DIY LEDs - The write-up

[Obi-dad]

I got a new meter (Extech) that has a very tiny temp probe. I put it on the dome of the XP-G LED, and it measured 150 degrees F after a couple minutes, with current at 350ma. This happens in both mounting methods I am using, one is screws with thermal compound, the other is glued with thermal adhesive. I haven't calibrated the thermometer to anything yet, I will put it in boiling water to get the 212 degree point, and icewater to get the 32 degree point.

Is it normal for the dome of the LED to be so hot?

 

[widmer]

Is the probe black in color? I wonder if it is absorbing the all of the radiation as well (ie the light) and that causes it to register a much higher temperature?

 

[sammy113]

I use a Ryobi meter and I can't get accurate reading until I use a piece of tape on the heatsink so it doesn't reflect or anything. Then I get a 115~120F reading

 

[bstohrer]

I got a new meter (Extech) that has a very tiny temp probe. I put it on the dome of the XP-G LED, and it measured 150 degrees F after a couple minutes, with current at 350ma. This happens in both mounting methods I am using, one is screws with thermal compound, the other is glued with thermal adhesive. I haven't calibrated the thermometer to anything yet, I will put it in boiling water to get the 212 degree point, and icewater to get the 32 degree point.

Is it normal for the dome of the LED to be so hot?

I've noticed that the dome of the XP-Gs are too hot to touch at ~700ma while XR-Es driven at the same current are only warm.

Bob

 

[laverda]

You can not use standard temp guns on metal. Especially shiny metal. The gun is looking at light. A shiny metal surface will reflect the views of cooler surrounding areas into your temp gun.

Sounds like we need to pay more attention to cooling our lenses and optic now.

I have used it on metal all the time and it has worked well, but not shinny like the aluminum U channels are. I was thinking that might have been the problem.

 

[stugray]

"I've noticed that the dome of the XP-Gs are too hot to touch at ~700ma "

Would you all cut it OUT?!?

You are NOT supposed to touch the emitter hot or cold.

From the handling document:

"Use tweezers to grab XLamp XP Family LEDs at the base. Do not touch the lens with the tweezers. Do not touch the
lens with fingers. Do not push on the lens."


Stu

 

[laverda]

stu
Good reminder. The reality is it is almost impossible to build a set up and not touch them at some point. Should the oils from you hands be cleaned off then with a little alcohol when finished?

 

[Seattleaquafarm]

How many leds should get for over a tank with the dimensions of 48"-24"-24"tall,

Should i get the dimable drivers or not? anything else I should know to would be amazing.

Thanks guys!

 

[kcress]

laverda; Yes, I should have been more specific - shiny metal.

Also, I would always expect the lens, that-I-would-never-touch, to be that hot. You have all that blue radiation hitting phosphors in the lens which then converts it to white light thru re-emission. This is not 100% so that means heating. Plus you have a very hot die mechanically connected to it. This is why you want that lens to have cool air around it if possible.

 

[westmoorenerd]

So, I'm in the planning stages of building my LED fixtures....I'm going to have 8 12" heatsinks with 30 LEDs a piece...After reading through this thread a few times I still haven't quite figured out the best way to power all of them. I'm not quite sure it would be that safe to run the necessary amount of ELN-60-48D I would need to run everything in series, but everything I've read here STRONGLY discourages parallel strings. Any LED gurus want to chime in on my best options? Maybe Meanwell's new LED drivers the HLG series?

 

[Obi-dad]

"I've noticed that the dome of the XP-Gs are too hot to touch at ~700ma "

Would you all cut it OUT?!?

You are NOT supposed to touch the emitter hot or cold.

From the handling document:

"Use tweezers to grab XLamp XP Family LEDs at the base. Do not touch the lens with the tweezers. Do not touch the
lens with fingers. Do not push on the lens."


Stu

I am only touching a couple of the extra LEDs that I have for experimentation. I wanted to see if mounting them with thermal adhesive vs screws/thermal compound made a difference, as well as mounting them to aluminum bar vs heatsink. I had been measuring the heatsink, but when I accidentally touched the lens, I was very surprised at how warm it was. By the way, the thermal adhesive and the screws/thermal compound give me the same readings within a few degrees.

So does anyone know what is reasonable temp for the lens? I was measuring while running at approx 350ma. 700ma is much hotter to the touch, but I haven't measured that yet with the thermocouple.

The 150 degree F figure I got with the tiny thermocouple I would bet is a lot more accurate than the IR gun I have tried (besides the emissivity issue, the little red dot that people think is measuring the temp is only a visible laser to guide you - the real measurement is taking place with infrared, and the measurement area is much larger than the little dot from the visible laser - an IR gun does not help us with small items like the lens of an LED).

 

[The Grim Reefer]

stu
Good reminder. The reality is it is almost impossible to build a set up and not touch them at some point. Should the oils from you hands be cleaned off then with a little alcohol when finished?

I wouldn't use alcohol, maybe a damp cloth. Alcohol can leave a greasy film unless it's the really pure stuff and even that can cloud come plastics.

Just an FYI I knocked the lenses off a couple LED's and while I am not using them on my fixture they still work fine. Makes them a REALLY wide beam

 

[HeneryH]

Isopropyl alcohol was in the spec sheet as the recommended cleaner.

 

[The Grim Reefer]

Isopropyl alcohol was in the spec sheet as the recommended cleaner.

Then just make sure it the pure stuff, I have some here I used to clean my heat sinks and it's only 70%. Made the mistake of cleaning my camera lens with it, left a nice film of something on it I had to clean.

 

[Jay1982]

Can anyone show me some pics of how they epoxied their optics onto their starboards?

I'm really not enjoying this. They are crooked and I'm using alot of epoxy.

It seems like I'm using a lot of epoxy and the stuff just isn't sticking:

 

[wld1783]

Paralleling LED Strings – NOT RECOMMENDED


But what other opitions do we have if we want to run 150+ leds and not have 12+ Meanwells that might burn the house down when they start up.
Or have a 240V driver? That would be scary.

This is what I did.

Parallel strings have issues but they miss the big picture...practicable application. The CREE XR-E has a max DC forward current of 1000mA. at 3.2 volts and 100% efficiency you would have a 32watt fixture. The LPC 35-700 has an output of 700mA. If you have 2 parallel strings of 12 in series 350mA will be going through each string at appx 60% less light. Note the PDF link page 8.
https://www.reefledlights.com/PDF/XLamp7090XR-E.pdf
Running the XR-E at 350mA X 2 is more efficient by a TLAR 10%. If the worse happens, Gremlin opens a wire in one parallel string then the other parallel string gets all 700mA well below the the max of 1000mA

Now my scenario.

1 CLG 150-48 driving 48 LEDs 4 parallel strings of 12 at the rated 3.2 A or 800mA per string. (This makes sense as the PAR was 9% better using this configuration vs 4 LPC 35-700)

Now the worse happens a Gremlin opens a wire on one of the parallel strings. The other three get 1060mA each 60mA over their rated max...Only 6% over. I feel that there is enough resistance in the circuit to handle 6% but to be safe it would be prudent to add a resistor or dial down the driver. It is constant current so this is easy and is what I would do and recommend using a KillAWatt Meter. To be safe I'll attach the meter and dial it down 15%.

Now if a Gremlin opens two strings the other two are toast. What are the odds.

With proper planning the only downside to parallel strings is you're not maximizing the output of the driver or the LEDs but Watts are Watts and PAR efficiency is realistically consistent. If one string is a little off the other string will be slightly brighter or dimmer no big deal. Its just a larger hardware cost for redundancy but better to be safe than drive them at 100% and risk damage due to a simple failure.

Also running parallel strings reduces the number of drivers. Heck you can run 96 LEDs on the CLG150-48 and still get over 160 PAR @ 24" using CREE XR-E LEDs on a 48"x8" heatsink.

Bill

 

[widmer]

Maybe now is a good time to add something I have been thinking about:

It is quite possible that it is very important you don't touch your optics with bare hands. This is because they rely on the the difference in refractive index of the plastic and air at the interface created by the sides of the optic, so that the light does in fact bounce up and out of the lens, rather than right through the sides. If a person touched this with his bare hands and left oil so that the interface was actually plastic : oil : air, I'm quite sure that the optic might not behave like it's supposed to, which would result in less light going where you want it to.

See http://en.wikipedia.org/wiki/Refraction

 

[westmoorenerd]

So, my tank is going to be lit by 240 LEDs, and I'm slightly worried about harmonics and just the price of drivers in general. I was looking at the specs of the HLG-240-20, it looks like it's enough to run 12 parallel strings of 6 LEDs at 3.2v at 800mA...So each one could power 72 LEDs in a parallel series right? Or is it better to go with a higher voltage, lower amperage driver and run more LEDs in each series?

The way I see it, with smaller series in parallel, if one string failed, it wouldn't affect the other strings as much... Say each 6 LED string in 12 is running at 800mA, if one LED (thus the whole string) failed the rest of the strings would still be running at 872mA which is perfectly safe right?

Please correct me if I'm wrong with this, I'm still pretty new to hobby electronics. I'm used to looking at diagrams and troubleshooting, not building from scratch.

 

[tahiriqbal]

Can anyone show me some pics of how they epoxied their optics onto their starboards?

I'm really not enjoying this. They are crooked and I'm using alot of epoxy.

It seems like I'm using a lot of epoxy and the stuff just isn't sticking:

---------------------------------------------------------------
Jay......

Bad idea..... It looks messy and it seems that you are spilling some light since they are not 100% fit over your XP-G's. What you could do is, get your dremel out (if got one) and grind the edges where your solder wires meet the optic holder. Its much easier and doesn't look bad at all....

 

[wld1783]

Try the Loctite No Mess Adhesive. Its like silicone and non toxic. You can get it at Home Depot

Bill