Designing and Building a LED Fixture

[Wasserpest]

Allow me to sticky this for now... excellent write-up Hoppy.

[CL]

Congratulations on the sticky
There's a lot of excellent info in this thread

[Hoppy]

Wow! Thank you. I just hope somebody will use this to make their own light.

[redman88]

already started to order the parts. spent 65ish dollars so far thats including shiping though i don't have to shipping on a few of the items yet. just need to order the LED's
(bump)

i am going to attempt to do this with just 21 LED's over a 29 gallon see how that works for me if needed i should be able to upgrade with another 21 at a later date

[Hoppy]

Well, the "adventure" isn't over

Yesterday morning the lights came on as the timer told them to. A few minutes later as I was starting to add the day's ferts, the lights suddenly dimmed. Only 6 LEDs, in one of the 12 LED series groups, remained on, with all 6 being together, the 1st thru 6th from the + lead. The other series group were all off. I disconnected it after checking that the cooling fans were on, and stewed about this the rest of the day.

This morning I tested the LEDs individually, and they all work, so it isn't a failure of LEDs.

I'm stumped, but I do have some ideas about the problem. Here is a question for the electronics skilled members here: If I run 3 LEDs in series, use a constant current device to provide them with 400 mA current, all 3 LEDs will light up as 400 mA current goes through them. Now, suppose my power supply voltage drops below 3X the forward voltage of each LED - for example, if they each run at 3.5 volts at 400 mA (10.5 volts total in series), and my power supply output drops to 10 volts, what happens and why?

I always think in terms of resistors in series or parallel, just from intuition. That thinking says the currrent drops until the voltage available equals the total voltage drop across the LEDs. But, the constant current device won't allow that to happen, so the current goes to zero. (?) With LEDs, each LED will drop the voltage available to the next one in series by 3.5 volts, so the first two will be ok, but the third one only gets 3 volts, which is too little. What does it do, and why? Will a LED supplied with current, but inadequate voltage pass the current along, but not light up?

In my actual circuit, I have about 45 volts available for total forward voltages of the LEDs in series, and 12 in series, each needing 3.5 volts, or 42 volts total. But, suppose a questionable solder joint adds a resistance in the circuit, dropping 20 volts at 400 mA? Would that allow 6 LEDs to light up, but leave the last 6 in the series just passing current without lighting up?

Any other ideas?

EDIT: I just found 3 solder joints at LEDs that were questionable, one of which I could break loose by hand. I resoldered the 3 connections, but that didn't help. Also, I realized that the "-" connection to the two series strings of LEDs is ground, and the heat sink is grounded. So, any current leakage to ground, perhaps at the 6th LED that is lit up, would shut down the last 6 of that string, and that string would then hog the current, leaving the other string underpowered, so it would be off. It may be that my soldering is the total cause of this problem, with the resistance at one or more joints increasing, or bits of solder making a connection to the heat sink, etc. causing the "failure". It is becoming obvious that a 15 watt soldering iron isn't up to doing a good job on mounted LEDs, so I may end up buying a 25 watt one.

[Hoppy]

Progress: The only type of failure I have been able to think of is a grounded LED in the string that has 6 lit up LEDs, so those "downstream" from that grounded one get no current through them, and the other string gets too little current to light up, since most of the current goes to the lower total voltage drop string with the ground.

Hip, hip, hurrah! Thanks to the troubleshooting tips from an experienced LED hobbyist on the Barr Report, I found the problem! First, I disconnected the ground connection to the string that didn't light up at all. Plugged it in, and nothing changed. So, I reconnected that ground, and disconnected the ground to the string that had the 6 still operating LEDs. Plugged it in, and the same 6 continued to be the only ones lit up. This had to mean there was a grounded LED in that string. So, I removed the mounting screws for the last lit up LED and placed a piece of paper behind it to keep if off the heatsink. Plugged it in, and the previously unlit string came on fully, and the string that had the 6 still lit ones was dark. Absolute proof that that LED was shorted to ground at the "-" end of it. I still can't see anything wrong on it, nothing to show that it is grounded, but there is no question that it was grounded. To double check this, I then reconnected both grounds, so both strings had the "-" end grounded, but left the faulty LED separated from the heat sink. Plugged it in, and both strings light up fully.

My tentative guess is that moisture between the star mount and the heatsink may have been involved, or the star mounting is not well done by the assembler. Tomorrow I will figure out what next. Any suggestions? Is there a relatively easy way to electrically isolate that one - with a mica shim, for example? Or would it then overheat?

[JDowns]

Mica would work, or Ferrite.

Here is just one link of grease epoxies that would probably work in your operating temperature ranges, giving you good heat transfer and electrical isolation.

http://www.epoxies.com/therm.htm

[JDowns]

I would call them for a recommendation given your application and then order a sample for evaluation.

[Hoppy]

So, the epoxy thermal compounds are good electrical insulators but also good thermal conductors. Frys has several types of thermal compounds in stock, so I may visit there and pick up the best one. I sure don't want to wait for some one to ship it. That is a good idea though. Sometimes you get so close to a project you can't see the things right in front of your face!

[JDowns]

There are quite a few epoxies that are very good at thermal conductivity while also being electrically isolated. Just not sure what your operating temperatures are at.

I would think you could apply the epoxy to your heatsink, let dry, then mount the star.

Frys more than likely carries such a compound or grease. I haven't personally looked in my hours of wandering the store looking for ways to waste money, but its probably worth a look


Another type product

http://www.omega.com/pptst/OB-100_OB-200_OT-200.html

A thin film of the OB-200-2 and that 2oz should go a long way.

[Hoppy]

Problem solved - for now anyway. The "failure" was current leakage to ground through the "line" contact where the screws held the one LED to the heat sink. Looking very carefully I can see slight discoloration around the screw notch in the star mount, where it contacted the conical back face of the screw head.

Today I found some Arctic Silver brand alumina filled epoxy heat conductive cement at Frys. I thoroughly cleaned the heat sink in the contact area, cleaned the back of the LED (using alcohol), and reinstalled the LED using the alumna filled epoxy. After a half hour (it sets up in 5 minutes) I plugged it back in, and only the 6 LEDs ahead of the failed one lit up. So, I removed the attaching screws entirely. Plugged it in again, and it worked fine. That is when I looked for and found the discoloration.

Now I suspect that I will get this failure again, as another LED leaks current at that spot. If so, I will then mount all of the LEDs with the ceramic epoxy and not use screws at all. In the future I won't try to use the screws if I make another one of these.

[JDowns]

Originally Posted by Hoppy

Problem solved - for now anyway. The "failure" was current leakage to ground through the "line" contact where the screws held the one LED to the heat sink. Looking very carefully I can see slight discoloration around the screw notch in the star mount, where it contacted the conical back face of the screw head.

Today I found some Arctic Silver brand alumina filled epoxy heat conductive cement at Frys. I thoroughly cleaned the heat sink in the contact area, cleaned the back of the LED (using alcohol), and reinstalled the LED using the alumna filled epoxy. After a half hour (it sets up in 5 minutes) I plugged it back in, and only the 6 LEDs ahead of the failed one lit up. So, I removed the attaching screws entirely. Plugged it in again, and it worked fine. That is when I looked for and found the discoloration.

Now I suspect that I will get this failure again, as another LED leaks current at that spot. If so, I will then mount all of the LEDs with the ceramic epoxy and not use screws at all. In the future I won't try to use the screws if I make another one of these.

You can still use the screws. Just use ceramic washers as a barrier.

Glad to see you got it fixed.

[Hoppy]

Originally Posted by JDowns

You can still use the screws. Just use ceramic washers as a barrier.

Glad to see you got it fixed.

Back during another "debate" over how to mount the LEDs, by another DIYer, he chose to use a silicone material to mount the stars. In retrospect I think that was a better idea than screws. It is true that you can use insulating washers, and bushings, to separate the screw from the star, but that is very much harder to do. For one thing, the notches are designed to fit #4 diameter screws, but with an insulating bushing, you would need smaller diameter screws, hard to find and much harder to drill and tap for. I would rather just use adhesive, easy and simple. Here is the one I used: http://www.arcticsilver.com/arctic_a...l_adhesive.htm Time will tell whether this was a good choice.

For a reef light, needing the LEDs to be driven at near maximum current, I think it would be worth the added effort to use insulated screws, but for our use, the heat that has to pass from the LED to the heat sink is much less and a lesser thermal compound should work fine.

[evilc66]

You can always use nylon screws.

[Regloh]

Nylon can only take 180-230F long term. If your fixture stays below that you are in the clear...
I have used ceramic screws before, but those are not really affordable on a DIY project