Grams per watt, LEDs outperform HPS

Corso, I've seen very decent results with Hydrogrow lamps. Just a quick run through their site claims a 95F operating temperature (enough to not kill the AllinGaIP reds) and use Bridgelux LEDs which are better than the no-name Chinese kind. Their 600w unit also goes for a cool $2000 so keep that in mind. The decent lights seem to be the most expensive.

But there is no "best". Thermal management is problem number one, followed by type of LED used and spectra. If they have those bases even remotely covered the light will work well and last.

Check out the difference in light output from this one test when the junction temperature gets 10C above it's limit. It's from Bergquist's website; they specialize in thermal management solutions for LEDs.

Junction temp goes from 32C to 42C, light output jumps from 630 lumens to 110 lumens!!!

.LED heat management.jpg

Originally Posted by bassclef

Corso, I've seen very decent results with Hydrogrow lamps. Just a quick run through their site claims a 95F operating temperature (enough to not kill the AllinGaIP reds) and use Bridgelux LEDs which are better than the no-name Chinese kind. Their 600w unit also goes for a cool $2000 so keep that in mind. The decent lights seem to be the most expensive.

But there is no "best". Thermal management is problem number one, followed by type of LED used and spectra. If they have those bases even remotely covered the light will work well and last.

Check out the difference in light output from this one test when the junction temperature gets 10C above it's limit. It's from Bergquist's website; they specialize in thermal management solutions for LEDs.

Junction temp goes from 32C to 42C, light output jumps from 630 lumens to 110 lumens!!!

.LED heat management.jpg

Note on top of the image it says: "Light output of the same LED die on different circuit board materials at a maintained die temperature of 80°C (176°F)"

better off having my brother (electrician) build me a couple?

Looks like the test come from this site or some like it:
LEDs - Temperature Effects
http://www.bergquistcompany.com/ther...mp-effect.html

It also looks like it's a test of thermal clad versus FR-4
and it sure looks like FR-4 is a crap -that's good to know.

But looking at the PCB's called FR-4 like this witch has holes in then where the LEDs heat transferring metal part goes through
I can't see why thees would be crap if mounted on a proper heat sink with good thermal paste. ?

And the reason this is getting my attention is that the new panels like Apollo 6-8-10


Is using what looks could be FR-4


I don't think the test is applicable to our grow light(1&2gen.) with the led mounted directly on the
heat sink and not on a pcb.

Rasser, I think it would be difficult to fill in the gap in the FR-4 holes there with something electrically isolating. Thermal adhesive might work, but you'd have to goop it in there; not really the best way. And if you use Kapton tape, you also suffer thermal transfer penalty.

The challenge is transferring heat quickly from the thermal slug of the LED (on the bottom) into the heatsink quickly with the thinnest layer possible while also electrically isolating it. Currently the Bergquist stars have the best thermal conductivity in this area, as most DIYers mount them to the heatsink with good thermal epoxy.

I personally use wakefield delta pads, they are much thinner than the bergquist stars (.007in) with the same thermal conductivity (1.3W/m-K) and electrical isolation. FR-4 thermal conducitivy is much less at 0.3W/m-K.

Originally Posted by Rasser

I don't think the test is applicable to our grow light(1&2gen.) with the led mounted directly on the
heat sink and not on a pcb.

Absolutely it's not, it wasn't meant to be a direct comparison. I just wanted to show how LEDs (especially reds) drop off a cliff with light output once they reach ~40C or so. A heatsink will reduce temps for sure. Problem is if the fans or enclosure isn't great at getting rid of heat, or the heatsink is too small, and the whole unit heats up.

Originally Posted by bassclef

Absolutely it's not, it wasn't meant to be a direct comparison. I just wanted to show how LEDs (especially reds) drop off a cliff with light output once they reach ~40C or so. A heatsink will reduce temps for sure. Problem is if the fans or enclosure isn't great at getting rid of heat, or the heatsink is too small, and the whole unit heats up.

Just thought for a moment that "Junction temp goes from 32C to 42C, light output jumps from 630 lumens to 110 lumens!!!" was correct.

Looking at the chart from the thread BS is doing it wrong I don't see that a dramatic effect on light output vs. temperature.