Calling out to Weezard for LED advice

[Weezard]

Don't think it's static. Hot and humid here, all year long. (That's why I need the LEDs.) Dirty power's probably at least as likely, and I don't think that's it either. The problems occur when I hook up multiple LEDs. And they never seem to "blow" while they're operating.

When I got started, I deliberated toasted a whole grab bag of Radio Shack cheapies just so I'd get a feel for how it happens. This is strange. They seem to be working okay, and when I go to start 'em up again, nothing happens. On exam, they're cooked. I 4 x color-coded the wires so I'd be able to easily tell 'em apart--pretty certain I didn't reverse-bias 'em, too.

Maybe I "milliseconded" them by accident. In that case do they keep working but refuse to fire up next time?

Let's think about this.

If you apply input power to a constant current regulator with no load.
It will "see" almost no resistance at it's output.
Whatever your pre-set current is, the supply will raise it's voltage as high as it can to try to hit that preset.
So a supply with no load will sit at whatever voltage you have preset.
That means you must have a load connected when you turn on the supply.
If you power-up the supply and then try to connect or disconnect a load, the supply may get confused and voltage spike your load.
It's a very short spike, but more than enough to punch a hole in a led junction.
When testing I start with everything tightly connected, use variable voltage setting and bring it up slowly until I hit my target current. Then record the voltage reading.
I turn off the supply and wait a few seconds for the supply to drain before disconnecting anything.
Have not fried a big led yet.
Did eat a couple 3 Watt mpcb when I was learning, though.

Keep in mind, that it's before 11 am my time. So take anything I write before noon with a boxcar of salt.

Not a morning person.

Breakfast is up, I'll be back later.

Weeze

[DreadedHermie]

Thanks for the continued support. The LED's may have won the first round (a Pyrrhic victory, for sure) but I am not giving up (yet). Taking the mutts for a walk to clear my head, and prolly reorder parts tonight.

Your analyses of the potential causes for the problems seem valid to me. All I know of operating this Mastech is the little pamphlet it shipped with. Doesn't cover much; includes a couple of recipes for LED disaster, apparently. I definitely was disconnecting the LEDs abruptly and this possibly resulted in some of my catastrophies, since the LEDs appeared to be working fine when I disconnected them. I think I was subconsciously (rather, consciously and ignorantly) trying to avoid a turn-off spike from the PS and in the process spiked them even more effectively.

Anything you can think of to help keep the next batch of LEDs alive will be appreciated.

[denialisback]

No worries, brah. I thought it was droll.

And your avatar is great!
Reminded me of my first stab at being the electrician, (or somebody like him), at about that age..
('cept I used a small, stainless steel, turbine blade that I held with a boxing glove.)
Jammed it in a broken socket 'cause I wanted to "magnetize" it.
Parental units were not amused.

So, it's time to outshine the Procyon.
Been saving my pennies and it's need to build some more Ledengin 15W arrays.

Been playing with an idea that involves 3/4" soft copper water pipe.
Want to coil it in a flat spiral, squish the spiral disk almost flat, sand and polish one side and attach the LEDs with thermal cement.
Then, if I hang it from the center I end up with a "cone-o-light"
Pumping water through the pipe and back into a res. will cool the leds and keep res. temps up in the winter for mainland folks.
I'll probably have to feed it from the tap and tie the output into my drip irrigation for the yard.
That will transport the heat out of the grow room.
Still a pipe dream, until I get to the plumbing supply store.

Howz your design coming?

Aloha,
Weezard

This is a very good idea, congratulations Water cooled led set, I was thinking of doing the same with my HID HPS 400w grolux but I decided to go CFL, possibly 20x50W or something heh.. who knows i might need this for the 50 20watters heh.

Peace,
Denial

[Weezard]

Mahalo!

Weeze

[Weezard]

Thanks for the continued support. The LED's may have won the first round (a Pyrrhic victory, for sure) but I am not giving up (yet). Taking the mutts for a walk to clear my head, and prolly reorder parts tonight.

Your analyses of the potential causes for the problems seem valid to me. All I know of operating this Mastech is the little pamphlet it shipped with. Doesn't cover much; includes a couple of recipes for LED disaster, apparently. I definitely was disconnecting the LEDs abruptly and this possibly resulted in some of my catastrophies, since the LEDs appeared to be working fine when I disconnected them. I think I was subconsciously (rather, consciously and ignorantly) trying to avoid a turn-off spike from the PS and in the process spiked them even more effectively.

Good analysis!

Disconnecting an alligator clip sounds "clean" until you see it on a 'scope.
Literally, thousands of "off n onnies" in a fraction of a second. Looks like a burst of rf.
If there's any inductance at all in the circuit, at these currents, the spikes will be higher than the dc level and will be in the wrong direction!
Once you get over a few amps, coiling a lead for strain relief is enough inductance to make trouble.
Then there are small, distributed capacitances...

Anything you can think of to help keep the next batch of LEDs alive will be appreciated.

So, next time, hard wire the leds and use the power switch on the supply to power down and we should be OK.
I've been running 13. 5 Watt leds from my Mastech for months with the supply on a timer.
Keep it set at 12.6V 3.4A.
So I bang the AC on and off with few problems.
(it did occasionally cause my UPS to "freak out" and beeeep at me.)

Had to plug the timer into a plain old socket and eschew the UPS for those leds because when it siezed-up like that it shut down the airpumps.
Not acceptable!

I feel your pain, brah.
Fried my first home-made geiger counter in the '60s.
Almost cried when the magic smoke got out.

Aloha,
Weezard.

[DreadedHermie]

LED wives and mothers are waving tearfully goodbye to their loved ones: the unfortunate few who have been selected to do glorious battle with the now-infamous Dreaded Hermie. No one speaks of it, but they all know: we might not survive this....

[attachment=o213864]

[DreadedHermie]

Disconnecting an alligator clip sounds "clean" until you see it on a 'scope.
Literally, thousands of "off n onnies" in a fraction of a second. Looks like a burst of rf.
If there's any inductance at all in the circuit, at these currents, the spikes will be higher than the dc level and will be in the wrong direction!
Once you get over a few amps, coiling a lead for strain relief is enough inductance to make trouble.
Then there are small, distributed capacitances...

So, next time, hard wire the leds and use the power switch on the supply to power down and we should be OK.
I've been running 13. 5 Watt leds from my Mastech for months with the supply on a timer.
Keep it set at 12.6V 3.4A.
So I bang the AC on and off with few problems.
(it did occasionally cause my UPS to "freak out" and beeeep at me.)

Weezard, thanks. These details/specifics are what I'm needing. I reordered my supplies and should be back blowing stuff up by the end of the week, although responsibilities in the "real" world may cut into my mad scientist time over the coming weeks. Getting older ain't so bad, but having to act like an actual adult does get tiresome.

I started looking at thermal switches. Most that are referred to that way on the 'Bay open the circuit at a given (rising) temp. The ones that close the circuit seem to be called "thermostats" or fan switches. From your brief description of operation I'm gonna assume the latter is what you're using.

Looking at a pic of your light, looks like you have 3 of them, all mounted in places you'd expect the sink to be the hottest--right near the leds. It also looks like you've got it wired so any one of the switches can complete the circuit by itself. Close? :detective1:

I found some switches that close when temps rise to 50 deg. C. Looks like the right range for me, although certainly I'll hard wire both fans at first.

[attachment=o213907]

DH

[DreadedHermie]

[attachment=o213910]

The attachment's just the pdf datasheet from Mouser. Couldn't get it to work, trying again. :wtf:

[Weezard]

[attachment=o213910]

The attachment's just the pdf datasheet from Mouser. Couldn't get it to work, trying again. :wtf:

STC 120 looks about right.

And you are spot on about;

"It also looks like you've got it wired so any one of the switches can complete the circuit by itself. Close? :detective1: "

Give that man a ceegar!:jointsmile:

Duty cycle depends on ambient.
Generally, the fan runs for 30 minutes and then shuts off for 10 minutes.
Found 'em cheap, so I bought a dozen.
Figure I can use them to ride herd on my peltier devices.(Long story.)

Aloha,
Weeze.

[DreadedHermie]

So I looked at 'em. Looks like a miniature heat pump. Great idea, if you have a place to exhaust the heat. The ones I saw drew like 2A, though. Is this the kinda thing you're talking about?