Top bin COB comparison 2
- Thread starter robincnn
- Start date
BOBBY_G
Well-Known Member
3000K
Li-Cor LI190 @ 12"
all with Arctic MX4 paste
all on PLC 140mm pinfins except the 1212s which are on Cutter 111mm
all@22C ambient, 120mm fan blowing down on pinfin
1212 looking good up to 30W, maybe could do ever so slightly better on larger heatsink
1818 matching CXB3590 @ 60-90W at just over half the cost
1825@60-120W dealing out 45-50% more photons at a given efficiency than the CXB3590@40-80W, at about 10% less cost per chip
Li-Cor LI190 @ 12"
all with Arctic MX4 paste
all on PLC 140mm pinfins except the 1212s which are on Cutter 111mm
all@22C ambient, 120mm fan blowing down on pinfin
1212 looking good up to 30W, maybe could do ever so slightly better on larger heatsink
1818 matching CXB3590 @ 60-90W at just over half the cost
1825@60-120W dealing out 45-50% more photons at a given efficiency than the CXB3590@40-80W, at about 10% less cost per chip
Last edited:
Airwalker16
Well-Known Member
That's crazy! ! Where do you even get citizens in us tho? NorthernGrowLights?
The Dawg
Well-Known Member
dandyrandy
Well-Known Member
You need to stop. Now I have to drill and tap more holes. It's a good thing I bought heat sinks large enough for all of these new cobs. And new drivers. Ugh
BOBBY_G
Well-Known Member
ive got a bunch, if youre looking for colors robin doesnt stock, pm me. ive got 1212s, 1818s and 1825s
Last edited:
amazing work .. amazing performance from you for all of us .. very thx
I most wonder comparison between 3500k and 3500k 90CRi 80 CRi Citizen, VERO, Cree
.. In the moment for me, probably the most considered option for the upcoming DIY build
2:1 (80CRi: 90CRi)
.. You have the possibility to get these COBs?
.. I think the use of thermal pads graphite could be a very good choice .. better than thermal grease .. excellent thermal performance bonus in the form of an adhesive layer version T62-1, good manipulation and grip with a very favorable price
If the truth stated in Cree datasheets .. could be a really good choice in thermal conductivity 15W / mKhttp://www.digikey.com/product-search/en?FV=ffec7cfb
Attachments
Last edited:
anyone can easily explain the thermal conductivity value when X-Y axis and the value in the Z-axis and the difference between them? .. I would say that commonly quoted is for real use of the Z-axis direction rather than X-Y .. what is important for contact between the heat source (COB) and heatsink only as between the layer of heat transfer due to roughness and imperfections in the material processing? The higher thermal conductivity X-Y better or if Z axis value is higher? My understanding is important that the Z-axis for instance thermal conductivity heat sink for the total intake quantity of heat for a three dimensional object = heatsink .. but if only to transfer heat to the heatsink of the COB is preferable to higher values XY therefore heat conductivity spread only on the surface and better as the lesser of Z -axis .. thus faster heat transfer to the cooler = higher value of thermal conductivity XY axis and Z axis lesser value? .. I can't read replies to this topic .. I do not have sufficient knowledge in this matter
.. it would be better to use the thermal pad T69 or T62?
.. it would be better to use the thermal pad T69 or T62?
Last edited:
BOBBY_G
Well-Known Member
heres a twist on the 3000K chart
-calculate the PPFD/$ for each measurement point based on current chip retail price
convert PPFD/W measurements to chip efficiency (when i compared with the efficiency formulas I derived from supra's "pulse" measurements i got efficiency as a %is somewhere between 5.8 and 6.8 X my PPFD measurements, so i multiplied all data by 6.6. Assumptions were made but assumptions were applied equally to all data)
-plot PPFD/$ vs efficiency
1212 leading as expected, but you'd need more chips! and if using a higher number of discrete smaller heatsinks that cost goes up too.
I can try to redo the data considering heatsink and holder and optic cost later. For now, lets all recall a much simpler time when big ass monolithic heatsinks were the thing.
-calculate the PPFD/$ for each measurement point based on current chip retail price
convert PPFD/W measurements to chip efficiency (when i compared with the efficiency formulas I derived from supra's "pulse" measurements i got efficiency as a %is somewhere between 5.8 and 6.8 X my PPFD measurements, so i multiplied all data by 6.6. Assumptions were made but assumptions were applied equally to all data)
-plot PPFD/$ vs efficiency
1212 leading as expected, but you'd need more chips! and if using a higher number of discrete smaller heatsinks that cost goes up too.
I can try to redo the data considering heatsink and holder and optic cost later. For now, lets all recall a much simpler time when big ass monolithic heatsinks were the thing.
Last edited:
Organic Miner
Well-Known Member
So, if I read thartt chart correctly ... if I drive my 1212 hard, I am getting a better PPFD/$, i.e. lower efficiency but better PPFD/$? Correct?
Rocket Soul
Well-Known Member
Jerry KB says new veros at kingbrite in August, still no prices.
bizfactory
Well-Known Member
I have not been following COBs as closely lately so got a dumb question regarding Veros..
Does Vero 29 Gen 7 = the "2.0" veros listed in the efficiency charts from Supra a while back? If so, are the "new" veros you are talking about Gen 8? And finally, what do the different letters mean when looking for the datasheets:
The only thing I could find in the model number key is possibly "array configuration" but not much else mentioned about it. Is that literally the layout of the individual diodes on the COB?
BM9AGS
Well-Known Member
So then what would multiple clu58 do?
BOBBY_G
Well-Known Member
for the most part take a point on the chart and double/triple/quadruple the wattage
i.e 4 x 1825 would be 8.8 ppfd/W @ 160W instead of 140
what you propose gets expensive. 1212s are just cheap as heck tho.
5 of the 1212s are $60 and its destroying the $46 CXB3590... almost 2.5x the watts at the same efficiency
cost per watt goes way down and youre not stuck running your expensive CXBs at 20-50W/ea
i.e 4 x 1825 would be 8.8 ppfd/W @ 160W instead of 140
what you propose gets expensive. 1212s are just cheap as heck tho.
5 of the 1212s are $60 and its destroying the $46 CXB3590... almost 2.5x the watts at the same efficiency
cost per watt goes way down and youre not stuck running your expensive CXBs at 20-50W/ea
Rocket Soul
Well-Known Member
Not sure, i dont think anyones done a proper workup on the efficiency of these new chips, as per spreadsheet but bobby_g and robin cnn have made real measurements here and maybe on some other thread.
Heres the link to datasheets:
http://www.bridgelux.com/sites/default/files/resource_media/DS93 Vero 29 Array Data Sheet Rev A 20160622.pdf
B,C&D versions: diffrent voltage and preformance on diff currents: b:50ish,c:70ish and d:38ish but they vary quite a lot with current.
BM9AGS
Well-Known Member
Gotcha....I could have figured it out if I used my brain. Basically 10 cxb 3590 at 10 watts each would crush it too.....and my bank acct
But 1212s are rocking for $60. Hmm. Are those reggies heat sinks?