineering and Design
Pathway Connectivity

Hey Gary....
I am surprised this grounding issue was not identified in the design process..say in August-September of 2000.
It might have been advisable to separate both DMX i/o sections using separate DC to DC converters or , less expensively, a dual winding secondary transformer driven from one of the spare counter/timer outputs of the processor(at say 10khz)....coupled with relays (or opticaly isolated mosfets) to connect/disconnect DMX pin one to ground ....depending if the port was an input or output.
Oh....wait.....
Well as long as "We remain focused" and "Have a gooey".....and in particular use the number one design rule "NO ONE would EVER want to do that".....
S