Top bin COB comparison 2

JorgeGonzales

Well-Known Member
At 700ma 1212 = 24.2w the 1818 = 35.4 thats 31% more power from same ampere. The 1212 costs in my country and yours 2 between 10-12 the 1818 start at 24. Yes yes :)
Man I have no idea what we are talking about. I was just saying the 1212 won't be ahead of the 1818 at any wattage, below or above 36W, if we are talking about light output. Maybe we aren't.
 

optzulu

Well-Known Member
Talking about ampere, Watt is for hps :)
For example you want led fixture that runs on 700ma and will cover 1m2, you say that the 1818 fixture will be better I say no.
you need either sixteen 1212 cobs for 170 euros
Ore eleven 1818 chips for 240 euro.To cover the 1m2 space now pleas don't tell me that the 1818 will have a better lm/W ratio and 16 cobs will not have a better ppfd and spread then the 11 cobs.
 

JorgeGonzales

Well-Known Member
Talking about ampere, Watt is for hps :)
For example you want led fixture that runs on 700ma and will cover 1m2, you say that the 1818 fixture will be better I say no.
you need either sixteen 1212 cobs for 170 euros
Ore eleven 1818 chips for 240 euro.To cover the 1m2 space now pleas don't tell me that the 1818 will have a better lm/W ratio and 16 cobs will not have a better ppfd and spread then the 11 cobs.
Shit now I have to figure this one out...yeah I can't do it. Why do you care about amps? Watts are for everybody. My TV, my microwave, my refrigerator, and my lights.
 

BOBBY_G

Well-Known Member
The 1212 will win from the 1818 from 1w to 36w then the 1818 will get ahead, but the 1212 costs also 50-60% less then a 1818 and only have 1/3 less power.
dont bank on that. ill be testing 1212 vs 1818 and 1825 in next few days but i would expect 1212 to lead on anything but price (which if running soft its prob "close to or equal" to the larger chips but cheaper on a $/watt
 

JorgeGonzales

Well-Known Member
dont bank on that. ill be testing 1212 vs 1818 and 1825 in next few days but i would expect 1212 to lead on anything but price (which if running soft its prob "close to or equal" to the larger chips but cheaper on a $/watt
I think it's safe to just count # of die for the Citizen line, 12*12 vs 18*18 is all the math needed. 9 1212s vs 4 1818s = same # of dies, same price, same performance.
 

BOBBY_G

Well-Known Member
At 700ma 1212 = 24.2w the 1818 = 35.4 thats 31% more power from same ampere. The 1212 costs in my country and yours 2 between 10-12 the 1818 start at 24. Yes yes :)
by your logic 72V crees would be twice as "powerful" as 36V when watt for watt they are identical
 

BOBBY_G

Well-Known Member
reran 1825 at low currents to clean up the data. in the future ill use a voltmeter to get two decimal places on the voltage instead of the one my power supply displays. sctter below 25W prob just rounding error when computing watts from V x I

upload_2016-7-11_21-1-53.png

upload_2016-7-11_21-0-30.png
 
Last edited:

BOBBY_G

Well-Known Member
from the cob efficiency spreadsheets thread:
For pulsed, i dial in the current then turn i off for a few seconds and then get the pulsed voltage current and PPFD. Will try the max on the DMM next time.
I just try to stay consistent while comparing cobs. I do not move the lamp or the PAR meter. Also i mount the cobs on different heatsinks of same size before i start testing so that i do not have to swap cobs during experiment.
unfortunately my adjustable power supply surges on on/off and i dont want to blow any chips up. thats what you get for $100... in any case even with a driver supplying power id be suspect of peak readings as id imagine youd kind of need a large resistive load or something to buffer it. my best meter is a fluke but its almost 20 years old, does not have peak hold (nor does my li-cor li250A)

im confident based on my observations that the numbers are very accurate below 2A (say within 1% which is prob in the neighborhood of the combined error of the voltmeter and par meter). 2A-3A def has some droop, but in any case id rather have repeatable numbers than absolute numbers of dubious validity, so in this case temp droop is included (for all chips in the test, if that helps). like robin i mount multiple chips on similar heatsinks in advance and have heatsink and sensor position marked so as to minimize error
 

Rocket Soul

Well-Known Member
Found this link on one of the threads around her that u cant seem to find again:.
http://www.futureelectronics.com/en/Technologies/Product.aspx?ProductID=BXRC30E10K0C73BRIDGELUX2073533&IM=0

If i understand it right they give prices for gen7 veros between 25-30ish $. No stock though.EDit: MOQ 50 is the bad news...

In this link the 70v version, looks very efficient. Im loving them for 6 of them on a meanie hlg185c500. They should squeeze in with lower voltage at lower current and give north of 200 lumens per watt w each chip around 33 watts. Compare to classic 8 cxb config on same driver but at 700mA:

You would pay maybe 180$ for 6 veros instead of 360ish for 8 cxb. I know citi looks really good aswell but eu supply situation is fucked up unless u are in poland w tme. I cant even get a quote from my distributor.

Also Vero 29c seems more efficient than the d version tested here, at least per datasheet. Im liking this new vero big time, can soneone with traction with Jerry@kingbrite get on his case to source these?
 

BOBBY_G

Well-Known Member
wow the tests i did today with graphite pads were the suck at high currents. an 1818 was topping out at 690 umol at 2600 mA and further currents made the PAR go down, so i knew i had a problem. replace with arctic MX4 > 917 umol@ 3A. that was a 3000k that got 7.39 ppfd/W @ 70W and 5.74 ppfd/W @159W which should fix the discrepancy seen with robins test from here (though i think that CXB might be on graphite as well...)

will never use graphite pads again they may work fine at low currents but not with higher dissipation. another cxb chip i had that was working "fine" at high currents with graphite gained 5% ppfd/W at 3000 mA with MX4

the 5000k tests i did yesterday had the top performers using mx4
the others used arctic silver pads (soft not like graphite) ill redo those with pads with mx4 and rerun, to see if it helps a little
 
Last edited:

DrBlaze

Well-Known Member
will never use graphite pads again they may work fine at low currents but not with higher dissipation. another cxb chip i had that was working "fine" at high currents with graphite gained 5% ppfd/W at 3000 mA with MX4

the 5000k tests i did yesterday had the top performers using mx4
the others used arctic silver pads (soft not like graphite) ill redo those with pads with mx4 and rerun, to see if it helps a little
I've been following along here for months now watching the evolution of these cobs, and have been surprised that so many otherwise obsessive DIYers have been giving up efficiency by using these pads. While I've never built an led fixture I've built plenty of pc's over the last 20 years and would NEVER consider using them. If you look at the larger tests, the pads are waay behind even the worst pastes on the market. The tests usually include something odd, like denture creme, for reference. The pads are basically a bit better than using denture creme.
 

Bubblin

Well-Known Member
I've been following along here for months now watching the evolution of these cobs, and have been surprised that so many otherwise obsessive DIYers have been giving up efficiency by using these pads. While I've never built an led fixture I've built plenty of pc's over the last 20 years and would NEVER consider using them. If you look at the larger tests, the pads are waay behind even the worst pastes on the market. The tests usually include something odd, like denture creme, for reference. The pads are basically a bit better than using denture creme.
You couldn't pay me to use a thermal pad on anything, ever. They're often worse than no paste.

Arctic silver 5 and Arctic silver ceramic both have amazing thermal conductivity and longevity. Aka take it apart a year down the road and there's not dried up chunks of flaky shit falling about. Many pastes degrade over time, often a lot.

Ceramic has almost no cure time either, Arctic silver 5 can take days to cure, but it's worth it imo.
 

MeJuana

Well-Known Member
Talking about ampere, Watt is for hps :)
For example you want led fixture that runs on 700ma and will cover 1m2, you say that the 1818 fixture will be better I say no.
you need either sixteen 1212 cobs for 170 euros
Ore eleven 1818 chips for 240 euro.To cover the 1m2 space now pleas don't tell me that the 1818 will have a better lm/W ratio and 16 cobs will not have a better ppfd and spread then the 11 cobs.
We need two values not just amps. It's amps * volts... For measuring device output we always use watts even with little dc electric motors we measure output in watts.
P(W) = I(A) × V(V) (WATTS EQUALS, AMPS TIMES VOLTS)
I(A) = P(W) / V(V) (AMPS EQUALS, WATTS DIVIDED BY VOLTS)

Just because we are using DC current exclusively now it doesn't change this.
 

weed-whacker

Well-Known Member
wow the tests i did today with graphite pads were the suck at high currents. an 1818 was topping out at 690 umol at 2600 mA and further currents made the PAR go down, so i knew i had a problem. replace with arctic MX4 > 917 umol@ 3A. that was a 3000k that got 7.39 ppfd/W @ 70W and 5.74 ppfd/W @159W which should fix the discrepancy seen with robins test from here (though i think that CXB might be on graphite as well...)

will never use graphite pads again they may work fine at low currents but not with higher dissipation. another cxb chip i had that was working "fine" at high currents with graphite gained 5% ppfd/W at 3000 mA with MX4

the 5000k tests i did yesterday had the top performers using mx4
the others used arctic silver pads (soft not like graphite) ill redo those with pads with mx4 and rerun, to see if it helps a little
even in cree's design pdf for the CXA they stated that grease was the best, pads 2nd and adhesive last
 
Top