New LEDS

[VegasWinner]

I'd like to check these boards, but I don't like the heat sink profile they come with. If the could run on lower profile and cheaper hs like the quantums I'd be great. Waiting for calculations.

it looks like a 3" profile high heat sink. his heat sinks can be water cooled or air cooled. I think they are more powerful than what anyone else has right now. If I was looking to build a light I would get that board and mount it on my own flat heat sink and use his four channel driver, cost is similar to a meanwell C1400B driver. I would look at that 175w board for a 2x2 tent myself. He mounted the driver on top of the heat sink. I would remote mount the driver and have a 6" head room requirement like this

 

[greg nr]

Yes, he will be selling both a 100w and a 200w version of the board. That's a crazy amount of heat. He is recommending active cooling, either with air or water.

 

[GardenGuy]

@VegasWinner too many watts packed in a small board imho. Not talking bad on them, they actually might be the next big thing, but not for guys on a budget.

i'm waiting for the quantums to build my rig, and the chilled boards got me thinking again, but i believe i can do the my thing with less $$.
If i had more money to spend, i'd love to make a custom watercolled rig! it would rock!

 

[Mullumbimby]

Thanks for your input guys.
Does anyone know the efficacy (Lumens/Watt) of the CXB3590 @ 40-50W?
So long as I beat that and spend less money than last time, I'll be pretty happy.
The Veros are looking pretty good at the moment, at around 630mA

 

[CobKits]

pct.cree.com be sure to select proper bin and junction temp.

junction temp goes from 40-50C (cobs run soft or aggressive active cooling) up to 85C+ for cobs run aggressively or passive or undersized heatsinks

 

[doz]

New DB bin 3590 @ 3500k gives about 8% more lm/w than the CD bin. CD bin @ 1.4a was already 56% efficient. DB bin should be about 58-59% efficient @ 1.4a 36v. Ill take that all day long for $36/chip. Whoever posted that the 3590 are $50, look elsewhere....

 

[CobKits]

CD bin @ 1.4a was already 56% efficient.

thats been disproven in every real world test. id guess at 50-55% at chip level at normal operating temperatures (just a guess). also there is no guarantee that a DB chip is 8% over CD, its unlikely that the bins perform at the center of the bin range in each case. 3-5% bump may be more resonable of an expectation assuming that the CD bins were performing toward the top of their bin

ive seen single DB chips out of the same batch differ by almost 5% in output

 

[Mullumbimby]

pct.cree.com be sure to select proper bin and junction temp.

junction temp goes from 40-50C (cobs run soft or aggressive active cooling) up to 85C+ for cobs run aggressively or passive or undersized heatsinks

What a great tool.
Sadly, it doesn't show anything below 1400mA but it does let you see what a big difference every variable can make Especially jct, as you said.

 

[CobKits]

any cxb3590 data in supra's spreadsheets below 1400 mA was estimated based on relative to 3070 performance (current goes lower in PCT for those.)

larger and smaller cobs def seem to converge on efficiency at lower currents - the thermal penalty is leveled a bit

 

[doz]

Gotcha. I ordered a lux meter so that should give me an idea of where its at. Ill also get a better reading on the temps. I know on the hottest spot, it does not get near 85c and the TC is near the corner of ht echip.

 

[Hybridway]

Per my measurements, the following chips are about the same efficiency as CXB @ 42W:

1212 gen 5 at ~35W
1212 gen 6 at ~40-45W
1818 gen5/vero gen7D @ ~50W
1818 gen6/cxm22 @ ~60W
vero C/1825 gen 5 @ 60-70W

Good info!

 

[Mullumbimby]

I still like the look of Vero 29C Gen 7 below 1A at 185+ Lumens/Watt

 

[CobKits]

Gotcha. I ordered a lux meter so that should give me an idea of where its at. Ill also get a better reading on the temps. I know on the hottest spot, it does not get near 85c and the TC is near the corner of ht echip.

Tc should be pretty uniform over the package (if you could measure behind the les it should be similar.

an accurate voltmeter is a good way to guess at Tj.

while all cobs have different forward voltage, you can take a quick voltage measurement as soon as you fire the cob up (like first 5 seconds) and call that 30C

once the cob reaches steady state temp measure voltage again and compare your voltage drop with whatever the datasheet/calculator would predict. its not uncommon for voltage to be off by one or two tenths for your cob vs datasheet, but this will give you a relative guess to if youre running the cob at 55,75,95C Tj, etc.

 

[BuddyColas]

Tc should be pretty uniform over the package (if you could measure behind the les it should be similar.

an accurate voltmeter is a good way to guess at Tj.

while all cobs have different forward voltage, you can take a quick voltage measurement as soon as you fire the cob up (like first 5 seconds) and call that 30C

once the cob reaches steady state temp measure voltage again and compare your voltage drop with whatever the datasheet/calculator would predict. its not uncommon for voltage to be off by one or two tenths for your cob vs datasheet, but this will give you a relative guess to if youre running the cob at 55,75,95C Tj, etc.

So taking the 5-second voltage as a start, percentage wise what would be a tolerable droop? Like if it was greater than"X" percentage droop should be getting a bigger heat sink.

 

[CobKits]

depends on the spec of the cob and the current.

more specifically, the best heatsink you can afford.

if it drops a few% in output and still works for you it wont hurt the cob if you stay within the limits