Jump to content

Easiest Terminators/Snubbers to make


chuckd

Recommended Posts

  • Replies 265
  • Created
  • Last Reply

Ditto!!!!!!!!!!!!!!!!!

To the moderators:

Just a thought but you may want to sticky this thread. The explanation here is the best I've seen. I have built snubbers and use them in my display, but never had an explanation as good as what was provided here. Just a thought.

Link to post
Share on other sites

First off, I believe that resistances much lower would work as well, if not better (faster), but at the expense of drawing more power. For example, you 'could' use a 1K resistor, but it would have to dissipate 14.4 watts of power! That would be a big sucker for sure. A 1K 1Watt resistor would probably flame out, which is quite spectacular if you've never seen a resistor do that before!

The capacitive voltage from our LED's is 'drained' over what is known as an RC time constant. That is, it is directly proportional to the R (resistance) and C (capacitance) values of our circuit. We of course have no clue how much capacitance our LED strings have (and indeed it is different for everybody). However, you can be sure that with higher value resistors, it takes longer to bleed off the voltage. If you used a 1Meg resistor, for example, I doubt it would do much good, since it would take relatively forever to bleed the voltage out. A 1 Meg resistor would take 20 times longer than our fictional 50K resistor to do the job.

Cheap (5% tolerance) resistors come in common values, like 33K, 47K, 4.7K, 470, etc. They're cheap that way. A 33K resistor draws .43 watts, so it gets a bit closer to our 'double the value for safety' that I mentioned earlier. Our 47K resistor draws about .31 watts. The fact is, 30K or 50K, if they were available, would probably work (although I'd bet 30K works better than 50K, since it would be faster).

I really doubt that whoever came up with 33K or 47K did a whole lot of science, but rather found a good value that would be small (1 watt or less), and do the job effectively. Sorry to say, but that's how a lot of engineering happens!

Link to post
Share on other sites

Seems to me that what I'm seeing in the scope traces, especially in the 24 strand trace, IS in fact the inability of the Triac to turn off, due to the amount of capacitance.

Normally, a traditional 'snubber' circuit for triacs is for big inductive loads (like relays). Relays are basically big electromagnets, and when you remove power from one of those, they kick all of their magnetic field back into the circuit. This creates a big current across your triac, and prevents it from turning itself off for awhile.

We, however, have big a big capacitive load with our LED's. You can kind of think of this as opposite of an inductive load, but works similarly. It might be helpful to think of it as a tiny and very fast rechargeable battery.

I see it happening this way:

1. If you string enough LED's together, you effectively get a huge capacitive load.

2. When the triac is asked to turn off (by the LOR microcontroller built into the 1602 box), it has to get to a point where there is very little to no current flowing in order to properly turn off.

3. When the AC is removed, our LED's turn their capacitive energy back into current, and this flows through the circuit back to our triacs.

4. If this doesn't go away sufficiently before the triac is asked to be turned on again, it will never actually 'turn off'.

You can see in the scope traces that the triacs are being asked to turn off at around 40V. When enough capacitance exists, enough current is generated by the LED's to keep the triac on. This is why in the 24 strand example, the upper half of the AC wave is fully intact. The lower cycle 'turn off' didn't happen due to the LED's generating enough current to keep the triac alive for that half cycle.

Another nasty thing occurs when adjacent channels are off or are on at low intensity. This kick-back current can (and does) flow through your box into the other channels. This is why you can see all your channels on a box flash even though you only dimmed one channel.

Link to post
Share on other sites

3. When the AC is removed, our LED's turn their capacitive energy back into current, and this flows through the circuit back to our triacs.
Are you sure that current is actualy flowing through the TRIAC since you're only measuring voltage with your scope? I see the TRIAC, in this case, akin to the series diode in a simple power supply with only a small amount of capacitance on the output. Current doesn't flow back through the diode when the input voltage drops to zero.

Also, are you sure that the capacitance is from the LED's and not from the massive amount of twisted wire? Wouldn't you also get similar results with a power cord that was tightly twisted and of similar length?

Link to post
Share on other sites

Well, no, I'm not sure about that. My days as a design engineer were spent with microprocessors and audio circuitry. Didn't exactly spend too much time with triac/power circuits, so it's slightly new to me.

I looked at what the traditional snubber circuit did for triacs, and it's mainly to kill off inductive kickback from something like relays or solenoids. The large spike that comes back can saturate the gate of the triac, creating a false misfire. Since my 24 strand example shows a complete upper half AC waveform, it 'seems' like a false misfire. Since inductive loads can do this, I'm assuming a large capacitive load could do the same in our case.

I have some twisted cord that I used for my LED strobes (500 feet). So I slapped on a plug and a socket on each end this morning, and put one LED strand at the end of the string. The waveform appears nearly identical to my one strand case, so I don't think it's the cord causing any problems here. As for circuitry in the LED 'warts'? I don't know, but I'm fairly certain my strings have no capacitors.

Link to post
Share on other sites

Ernie, so obviously the triac needs a zero current condition to turn off. Since that waveform is so 'clean' in the upper half, I'm assuming that the triac had to be falsely triggered. Doesn't the triac gate have to be pulsed in the opposite direction to get the 'other half' of the cycle, unless our light strand is providing a holding current somehow?

Do you think those rounded edges are an additive effect from the strands? I don't think the triac is off at that point, rather the AC voltage has dipped below the level sustained in the LED's, so they're dumping back into the circuit.

I think it's reasonable to assume that the LOR controller isn't being 'fooled' into triggering this way with noise, as it's simply too clean a waveform.

For simplicity, we could say 'if one terminator doesn't work, slap on another' (up to a point). Two 47K terminators would make a 23.5K terminator with 2 watt capacity (since they are in parallel). Four would make a 11.75K, 4 watt terminator. That's alot easier than building different value terminators.

Link to post
Share on other sites

FYI.

I only had 5 channels acting up during fades and after searching around the house, I found 5 "Plug-Ins" laying around.

I plugged them in somewhere along the string for each channel

and....

VIOLA!! Problems solved. Fades and flicker GONE!! :eek:

I can definitely see where if you had several 100 channels where this was an issue, then the solution outlined in this post is EXCELLENT.

I may add a snubber on every channel next year, whether it is a problem or not.

Link to post
Share on other sites

I've noticed that terminators help all my LED fading situations, even those that don't 'flicker'. You'll notice the fade being a lot more linear. You're probably used to the situation similar to 'anything between 60% and 100% looks the same, then it drops off like crazy'. Well, with a terminator, you get a lot more variance with your fade.

At least with my LED's, it's quite noticeable. I have two 70 foot twin mega trees. On one song, they do exactly the same thing, with many dramatic fades. I put terminators on one tree, and none on the other. The difference was huge, where one had these gorgeous, graceful fades and the other just kind of dimmed then turned off. Each channel on those trees only has 4 strands end-to-end, so it wasn't a worse case scenario.

I'm of the opinion that you should probably terminate every LED channel as well.

Link to post
Share on other sites

I have to say, this is one of the best threads I've read in a while! I have 26 strands of 50ct. LEDs on our front trees. I put them all together on one X-10 module (a Lamp module) and they didn't turn off all the way! They were very noticeable when "off". I proceeded to fade them through LOR's Hardware Utility, and they never were clean and didn't turn off all the way. So, I made a c9 bulb and put that on the end of them, and wala!! It worked!

I never really understood this...but I new we needed some resistance. I had the same issue this past summer in my bathroom with a strand of LEDs on a X-10 unit.

Now it totally makes sense! I will keep this in mind as I start computerizing more LEDs. Great info, Chuck! Thanks a ton!

Link to post
Share on other sites

I never really understood this...but I new we needed some resistance. I had the same issue this past summer in my bathroom with a strand of LEDs on a X-10 unit.

I am not going to even try to imagine why you have x10 controlled LED lights in your bathroom... :P

But I DO think that it would be a perfect application for the "Plug in" fix! ;)

Link to post
Share on other sites

FYI.

I only had 5 channels acting up during fades and after searching around the house, I found 5 "Plug-Ins" laying around.

I plugged them in somewhere along the string for each channel

and....

VIOLA!! Problems solved. Fades and flicker GONE!! :eek:

I can definitely see where if you had several 100 channels where this was an issue, then the solution outlined in this post is EXCELLENT.

I may add a snubber on every channel next year, whether it is a problem or not.

Just a little history... the Glade "Plug-In" fix came out of a post by the "Crazylightlady" (Annalisa) looking for an alternative to the resistor "snubbers" she was making for LED strands about 5 years ago... the little light bulb went on over my head and I grabbed one of the little warmers and checked the back after thinking about the heat the resistor was going to produce on the homemade snubber and sure enough it was a good fit for the application.

I posted it up and she said I had won the k.i.s.s. award of the month! :D

It would be nice to have these snubbers molded up offshore in bulk for next year...

(the little light just went on again....) ;)

Link to post
Share on other sites

What do the scope pictures show? Is this the voltage across the LED string(s)?

Have you ever tried to see what current is flowing at various moments in time? That is probably more important than the voltage, since the triac turn-off depends the current flowing through them.

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...