Control/Dimming Repairing Chauvet DMX-4 dimmers.

[JD]

Well, the triacs could be popped. If they are being gated (usually those channel LEDs are in series with the opto-cuppler so looks to be the case), then there is not much more to the circuit. The triac output usually goes through a choke (very low failure rate) then through the fuse holder and then to the receptacle. The only other problem could be cracked PC lands on the board around the connections. I did see a pack once where the caps on the fuse holder were the wrong ones and the fuses were not making contact, but it more than likely sounds like the triacs are open. I thought Elation had a rebuild-replace through their customer support, so I would put a call or email in there first.

Hopefully someone else in the Central NY area will chime in on local service up there.

 

[danhr]

Unfortunately I don't remember the sequence of events before the failure- it was a while ago and I took it out of service and did the basic checks and then set it aside. It's odd that all 4 channels would fail simultaneously- we have used these packs for years with few problems. The triac is usually the weak point- even in home brew units I built (using diacs not optocouplers) I had only triacs fail. (saving $$ by using 15 amp ones, I know....) I checked for bad solder joints, ribbon cables, etc. and all looks good. I suspected it might be the main chip since it drives all 4 channels but if the leds follow the optos that wouldn't the case. I'll try a new triac in one channel just for s***s and giggles but we might just end up with spare parts here......

 

[JD]

Unfortunately I don't remember the sequence of events before the failure- it was a while ago and I took it out of service and did the basic checks and then set it aside. It's odd that all 4 channels would fail simultaneously- we have used these packs for years with few problems. The triac is usually the weak point- even in home brew units I built (using diacs not optocouplers) I had only triacs fail. (saving $$ by using 15 amp ones, I know....) I checked for bad solder joints, ribbon cables, etc. and all looks good. I suspected it might be the main chip since it drives all 4 channels but if the leds follow the optos that wouldn't the case. I'll try a new triac in one channel just for s***s and giggles but we might just end up with spare parts here......

Is odd all four are out. Still, the controller chip does everything, and since you are getting drive signals and the channel LEDs respond, the chip sounds like it's working. Optos are pretty reliable, and to have all four go would be more than strange. Sure there's not a trace blown off the back of the board? As I said, usually when a triac goes, it shorts leaving that channel on. I think the problem is something a lot more simple.

 

[danhr]

All circuit traces are fine, wiring is fine...earlier I found a cold-ish solder joint which fed the power transf. but ruled that out since the data side is OK and the power side is always hot in these packs. My thought is some component has failed on the data board that is common to driving the optos on the main board. This is certainly strange, there isn't much to these units to go bad and when something does it is usually pretty obvious.....

 

[danhr]

P.S. A new triac, same issue.....

 

[gafftapegreenia]

Felt the need to mentioned that I just finished replacing the 12 Triacs in my packs. Great success! Also, at least two were stuck in off position, the rest in constant on.

 

[JD]

Unfortunately I don't remember the sequence of events before the failure- it was a while ago and I took it out of service and did the basic checks and then set it aside.

At this point, those clues become more important. Much of what you said indicates all the major parts of the pack work, but it doesn't!
Do you remember if all four were lost at the same moment?
Transformer, power supply, DMX decoder, general processing, and (with the one you replaced) triacs are working. Something like a common buss is out. The dc regulator is working, so that moves us to something post-opto-cuppler.

Because we don't want to get anybody killed, the live buss testing should be handled by someone who is qualified, but we appear to be at that point. This is going to turn out to be something simple but quirky, like an opened neutral connection on the receptacles, or something just as far off the wall. Try to remember back to when it first went.

 

[danhr]

No worries- I have been concocting gadgets for 40+ years and safety is job #1. I can't even remember back when the pack died- at least 1-1/2 years ago plus it was a VERY chaotic 1st tech and I was trying to do several things at once (and we all know how well that works....). Triacs are getting power; 120v on both HV leads and 0v when a load is put on any one. So it appears something isn't getting from the data board to the power board. I even rang out the ribbon cable between the two boards. I found replacement data boards from American DJ for $36 but I'm not ready to go that route just yet....especially since I just got a used pack on eBay today for $50.

 

[JD]

Basically, the output of the processor feeds through the channel LED and into the LED half of the opto-isolator. The other side of the opto is a photo-triac which feeds through a couple of resistors to the gate of the triac.
So, if the channel LED is on, then the LED inside the opto is on, and the photo-triac is in conduction allowing ac to feed through a dropping resistor to the gate on the power triac, which should turn that on as well.

Something in this simple loop is open, and it is open for all four channels. I have no first hand knowledge that would tell me for sure that you would be safe making these tests.

If the unit was on my bench, I would jump MT1 and MT2 on one of the triacs and make sure that channel came on. If it did, I would remove that jumper, and next jump the output of the photo-triac and verify the channel would still come on. These two tests would clear the output circuit and the post-opto firing circuit. I would next see what the voltage was across the photo LED in the opto when the channel was full on. If higher then 2vdc or lower then .5vdc, then the LED in the opto is bad. If the voltage was about 1.5vdc and the output was still not firing, then the photo-triac in the opto was bad.

There are a lot of variations on the specifics of the firing circuit, and it even has varied across production runs. Because of that, and the fact that I am going over these on a chat room, with no visual conformation on anything, I really feel this is not the best way to handle this.

Another alternative would be to get some good pictures of the boards inside the unit and post them up here. Might be something that jumps out at one of us old-timers.

 

[danhr]

I did the M1-M2 jumper test last night with a load and got output so the HV wiring is OK. As I mentioned before I have built dimmers using solo diacs to fire the triacs but not optos so I appreciate the test procedures for optos. (I need to find a schematic for these...although they're not that different from the ones I built) I will check the readings tonight and see....I'll see also if I can post some pics. And if I get this working again I'll use it as a test subject for a complete 25A triac upgrade....

 

[JD]

Well, just a few seconds of theory behind design- (This coves most theatrical dimmers.)
Unlike the analog phase shift dimmers that use a diac, these designs use a circuit called a "Zero Voltage Cross Detector." When the AC waveform crosses 0 volts, the detector starts a countdown in the processor. Each half wave is basically divided into a 255 count. The processor compares the count to the DMX value. When the numbers are the same, it turns on the gate signal to the opto, therefore turning on the triac. The higher the DMX number, the earlier in the waveform the triac is turned on and the brighter the light is. Now, there is a lot of variation in more expensive dimmers, such as programs that anticipate and alter the count so that there is a more natural curve to the lamps brightness, but the basics are the same.
Because shoebox dimmers are built to be inexpensive to manufacture, a common shortcut is to put a "channel indicator" LED in series with the output of the processor. It too will show relative brightness depending on the length of the gating signal.
Since the LEDs respond to the dimmer controls, it actual gives us a good diagnostic point. If the cross detector or processor was not working correctly, the LED would not respond correctly.
Getting actual schematics can be a real problem, but I hope this info helps.

 

[danhr]

I jumped the opto outputs and all the triacs outputted ok. I don't get a voltage reading on any of the opto inputs so unless all 4 optos are dead I think there may be a dead transistor on the data board that feeds the common to the optos. I will try swapping out the data board when the new pack comes in (I can't spare any others) and see if that's it. Thanks for all your advice- I now have a better grasp of how these work. I think we may have gone as far as possible with this without violating forum rules and overstretching the limits of my bench repair skills, especially since a lot of these components are surface mount.

 

[danhr]

OK last post on this. I swapped out the data board from another pack and the dead pack works fine. Soooooo....it apparently was the transistor which feeds the common to the optos on the other board. Too small for my fat fingers to replace (if I could get a replacement) so it looks like I'm ordering a new data board......

 

[Dscheuerman]

Hi cb, first time poster. Lighting noob. I use an Obey 10 to control a Chauvet dmx4 dimmer pack. Last night after sound check, all the lights mysteriously came on full blast, and would not respond to the faders on the obey 10. Nothing changed I'm aware of. Checked manual for dmx4 and nothing seems to be set up wrong. Any thoughts?

Edit: forgive the autocorrect error. Typing with thumbs.

 

[AxlD1234]

My first thought is Did somehow the lights get parked on Accident?

The only other time i have seen that happen is if your main DMX Line coming out of your board was interrupted somehow, and the lights do not have DMX Signal going to them

 

[Dscheuerman]

My first thought is Did somehow the lights get parked on Accident?

The only other time i have seen that happen is if your main DMX Line coming out of your board was interrupted somehow, and the lights do not have DMX Signal going to them

Thanks for replying. I'm not familiar with the term parked, but the lights powered by the dmx4 are shop lights, which have only on and off settings. They are controlled by chases on the obey 10. Incidentally the obey 10 is working fine otherwise to control our pars.

 

[AxlD1234]

The Term Parked means That the lights are "frozen" to whatever Level you put them on So even if you cleared the board the lights would stay at whatever it is at. But i just looked at the controller that you are using and it isn't possible to Park dimmers on that controller.

 

[JD]

Time for some diagnostics:
1) Unplug the DMX cable. Are the lights still all on?
2) Change some of the pack settings, do they all still stay on?
If both answers are "yes" then somehow I think you popped all the triacs in the pack. DMX 4 is good to about 600 watts per channel. What were they loaded at? (How many watts.)
If the lights go out when you unplug the DMX cable, then somehow the board is telling the pack to keep all the channels on.

 

[Dscheuerman]

Thanks for replying. The answer to both questions was yes. We've blown triacs before by overloading, but the channel just stopped working. Are you suggesting that if all four triacs break, it defaults to always-on? I will go ahead and try to find some replacement triacs. Any physical store that might sell them?

Ps there's only 500 watts running off each channel. Shouldn't have blown... Oh well.

 

[Les]

We've blown triacs before by overloading, but the channel just stopped working. Are you suggesting that if all four triacs break, it defaults to always-on?

It can go either way - just luck of the draw. Is there any way you could connect the pack to another controller to confirm the issue?