The plot sickens......

[chuckd]

Anybody know what is driving the CP (clock pulse) line of the 273's? That signal is jittery as heck, during a constant 30% intensity. Since this drives all the triac controllers, that would definitely cause the problem I'm seeing.

My suspicion is that this is driven by the output of the AC line zero crossing detector, but I don't know.

[Jeff Sand]

Ah, I wasn't realizing that this was happening on all 16 channels simultaneously. In that case, I agree we are looking at either a controller issue or a zero-crossing detection issue. Can you tell what exactly the 273 is being used for? I don't have a board in front of me right now, but I've got a couple V6 1602W's in my parents' basement that I can go grab tomorrow and try to take a look at. Having a jittery clock signal isn't necessarily a sign that something is wrong; the 273 is just an 8-bit register, and at this point we don't know what its purpose is. Question: If you set your channels at constant level (30%) is the CP frequency staying constant, or does it look like it's all over the place?

[chuckd]

From what traces I can see, all the Q outputs of the 273 are tied to the anodes of the triac controllers (as sources, instead of sinks I guess?). The D inputs seem to tie directly to the PIC. The D inputs seem OK, but the Q outputs don't. Thus the CP signal weirdness. There is a resistor connected to the CP signal, this implies to me there's some analog circuitry involved here.

I could visualize a design where the 273 is used to not only drive the anodes, but the CP signal derived from the AC waveform would synchronize everything with the AC line. Just a guess, though (no schematic....

[Jeff Sand]

Huh, you'd think after working with this stuff for 6 years I would have looked more closely at the output stage. They're connecting the anode to a logic gate? I'm gonna have to take a closer look at this before I dig myself any deeper and say something dumb.

I looked up the thread and couldn't find it, is this a D or a PC board?

[chuckd]

From what I can tell the Q outputs go through 470 ohm series resistors, straight to the anodes. The cathodes are tied to ground. Not what I'm used to seeing either.

The board is a CTB16D V7

I'm figuring the crossing detection is done by the PIC via a simple current limiting resistor, but I'm not sure yet,

[LightORamaDan]

We are working on this but have not yet replicated the problem that is being seen. We are testing with 100s of LEDs with/without terminators, regular incandescent loads, etc.. and it is all working but we are continuing to work on it!

Before this thread I had heard something from someone in a 50hz country that was having a fading problem with LEDs with V7... Are any of these examples in 50hz countries?

Dan

[Jerry Budelman]

Hi folks,

I'm new to participating in this forum, but I have some similar problems and I'm a retired electrical engineer as well. One of my (4) controllers exhibits occasional random flashing during playback, but otherwise responds properly to its programming. All 4 controllers (1602W showtime units, blue board) were recently upgraded to 4.32 firmware. S2.6 software.

Most of my lights are LEDS and I use a C7 bulb as a terminator for all LED strings. I had no luck using any reasonable value resistors since they don't generally provide enough holding current for the triacs to latch especially at low dimming ratios.

Looking at the triac waveforms posted earlier, it seems as if the current through zero crossing never falls below the valley point where the triac resets to a non-conducting condition. This can happen if there is an inductive load somewhere. The inductance might even be part of the primary power circuit (any transformers in the line??).

Alternately, it could be a firmware problem. I had no issue last year with the same controller running downrev firmware, but there were no LEDS in the array that time.

Any ideas?



Jerry Budelman

[chuckd]

Oklahoma is kind of behind the times, but we're not exactly a 50 Hz country!

My unit does this with LED, incandescent, soldering iron, blender, whatever. LED withOUT terminator is actually the least blinky. Those nice flat transitions make the problem worse for some reason.

Dan, want any scope traces from around the board anywhere in particular?

[Jerry Budelman]

Good news...I figured out the problem that was causing the lights to blink. One of the channels in the controller was used to activate a small motor that made the wings on an angel move. The channel was simply turned on (full) at the beginning of the show. However, the presence of this motor caused the entire controller to get flakey. I suspect it's either some spikes created by the motor, or it's inductive nature. Putting a 10 watt incandescent lamp in parallel with the channel cleared it up (as did disconnecting the motor).



Folks might want to be aware of this potential issue. An inductive load on any channel seems to mess up the entire controller if the power factor is too low.



Jerry Budelman

[chuckd]

I wish mine were that simple. Just connecting one resistive load to one channel and we're blink crazy. I would send the unit back (it's a brand new unit), but from what I hear, the turnaround this time of year is many weeks. It'd be better if I could fix it myself.

[cmoore60]

Since LOR cannot seem to replicate. You might suggest that they ship you a unit that is not of this revision and you ship your problem child to them for further evaluation.

Just kind of thinking out loud if you know what I mean.

Chuck

[Jerry Budelman]

Hi Dan & others,

Here's another random tidbit from my show preparations. As I was waiting for the scheduled test show to start, I noticed random flashing of many of the channels. The flashing continued during the show; most noticible when a particular section of lights were not active. This behavior was not evident when the yacht was in the boathouse, so I went looking for things that were different. It turned out to be a small dimmer that I had used for a few LED/mini lights in the boat pilothouse. The dimmer was on the same primary circuit as the controller that was acting up. Moreover, the problem only manifested itself at a very particular setting of the dimmer; no problem at full-on or off, or most other settings. It looked like some kind of feedback effect. I put a line filter (Corcom) in front of the dimmer/LED string with no effect.

This seems to be another indication that the trigger or zero-crossing circuit is very sensitive to line noise. It would seem a narrow bandwidth Phase Locked Loop (PLL) filter would do wonders for this problem as well as the difficulty with running LOR on a generator.

Dan- if you would provide some schematics, I would be happy to analyze them and provide some suggestions.



Jerry Budelman

[chuckd]

I too get the random flashing when there's nothing going on at all. From what I can guess from poking around the circuit, I think there's simply a current limiting resistor out of one of the transformer taps, going directly into the PIC. Since the PIC has clamp diodes, this is commonly done for cost and simplicity issues. However, it does make zero crossing 'software dependent', and a good software filter is a must.

This to me would only make the unit be 'off' a bit from where you expect the sequence to be. I can't figure how this would cause random flashing when no sequence is being run.

Good luck on the schematics. I tried that previously to no avail. I'd love to have them, as it's the only way I'm going to get my unit repaired this season.


I'd like to see the circuit become something like a '30V crossing circuit', rather than a zero-crossing detector. Eliminates a lot of problems that way.


Chuck

[Jerry Budelman]

Chuck et al,

I spend some time on the phone with one of the fellows from LOR to discuss the symptoms I was having (random flashing). I'm becoming convinced that my issue at least is the result of noise on the power line which creates a dV/dT triac turn on event. Triacs (and SCRs to a lesser degree) are susceptible to turning on if a sufficiently large and/or fast voltage spike occurs across the device, independent of a gate trigger. My display is on a boat in a marina where there are many other boats lighted and some have sequenced lights (not LOR but basic off the shelf stuff). The LOR fellow also said CFLs and ESPECIALLY neon signs drive the box bonkers. He has had good luck with noise suppressors designed for the X10 network stuff. The units are XPPF-X10 filters and are availbale on ebay for about $5 each. He uses them to supress noise coming FROM an offending source. He said he hasn't used them in front of the LOR controller as yet. I've ordered some and will see if they have any effect. Don't know what is in the filter, but I might sacrifice one to find out .

I asked if LOR had considered using snubbers on the triacs, and he said they looked at it, but they caused more problems than they solved, so they went with "snubberless" triacs.

My (apparent) near-term solution was to insert a large isolation transformer (3 KW)into the shore power cord before connecting to the marina electrical system. When I tested the system last night, all was perfect. Maybe I got lucky, or the stars were aligned...but it seems as if the transformer is blocking some or all of the spiky stuff. It also isolates the grounds, so no ground loops which may also have been contributing to the issue.

You may want to look to see if there are any noise issues causing your problem.

Regards,

Jerry