DIY LEDs - The write-up

[staggeringwade]

I'm about to flip out!
I have the leds screwed down to the heat sinks with metal screws and have rubber washers under the screw heads to prevent shorts. The stars came pre-soldered, so I did a good job soldering the wires on-not sloppy blobs. I have yet to fire up the mean well Eln6048D because I'm afraid to let out the magic smoke. (can't afford too).
I have been testing and testing with the multimeter trying to understand If I have any shorts. I have had a couple of good answers in regards to ohms and resistance, but still lost.
Question #1 -If i take the red lead of the meter and place it on the + of the led star, do I place black lead of the meter to the heat sink or - of the star?

Question #2- If there is resistance meaning any reading other than OL on the meter I have a short somewhere? oo.o or high numbers are all bad? Some Leds are O.L. some are 0 and some are very high numbers- there is no consistency.

Question #3 -If there is a short, It does not necessarily mean there is a short at that tested led, it could be anywhere in the string?

Question #4- I have 12 Cree Q5's in series, and the svr 2 potentiometer turned all the way down in the mean well driver and I decide to plug in the meanwell, if there is a short is there a good likelihood that I won't damage the LEDS or will I destroy the leds?

Question #5 -If i do the two AAA battery test for every led, will that show anything since all leds are in a string now? Will the AAA batteries illuminate the led if there is a short?

Guys I really am sorry for all of these questions. Like I said before I have used multimeters all my life and have worked with electronics as long, but this is truly killing me! I obviously don't know jack about meters.
Thanks
Wade

 

[TheFishMan65]

1) From any pad one te star to the heatink shoud have a very high resistance. The starts I used (and I htink most) have two - and two +. Use the pad without the wire start on one start with the - unused pad it should have a very low resistance to the + unused pad on the start the wire is connected to. I hope this makes sense. IF there is a wire lowresistance no wire high resistance.

2) HIgh and low are both good it just depends see answer 1.

3) The only short too worry about (Ihtink) is to the heat sink. So if you get one it has to be one of the 4 pads listed in 1.

4) Chances are ou will be fine. But I can't guess at all possible failures.

5) The will still illuminate the LED. The onl ime they wouldn't would be if the + and - of the LED under test are both connected to the heat sink. Then the current will go through the heat sink rather than the LED.

Hope this help. I am sure someone else will post some more good ideas.

 

[der_wille_zur_macht]

Wade, keep in mind that the basic goal of testing with a multimeter is to determine if there is a conductive path where there shouldn't be. So, think through the entire rig.

The electricity comes from the driver. It'll flow through anything conductive attached to the leads coming out of the driver. There are things that you WANT in the circuit (your wiring, your LEDs, etc.). There are also things you DON'T WANT in the circuit (the heatsink, the screws, mounting brackets, anything else metal).

So, when you probe with the multimeter, your basic goal is twofold:

1) Make sure that all the things you want in the circuit are actually in the circuit.
2) Make sure that the things you don't want in the circuit are not in the circuit.

So, start by putting the multimeter in conductivity mode. Or, put it in the lowest resistance mode if it does not have an explicit conductivity mode. In conductivity mode, most meters will beep when the probes are touched to something conductive. In resistance mode, the meter will display the resistance when the probes are touched to things that are conductive. In either case, the meter will show some default value (like "OL") if there is no conductive path between the probes.

So, put one probe on something metal in your build. Pick anything. Now, put the other probe on something else metal in your build. Again, for now, just pick anything. If the meter beeps or shows a reading, then there is an electrically conductive path between the two things you touched. If it does NOT beep or show a reading, then there is no path between the probes.

Now you should know how to test for the two conditions above. Most important is testing for #2. To start, touch one probe to something you DO want in the circuit - for instance, either pad on any of the LEDs, or any exposed wiring or solder joints. Now, touch the other probe to something you DON'T want in the circuit - i.e. the heatsink. If the probe beeps or shows a reading, you have a problem. If it does nothing, you're OK. Keep the one probe on the heatsink, and go around the array with the other probe, touching it to a + and a - pad on each star, in turn. If it ever beeps, you know you have a short from that LED to the heatsink.

Your multimeter may or may not show conductivity through the LEDs themselves depending on the polarity of the probes, but don't worry about it - probing the entire fixture should take only a minute and will confirm that there are no shorts anywhere.

Question #4- I have 12 Cree Q5's in series, and the svr 2 potentiometer turned all the way down in the mean well driver and I decide to plug in the meanwell, if there is a short is there a good likelihood that I won't damage the LEDS or will I destroy the leds?

If there is a single place where the circuit touches the heatsink or otherwise conducts to something that it shouldn't, you will likely not damage the LEDs but the fixture will be unsafe to operate (as you may inadvertently complete the short and hurt yourself or damage some piece of equipment). If there are multiple errors, you may indeed damage LEDs - for instance, if the first LED in the series shorted to the heatsink, and the last one did too, the driver would essentially just be driving those two LEDs (or maybe just one of them, depending on where the short was exactly). But this is pretty unlikely.

Question #5 -If i do the two AAA battery test for every led, will that show anything since all leds are in a string now? Will the AAA batteries illuminate the led if there is a short?

The battery test really just confirms that an individual LED is not fried. It really does not help diagnose your wiring.

Guys I really am sorry for all of these questions. Like I said before I have used multimeters all my life and have worked with electronics as long, but this is truly killing me! I obviously don't know jack about meters.
Thanks
Wade

Don't be sorry! We all started somewhere, and these forums exist to share info and answer questions!

 

[der_wille_zur_macht]

1) From any pad one te star to the heatink shoud have a very high resistance.

I would further qualify this statement by saying that the resistance between any pad on a star and the heatsink should be immesurably high. In other words, your multimeter shouldn't even register a value between an LED's solder pad and the heatsink - it'll show "OL" or whatever the default is.

 

[staggeringwade]

Guys- that was most informative!
After reading this I feel that I was worried over nothing, but at least I understand how to use the meter better now.
I the past I had the meter set to the conductivity mode where it beeps and I found no shorts, but the ohms / resistance setting is what confused me and had me very worried.
I will try and test things tonight and see what happens!
Thanks.

 

[der_wille_zur_macht]

Keep in mind that most people don't even bother testing an array like that, at least unless they find a problem when they power it up.

 

[pheinzig]

pheinzig; You should note that the human eye is lousy with respect to brightness judgments. So statements like "nearly as bright" are crazy subjective. For a photosynthetic creature it can make a vast difference. I'm afraid your test while interesting has tangled up color with photon flux. So one can't make any kind of statement about either.

You would need to have twice as many whites, essentially to replace the missing blues, to allow a statement about the blue verse white. And as CJO points out best would be a PAR meter that allows you to readjust to the same PAR in white as you had in the mixed scenario.

And knowing is half the battle. Thx for the lesson :thumbsup:

 

[staggeringwade]

Guys- that was most informative!
After reading this I feel that I was worried over nothing, but at least I understand how to use the meter better now.
I the past I had the meter set to the conductivity mode where it beeps and I found no shorts, but the ohms / resistance setting is what confused me and had me very worried.
I will try and test things tonight and see what happens!
Thanks.

I fired the leds up and everything is fine.:dance:
The rubber washers are trouble though. :mad2:The tighter i screwed the screws in, the more the washers would squish and cause shorts. I new better than to use them, but I live in area where there are no nylon washers in bulk. So I'm going to look at the web and see if I can't hunt some down and replace all the rubber washers.
I'm afraid heat will cause the washers to move and start shorts. Besides I can't get the optics on without trimming the washers, which will cause a short.
I tried to take a picture with the lights on, but it was so bright the camera couldn't take a good picture!
Thanks for your help everybody!

 

[jusmee]

I fired the leds up and everything is fine.:dance:
The rubber washers are trouble though. :mad2:The tighter i screwed the screws in, the more the washers would squish and cause shorts. I new better than to use them, but I live in area where there are no nylon washers in bulk. So I'm going to look at the web and see if I can't hunt some down and replace all the rubber washers.
I'm afraid heat will cause the washers to move and start shorts. Besides I can't get the optics on without trimming the washers, which will cause a short.
I tried to take a picture with the lights on, but it was so bright the camera couldn't take a good picture!
Thanks for your help everybody!

Maybe fibre washers would be a good idea?

 

[jusmee]

I would further qualify this statement by saying that the resistance between any pad on a star and the heatsink should be immesurably high. In other words, your multimeter shouldn't even register a value between an LED's solder pad and the heatsink - it'll show "OL" or whatever the default is.

Beware of fingers. If you are touching the probes, it will read your resistance and make you think you have a problem. Modern multimeters will only show OL or something like that if you keep the fingers off the metal

 

[jusmee]

If there is a single place where the circuit touches the heatsink or otherwise conducts to something that it shouldn't, you will likely not damage the LEDs but the fixture will be unsafe to operate (as you may inadvertently complete the short and hurt yourself or damage some piece of equipment).

I have one array that has one such short. It is internal to one of the star led MCPCBs. I will replace it eventually, but as it is powered from a fully floating power supply (no ground reference) it can't cause a problem.

 

[jusmee]

I did an experiment for 1 week:
Had all my RB off for 7 days(originally I had a 50/50 split XRE white/XRE RB)
Basically the colonies where under a yellowish light(much like a 10k MH).
When only the XRE whites were up and running.

In 7 days I had browned off and almost killed 3 SPS colonies.
Adding my RB back to the blend, and Ive stabilized 2 of them...not sure if the third will make it.
.

7 days of slightly reduced light won't kill coral IMO. If it did, reefs would die after one week of bad weather.

I worry that you describe 10000k as yellowish light. Cool white LEDs are lower than that (say around 5000 - 6500K), and are called cool because they look bluish. 10000K should look distinctly light blue.

 

[TheFishMan65]

Nylon washer can be obtained from mCmaster carr

 

[staggeringwade]

Are nylon washers the recommended choice for screwing down leds to prevent shorts? Is there some other material that is better suited for the job?

 

[theatrus]

If you're worried about temperature of the washers, Mica is the way to go. Expect to pay 10x the price for a washer though

 

[staggeringwade]

If you're worried about temperature of the washers, Mica is the way to go. Expect to pay 10x the price for a washer though

I'm not worried about heat too much I have a large heat sink.
I just could not get the stars tight to the heat sink with my screws and washers because the washers would squish outward and create a short.
Thanks for the advice.

 

[der_wille_zur_macht]

Nylon is fine. If your current arrangement won't allow a tight clamping force I'd rig something else before running the array longterm.

 

[staggeringwade]

Nylon is fine. If your current arrangement won't allow a tight clamping force I'd rig something else before running the array longterm.

I just ordered nylon washers from US plastics- they were cheap.
I will not run my array until everything is just right!
I fired up the array with a 9 volt battery on the dims of the meanwell for testing- i set the internal trim pot at 650 mah. Those trimpot screws are very sensitive-just barely even touching them with the screw driver went from 200 to 750 mah.
I received the Digital Aquatics ALC last week. I just ordered my RKE as the upgrade from the RKL.
So I'm trying to figure out how I'm going to set the amperage trim pots again once everything is running on the DA system (going from 9 volt battery test to 10v af the ALC)?
The array is in the garage and fish tank is in the house. I guess it might be easier to take the DA system offline for an hour and take the gear to garage and do the final mah tweaking before I button everything up. Doing this over the aquarium is a disaster waiting to happen!
ORjust leave it as is and assume that the difference from .650 mah at 9 volt will equal 750 mah at 10 volt thru the ALC? Is that too careless at this point? The leds are the cree XR-E kit from rapidled.com.
Wade

 

[der_wille_zur_macht]

just leave it as is and assume that the difference from .650 mah at 9 volt will equal 750 mah at 10 volt thru the ALC?

If you were really picky, you could look at the driver datasheet, find the %intensity at a 9v signal, the extrapolate up to what that would mean at a 10v signal. It's not QUITE linear, but close. If you aren't picky, you can just add 10% and you'll be close. So if you set it at 650mA on 9v, it'll probably be around 720mA on 10v.

Of course, that's all assuming that your 9v battery is exactly 9v, and your ALC is exactly 10v, and I doubt either of those statements are true.

 

[TheFishMan65]

IIRC it was nearly linear so going from 9 to 10 should increase 11% to about .7215. That won't be noticable to the eye, but the corals might care. Of course the coral might not like .75. So leave and see how the corals behave. $0.02