Is this a Vero 29 killer?

optzulu · May 23, 2016

I had today contact with the distributor of these lights, they cant sell me now the 1212 it wil take about 4 weeks. They will be 11€ for a piece but he had the 1812 on stock for 15€.

How much better is the 1812 then the 1212?

BOBBY_G · May 23, 2016

datasheet would tell you all you need to know in terms of lumen output and efficacy

BOBBY_G · May 23, 2016

JorgeGonzales said:
@BOBBY_G If you have to compare at Tj=39 to make the PCT say what you want, there is something seriously wrong with the PCT. Now I know that @SupraSPL says that it's because Tj is actually mislabeled, and is Tc, which if it's true means Cree has pretty terrible thermal resistance, plus they don't report current droop right, plus they under report lumens, not to mention they don't fix bugs.

This is why we need integrating sphere testing.

By the way, your 166lm/W @ 85C was spot on, but sort of pure luck. It's 2712lm/16.3W/32.6V. Eyeballing charts is just impossible.

Anyway, fuck that noise. The PCT is untrustworthy, or Supra is. If we can agree that the SupraSPL datasheet is correct, and Citizen's simulator is correct, it gets really easy to compare. And sort of fun.

For instance if you run 2x CLU048-1212, you get about 155lm/W at 6568 lumens. The exact same 13137 lumen output as a single CXB3590 at 2.4A, for $25. If you use a monolithic heatsink, the costs for cooling are identical, and I would argue in this case the cost to cool one cob at 87W is about the same as two cobs at 42W. Heatsinks just scale up with weight, basically.

At 500ma, I have no idea, you are probably right. I just thought the numbers were a little wonky. As in, not agreeing with the PCT at Tj=85C, or Supra's spreadsheets, or the Citizen datasheets.

But lets look at it at 50C, which is an easy number to hit at 500ma. CLU048-1212 makes 175lm/W at 500ma. Cree makes ummm...178lm/W at 1.5ma drawing a straight line between 1.4ma and 2.1ma. Equal costs, same efficiency (Citizen slightly ahead in umol/J), 3x the spread. You are probably right if you did individual heatsinks the Citizen solution would cost more, but probably not much. An extruded solution like heatsinkusa would cost the same regardless of number of cobs. Labor is higher, but people love that spread, yo.

This is way too long. But it was fun to do the math. At higher Cree currents, a couple $12 Citizens pull ahead. That's pretty cool. And it's a horse race at 500ma vs 1.5A.

For everybody's future reference here is the Supra spreadsheet:https://docs.google.com/spreadsheets/d/1nKfc12OCuDu7puJCa_6maefyCVm3gS5sBmzAXOlh06w/edit#gid=1157681753

And the Citizen simulator:
http://ce.citizen.co.jp/lighting_led/dl_data/sim/CITILED_ver5_Selection_Tool_WEB.xlsx

i hear ya on the PCT but again its because they use Tc and Tj differently on the datasheets than PCT. really only a side-by-side will tell us (my citis arrive tomorrow supposedly)

i havent been able to use the citi tool as i am still using office 2003 as i hate bloatware. this would be the first time in my life as a technical professional that ive had a need to use all the features of an .xlsx file so be it.

in any case they are close and you get diffferent results if you compare them run soft where crees shine or run hard where citis shine. we need to get them on a supra-style ppfd/W chart then we'll see whats up. i should have ordered a 1212 for science sake but didnt. all the ones i have coming are 1825

JorgeGonzales · May 23, 2016

optzulu said:
I had today contact with the distributor of these lights, they cant sell me now the 1212 it wil take about 4 weeks. They will be 11€ for a piece but he had the 1812 on stock for 15€.

How much better is the 1812 then the 1212?

Question is irrelevant, you can only purchase one. It's cheaper and better than a Vero 29. The answer is a about halfway between a 1212 and 1818.

I would have used an 1812 if it was in stock.

JorgeGonzales · May 23, 2016

@BOBBY_G Try this in Excel online, it's what I've been using, or gf's work computer. It's goofy as hell with no dropdown menus, but it works. Uploading here because I had to unprotect the worksheet by manually editing xml data to get it to even open.

BOBBY_G · May 23, 2016

still says #NAME? in all the cells

nogod_ · May 23, 2016

I dont have msoffice, i use WPSoffice for android and it works fine.

BOBBY_G · May 23, 2016

open office calc wouldnt work either

BOBBY_G · May 23, 2016

ok here's Excel Online

am i doing it right? selecting #7 for 3500K 80 cri?

the 1825 dominates at all currents. 158 lm/W @ 130 W for $40 is amazing

JorgeGonzales · May 23, 2016

BOBBY_G said:
ok here's Excel Online

am i doing it right? selecting #7 for 3500K 80 cri?

the 1825 dominates at all currents. 158 lm/W @ 130 W for $40 is amazing

Yeah, I have no idea what the middle 2 is, but the 7 is your cob, and the right 2 can also be a 1 which lets you specify lumens instead of mA, which makes comparisons at lumens pretty simple.

nevergoodenuf · May 23, 2016

You guys are finally seeing what I have been looking at for months. Higher power, less COBs, over a larger area. I guess I am just flying under the radar

exkalapalatt · May 23, 2016

The 1212 CoBs have 12 diode in a string and the CoB contains 12 string. The first two number is the numer of the diodes in a string and the second two numbers are the strings.
So a 1812 cobs contains +50% diodes than a 1212.

When you drive a 1825 CoB with 2400mA that is less than 100mA/diode or string. At this point a 1812 get 2400/12=200mA/diode.
Drive the CLU058 with 4800mA and lets see is it better than 130lm/W..

BOBBY_G · May 23, 2016

4000 mA

4800 mA

BOBBY_G · May 23, 2016

but wait were not done.

5750 mA!

BOBBY_G · May 23, 2016

ill try to run a bunch of sims and make a graph. i like graphs.

optzulu · May 23, 2016

I really hope these numbers are correct the clu58-1825 is indeed the king, with 141lm/w= 43% effic @4000ma is insane you are putting 228watt of power from each cob. this could be the HPS killer lol

guod · May 23, 2016

BOBBY_G said:
am i doing it right? selecting #7 for 3500K 80 cri?

yes, and all nr.

nevergoodenuf · May 23, 2016

Well, since you guys are running #'s and I don't have Excel. What is you output for the CLU550 (3000K, 70cri and 4000k, 80cri) @ 1.05 amps with a Tc of 50*c? I am rewiring my drivers right now to run them super soft. I am putting one COB per HLG120-c1050b. I also have the ones in the pic above that will also run on the same driver.

JorgeGonzales · May 23, 2016

BOBBY_G said:
ill try to run a bunch of sims and make a graph. i like graphs.

You can do SupraSPL style charts too if you want, we have pretty good LER guesses here and there.

Alesh last gen

Some I did using Alesh's math/spreadsheet. Interesting that he has a higher QER at 2700K, but could also just be sampling noise. I was happy with the digitizing, but picking the right zero point can alter things slightly, and all these graphs have lines that color outside the borders.

CLU048 2700K 80 CRI

CLU048 2700K 90 CRI (H6K Below BBL)

CLU048 4000K 80 CRI

JorgeGonzales · May 23, 2016

nevergoodenuf said:
Well, since you guys are running #'s and I don't have Excel. What is you output for the CLU550 (3000K, 70cri and 4000k, 80cri) @ 1.05 amps with a Tc of 50*c? I am rewiring my drivers right now to run them super soft. I am putting one COB per HLG120-c1050b. I also have the ones in the pic above that will also run on the same driver.

I'm not sure the CLU550 has a spreadsheet like the 0x8 series does. Oh, there it is on the old 0x6 spreadsheet.

I think we are all running the free Excel online because nobody has Excel

Is this a Vero 29 killer?

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Attachments

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Active Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member