DMX problem

[sunray]

My semi regular Sunday dinner event yesterday where a singing teacher is able to give her students the chance of being on a stage.

I use a fairly simple setup, a pair of profiles and pair of fresnels with no gels on 4ch QTX pack for solo & same again for flood. About a dozen cheap LED PARS to add some colour and flashing.

This time round they sung 'Mama Mia' and 'Attention' so used 'the steps', basically mid stage (4 steps normally used front of stage) so I added 2 profiles, 2 fresnals, pack and 3 LED PARS on an additional desk/universe. One of the PARS (3rd in the line of 5 devices) was flickering.

As can be seen the sheath has shrunk back and screen wire has fatigued as the end of the sheath/strain relief junction.

So this is not the first time I've trouble shooted on DMX runs however this lead was between 4th & 5th fitting but it seemed it was only only the 3rd fitting affected.  I did change the fitting and also swapped 2 over to rule out fittings.

I haven't had chance to remove all leads and test but is that a normal symptom? I'd have expected all or maybe those after the break to be affected, not just one.

 

 

 

Edited July 4, 2022 by sunray
I'm also surprised the lack of screen created this problem.

[timsabre]

Sounds like a data reflection at the fault point, I have seen this happen before where the fixture just before or 2 before the fault flickers. 

[sunray]

2 hours ago, timsabre said:

Sounds like a data reflection at the fault point, I have seen this happen before where the fixture just before or 2 before the fault flickers. 

I understand the theory behind that and assumed it's something of the ilk but knowing some of the abortions I've come across; twin flex, choc bloc, branches, one legged etc which have seemingly worked perfectly it just seems a little surprising this happened.

For what it's worth the system was working properly other than the one fitting gently flicking when off or on.

[timsabre]

Did you check the rest of the cabling, I agree it sounds unlikely. I too have seen DMX work perfectly over some awful wiring including multiple passive y-splits and lawnmower cable. But I've also seen reflected twitching like you describe... 

[DrV]

I think this serves as an object lesson in why you should always follow the rules regarding cabling and terminating DMX wiring. Sure, it may work with noncompliant cabling but when it fails it doesn't necessarily fail in the way you'd expect.

As an aside, it would have been very interesting to look at the signals with a scope to see exactly what was happening on the wires.

Dave

[sunray]

No not yet unfortunately,  it may be several weeks before the hall and my calendar coincide.

I once found a whole, albeit only half a dozen or so dimmer packs, wired with twin speaker flex, the wires were twisted tgether, folded over and stuffed in the XLRf holes. The packs were dotted around the venue so plenty of chance of interference.

judging by the layer of matted dust on everything it seemed to have not been touched for a while.

[sunray]

2 minutes ago, DrV said:

I think this serves as an object lesson in why you should always follow the rules regarding cabling and terminating DMX wiring. Sure, it may work with noncompliant cabling but when it fails it doesn't necessarily fail in the way you'd expect.

As an aside, it would have been very interesting to look at the signals with a scope to see exactly what was happening on the wires.

Dave

I do agree and carry a terninator in various bits of kit, if I think about it at the time I'll fit one but all too often I reach a 'that'll do' position with adding fittings and adding a terminator is very rare for me.

As mentioned this is the first time I've had this sort of problem on my 'toy' size systems and to be honest I don't 'expect' faults on such a small scale.

That said I've been around long enough to not be surprised when cables break with the sort of handling they get.

[DrV]

On reflection (oops!) the thing missing here is the reference connection. The system would thus be very vulnerable to mains - earth leakage. Without a reference, there is a stronger possibility of the DMX receivers reaching their common mode limit (7 volts from memory).

[sunray]

13 hours ago, DrV said:

On reflection (oops!) the thing missing here is the reference connection. The system would thus be very vulnerable to mains - earth leakage. Without a reference, there is a stronger possibility of the DMX receivers reaching their common mode limit (7 volts from memory).

That was heavily in my thought process but the 3 LED PARs are plugged in to a 4 Way extention lead so I'd mentaally ruled it out. Unless there is an earth wire fault too which I hope is not the case as I did a continuity test on each fitting prior to use.

[Jivemaster]

Rarely do DMX faults happen inside a box (desk, splitter luminaire etc. ) Faults usually happen in a lead and usually at a connection.

[DrV]

1 hour ago, sunray said:

That was heavily in my thought process but the 3 LED PARs are plugged in to a 4 Way extention lead so I'd mentaally ruled it out. Unless there is an earth wire fault too which I hope is not the case as I did a continuity test on each fitting prior to use.

This means you are relying on the mains earth to provide the reference. There could be all sorts of noise being introduced into the reference from high frequency currents which are being dumped to earth.

Look at a typical smps and you will find capacitors between the extremely noisy high voltage side and mains earth. The current flowing through them is going to be superimposed on your reference line. And even though the connection via that route may look like a low impedance path at DC, at high frequency it may be anything but. So the upshot is that the mains earth on one fixture may be flapping up and w.r.t. the mains earth in another fixture.

Also, the DMX reference (pin 1) will not (should not) be directly connected to mains earth anyway, other than at the controller.

Actually those two paragraphs should be in the opposite order but I'm doing this on a phone and it's late!

[sunray]

8 hours ago, DrV said:

This means you are relying on the mains earth to provide the reference. There could be all sorts of noise being introduced into the reference from high frequency currents which are being dumped to earth.

Look at a typical smps and you will find capacitors between the extremely noisy high voltage side and mains earth. The current flowing through them is going to be superimposed on your reference line. And even though the connection via that route may look like a low impedance path at DC, at high frequency it may be anything but. So the upshot is that the mains earth on one fixture may be flapping up and w.r.t. the mains earth in another fixture.

Also, the DMX reference (pin 1) will not (should not) be directly connected to mains earth anyway, other than at the controller.

Actually those two paragraphs should be in the opposite order but I'm doing this on a phone and it's late!

Although I agree with all this, especially the interference

On 7/5/2022 at 9:04 AM, sunray said:

I once found a whole, albeit only half a dozen or so dimmer packs, wired with twin speaker flex, the wires were twisted tgether, folded over and stuffed in the XLRf holes. The packs were dotted around the venue so plenty of chance of interference.

Theoretically the earth should make no difference to a balanced data network, in fact I can straightaway think of 2 other RS485 systems where unscreened twisted pairs are recommended and instructions say to only terminate any screen/earth wire at one end to avoid any earth loop. In pro audio situations perfect balanced audio is frequently done with cat2/3/4/5 cable (think BBC patchbays).

I stress I know real world and theory are usually poles apart.

 

Just to bring this back to the OP, I started this thread more to see if this is a typical symptom of the fault rather than getting into technical discussions, that said I always welcome these educational threads.

[DrV]

11 minutes ago, sunray said:

...Theoretically the earth should make no difference to a balanced data network, in fact I can straightaway think of 2 other RS485 systems where unscreened twisted pairs are recommended and instructions say to only terminate any screen/earth wire at one end to avoid any earth loop. In pro audio situations perfect balanced audio is frequently done with cat2/3/4/5 cable (think BBC patchbays)...

RS485 is only a balanced network if you stay within the common mode limits of the receivers.

[Jivemaster]

We all know of a DMX network that works but doesn't comply (I knew a well paid DJ who fanned out DMX in twin flex and it worked for years) BUT the first fault finding needs to be the cable and the connectors. Because DMX is effectively RF reflections can occur at damage or breaks in cables these can be reflected back down the line and confuse or conceal the sent DMX -hence the frequent finding that the fault shows up somewhere else. 

[timsabre]

7 hours ago, DrV said:

RS485 is only a balanced network if you stay within the common mode limits of the receivers.

RS485 isn't a balanced signal in the same way audio is, it's a differential signal and it absolutely needs the ground reference connection. As you say, you'll get away with it providing all the end points float within the 12V common mode voltage range of each other. 

Edited July 6, 2022 by timsabre