The curious part of that is that Rack 1 CEM is not showing a Port B error, just the no data warning.
Sensor Rack Woes - DMX Input Error
- Thread starterStradivariusBone
- Start date
The curious part of that is that Rack 1 CEM is not showing a Port B error, just the no data warning.
Mac Hosehead
Well-Known Member
Blue corrosion is bad. It means these racks are located in a humid environment. I would definitely mention it to ETC. I would think any connectors, including termination PCB, on the back-plane are suspect for failure. I have seen Sensor and Unison processors need to be replaced because of this. I often see this in electrical rooms where the only access is through an outside door and no air conditioning present.
It is in a humid environment and at one point there was leaking in that room. They put in a seperate AC unit to maintain it at 65 constantly to eliminate that issue, but I know a lot of damage was done.
I cleaned the terminals and lightly polished the copper with some high grit sandpaper. They are making good contact with the PCB. No change in behavior.
Talked with ETC for about an hour, I'm going through dimmer by dimmer in rack 1, turning each on to 75% one at a time and making notes on which other dimmers seem to be affected. So far almost every dimmer seems to be making another flicker on (sometimes 8-10 others) and all of them at least make themselves flicker. The interesting thing is the ones that are active don't go to full, they merely cut out. I was watching this on the rack for the most part, and the signal lights would cut out, almost opposite the other random dimmers coming on. Not sure if that's a clue.
I cleaned the terminals and lightly polished the copper with some high grit sandpaper. They are making good contact with the PCB. No change in behavior.
Talked with ETC for about an hour, I'm going through dimmer by dimmer in rack 1, turning each on to 75% one at a time and making notes on which other dimmers seem to be affected. So far almost every dimmer seems to be making another flicker on (sometimes 8-10 others) and all of them at least make themselves flicker. The interesting thing is the ones that are active don't go to full, they merely cut out. I was watching this on the rack for the most part, and the signal lights would cut out, almost opposite the other random dimmers coming on. Not sure if that's a clue.
BillConnerFASTC
Well-Known Member
Sounds like an excellent time to upgrade to Sensor 3.
I actually have good news, Bill! Remember that thread from maybe 3 years ago about my electric and the issues with it? Our district finally allotted funds to fix it!
We'll get on the Sensor 3 sometime in the 3rd quarter...of 2021...
Jay Ashworth
Well-Known Member
Jay Ashworth
Well-Known Member
But after catching up on the thread, I have a different question:
Is the access from an outside door?
Cause really cold metal will just condense the crap out of incoming humidity; the AC has no hope of knocking it down that fast if you open an exterior door...
Is the access from an outside door?
Cause really cold metal will just condense the crap out of incoming humidity; the AC has no hope of knocking it down that fast if you open an exterior door...
Hey Jay- I looked it up and the local Catholic radio station just moved their studio maybe 300 yards away from our building, but their transmitter is about 5 miles from us. It was recently upgraded to a higher power rating, but it's still pretty far away.
The room is all interior, like I said they had a ton of leak problems with this building and at one point several modules were damaged from water and had to be replaced. I feel right now that the problem is in the rack itself, we've eliminated anything else on the input side that could be causing it. There is a large fan coil that runs off our chiller and a residential style HVAC unit that runs 24/7 (but only kicks in when the fan coil turns off and the temp rises) so you could hang meat in that dimmer room.
The room is all interior, like I said they had a ton of leak problems with this building and at one point several modules were damaged from water and had to be replaced. I feel right now that the problem is in the rack itself, we've eliminated anything else on the input side that could be causing it. There is a large fan coil that runs off our chiller and a residential style HVAC unit that runs 24/7 (but only kicks in when the fan coil turns off and the temp rises) so you could hang meat in that dimmer room.
I hate to say it, but cleaning contacts with sand paper is a very bad idea. Things may look shiny, but I've seen many shiny and bright connections that were electrically lousy. If that was only done to high voltage, copper to copper connections, then it probably did no harm. For cleaning big, copper connections I like to use Scotch-Brite 7447 or 8447 pads, then liberally flush with a cheap, low residue, aerosol contact cleaner. If sandpaper was used on low voltage, low surface area connections, then I would be very concerned about embedded residues and scratches.
I don't know anything about the internal layout of Sensor racks, but I have run into problems with electro-mechanical connection problems so many times in my career. Given the environmental issues the rack has been exposed to (high humidity, heat, salt air, pollution) I would bet that's the root cause here. Those conditions are ripe for oxidation, which is often invisible on tin and gold plated connectors. (This is an area where most tech support folks, that don't have much field experience, have no idea what happens to electronics out in the real world.)
Every low voltage, electro-mechanical connection in the system is suspect at this point. Every card edge connector, every ribbon cable, every socketed IC, every pcb board to chassis, and chassis to chassis screw that might be relied upon for grounding.
The best way I know to clean connections is to flush with Caig DeoxIT D5. If any low voltage connections that got the sandpaper treatment, they should be scrubbed with a clean, cotton cloth or cue tips wetted with D5.
The D5 will go all over, and leave a residue, which does not hurt anything, so don't be afraid to be messy with it. The main thing is to flush contaminants away.
After the D5 drys for a few minutes coat with DeoxIT G5, which is a protectant.
Residues will eventually dissipate.
I cannot over emphasize the good results I have had from the Caig products. They work when others fail and are essentials in my electronic tool box.
Jay Ashworth
Well-Known Member
Well I learned something today. Bear in mind I was using 600 grit which is closer to polishing than anything else, but I'll take your experience over mine in that area. I had a bottle of Deoxit somewhere at my house, I'll have to dig it out and go to town with it. Would you think that a problem of this nature would cause a working CEM to fail worse over time as it was exposed to the rack but not fail in a bench-scenario?
That CEM relies on a bunch of external connections to function. Bench testing the CEM eliminates the rack completely by mating it to their test jig. If the CEM passess bench testing, it points back to the rack, or another module in the rack, or the power conditions. The rack itself is nothing but a bunch of wires and connectors of various types, so the answer to your question is yes.
I recall that in the past, you described that pulling things apart and reassembling would settle things for short periods of time. That screams mechanical connection issues.
Do yourself a favor and buy the aerosol versions of D5 and G5.
I recall that in the past, you described that pulling things apart and reassembling would settle things for short periods of time. That screams mechanical connection issues.
Do yourself a favor and buy the aerosol versions of D5 and G5.
Last edited:
JD
Well-Known Member
Ding! Ding! Ding!
I think you hit on the answer. 20+ years in electronic service and in almost every instance (90%) where physically moving things around fixed a problem for awhile, the ultimate answer was connector/connection related. Low voltage/low current connectors are especially prone as there is not enough current/voltage to break through the layer or two of oxide molecules that form on the surface of metal. Often, the offending connector looked bright and shiny, but cleaning and reseating did the trick. Although I have been known to use a pencil eraser, you don't ever want to use anything that is abrasive. My preferred cleaners are all banned now, so I will defer to cleaners mentioned above. The only other offender was the older style of plated-through holes on non-G10 board, but I do not believe this applies to the CEM.
So the Deoxit treatment was a failure. Treated the card edge connectors on both CEM's and the sockets on the backplane and nothing changed. I'm at a loss. We are going to have to program backup looks and have someone sit in the dimmer closet for our next show. This is getting ridiculous. It is essentially unusable at this moment and I've got a full docket of shows cranking up in a week.
On the backplane there's a dipswitch module that has a bunch of passive components on it that routes to a 6 or 8 pin card edge connector. I noticed some corrosion on that piece where the ribbon cable is soldered in rack 2 (the mostly unaffected rack). What do those dipswitches do? Rack ID?
edit- that's the terminator, I'm wondering if a failure there could be related to these issues.
On the backplane there's a dipswitch module that has a bunch of passive components on it that routes to a 6 or 8 pin card edge connector. I noticed some corrosion on that piece where the ribbon cable is soldered in rack 2 (the mostly unaffected rack). What do those dipswitches do? Rack ID?
edit- that's the terminator, I'm wondering if a failure there could be related to these issues.
Last edited:
So I connected a DMX line directly into the backplane, bypassing the existing cable and disconnected rack 2, isolating just rack 1 and still had errors. Connected everything back up and swapped CEM's from rack 1 and 2 and the errors continued to follow the CEM. Troubleshooting still points at the CEM, but ETC says nothing is wrong with it.
Without knowing what the CEM looks for from the switches, I'm not wanting to mess with them. Going to call tech support again in a bit. I tried to upload the dimmer test they had me do. Typically there is a pattern to the dimmers that "engage" other dimmers during an error state. Usually the numeric dimmer before or after (e.g. 89 will mess with 88 and 90) and it's usually very consistent about which dimmers are affected.
I tested the signal coming from my controller at the rack by using the input side of an ENTTEC USB DMX Pro and had no issues. It's getting nice and clean serial data.
Without knowing what the CEM looks for from the switches, I'm not wanting to mess with them. Going to call tech support again in a bit. I tried to upload the dimmer test they had me do. Typically there is a pattern to the dimmers that "engage" other dimmers during an error state. Usually the numeric dimmer before or after (e.g. 89 will mess with 88 and 90) and it's usually very consistent about which dimmers are affected.
I tested the signal coming from my controller at the rack by using the input side of an ENTTEC USB DMX Pro and had no issues. It's getting nice and clean serial data.
JD
Well-Known Member
The boards I have seen have a plethora of connectors, headers, and IC sockets. (not sure about the newer ones.) Any one of these could be a source of problems. Also, there are a couple of high pin density processors. Anybody who has serviced flat screen TVs know how these can develop pc land cracks at the pin connections. Basically, the large pack does not flex, but the board can. All these problems may disappear for a while once the board had been transported. It may have checked out ok, but swapping between racks kind of confirmed it's on the CEM mod.
Mac Hosehead
Well-Known Member
ETC won't send you another loaner? I had a CMEd once that would not work in its rack. ETC could not find anything wrong with it. After sending it back 3 times they sent me a different one. That one worked.
I see CEMs for sale on e-bay but that might be more of a last resort. You would have to make sure the software version matches your other CEM.
I see CEMs for sale on e-bay but that might be more of a last resort. You would have to make sure the software version matches your other CEM.
This rack has been giving us trouble for the better part of a year. We sent this current CEM back (this is not the original one) and they swapped out the 7517 and optocoupler IC's, but could find nothing else wrong with it. It seems that putting a working CEM will be fine for a while, but then degrade over time. But the fact that it won't replicate the problems in other racks or on the bench is very odd.
Something I noticed yesterday, there is no common between rack 1 and 2 (pin 3 on connector J2). It's labeled as ISO common on the schematic for the rack, and it's duplicated on pin 12 which is connected between rack 1 and 2. On the schematic it has pin 3 wired to the shield of the cable coming into the rack (which it is in rack 1) and then wired in parallel to rack 2, J2, pin 3 (which it's not).
I did talk with ETC again yesterday and they have some things we'll try out today. They've been (as always) super patient and helpful.
Something I noticed yesterday, there is no common between rack 1 and 2 (pin 3 on connector J2). It's labeled as ISO common on the schematic for the rack, and it's duplicated on pin 12 which is connected between rack 1 and 2. On the schematic it has pin 3 wired to the shield of the cable coming into the rack (which it is in rack 1) and then wired in parallel to rack 2, J2, pin 3 (which it's not).
I did talk with ETC again yesterday and they have some things we'll try out today. They've been (as always) super patient and helpful.
Similar threads
