First CXB3590 - need a little advice

welight

welight

Well-Known Member
sixstring2112 said:
But what if the new cree are 60 bucks a chip,should he still wait?
Or what if the new lower volt cree drops prices on the older 3590 to where they are as cheap as veros and shitizen,what cob would he want then ?
I want a new pc but im waiting for windows 20 :roll:
Get the goodz and get ya growin :weed:
Click to expand...
price wont change on 3590 for lower volts, only volume gets a price change, order a 1000 today and I will get you CXB Db in 3500k for 28.50 each:cool:
Cheers
Mark
 
BurgerBro

BurgerBro

Member
CobKits said:
the only other advice i would give is if youre married to cree wait it out till @welight has the newer low voltage versions

https://www.rollitup.org/t/cree-reduce-forward-volt-on-cxb30-35-series.938864/

those should hang with the best of the best citis and veros out there
Click to expand...
Wait... Are you saying that current Vero and Citizen COBs are more efficient than CXB3590s when run at low current? My understanding is that they had recently become as efficient as Crees, provided that both are run at 1.4 ~ 2 Amps.

I am not married to any brand. I only want what is best for my particular situation.
 
CobKits

CobKits

Well-Known Member
BurgerBro said:
Wait... Are you saying that current Vero and Citizen COBs are more efficient than CXB3590s when run at low current?
Click to expand...
yes, some are. there are all different sizes of citis and veros. usually anything over $15-$20 will be as efficient/more efficient than the current cxb3590. im hoping the new crees are in the clu058 gen5, vero29C, and luminus cxm22 range from 30-75W.
 
Last edited:
BurgerBro

BurgerBro

Member
CobKits said:
yes, some are. there are all different sizes of citis and veros. usually anything over $15-$20 will be as efficient/more efficient than the current cxb3590. im hoping the new crees are in the clu058 gen5, vero29C, and luminus cxm22 range from 30-75W.
Click to expand...
Well, now you've really sparked my interest. If that is true, it opens up a lot more possibilities. I could get 10 COBs instead of 6 and run them at ~25W instead of ~38W, or I could get different ones for flower and veg, and the price would not get too much higher, or might even stay the same...

What would you suggest to fill out an HLG-240H-C1050B or HLG-240H-C700B that would be a more efficient setup than the CXB3590s I was planning to buy? I don't mind the purchasing price too, too much, but wall efficiency is really important. Keep in mind that I pay a very high price for electricity.

And thanks for helping out!
 
CobKits

CobKits

Well-Known Member
at 25W citi 1212s are good chips an are $10-$12 each. they are great chips for those low currents, not the top dog in efficiency but theyre cheap and you can stack up more

luminus cxm22 are better than the CXB3590 across the board and are about $17 each

10 of either of those would give you more light than 6 CXB. 10 luminus would give *a lot* more light, maybe 10-15% more efficient than 6 CXB running the same wattage.
 
SupraSPL

SupraSPL

Well-Known Member
I have been out of the loop a bit but I have worked up the numbers for the large Citizens and Samsung S6 boards, looking for something that might top the efficiency of the CREE CXB3590. So far I have not found it, but I am searching for it so please bring me up to speed if I am missing something.

Going by CREE's 50W PDF data, the CXB3590 3500K DB at Tj 50C is 197lm/W (typical) 2.83umol/s/W, for $1.28/PAR W. Tj 50C should be no problem for an active cooled setup.

I think most growers should be looking in the ~50W/COB range, where the CXB3590 might have some decent competition. But the CXB3590 really excels at reaching very high lumens/W at LOW current and that is what I need for my setup. The numbers in this chart below 50W are extrapolated. I run CXB3590s as low as 11W when dimmed in veg or at the end of flowering but most of the time they are at 23W which cost me $2.40/PAR W (Thank you @REALSTYLES )

CXB3590 3500K DB.png
 
Last edited:
welight

welight

Well-Known Member
SupraSPL said:
I have been out of the loop a bit but I have worked up the numbers for the large Citizens and Samsung S6 boards, looking for something that might top the efficiency of the CREE CXB3590. So far I have not found it, but I am searching for it so please bring me up to speed if I am missing something.

Going by CREE's 50W PDF data, the CXB3590 3500K DB at Tj 50C is 197lm/W (typical) 2.83umol/s/W, for $1.28/PAR W. Tj 50C should be no problem for an active cooled setup.

That said what the CXB3590 does best is reach very high lumens/W at low current. The numbers in this chart are extrapolated below 50W. I do run my CXB3590s as low as 11W when dimmed in veg or at the end of flowering so I am interested in the low current performance.

View attachment 3935725
Click to expand...
Nice to see your input again SupraSPL ,its been too long
Cheers
Mark
 
SupraSPL

SupraSPL

Well-Known Member
Regarding the accuracy of this chart, I used the PDF data and extended the droop curves as best I could. CREE has always been known to be top notch and hit their marks, going back to the days of CandlePowerForums flashlight modding so I do not doubt their capability. I do recognize that there are some weird inconsistencies with some of the manufacturer's PDF data, especially when it comes to Vero. Although Vero tests very well in real world conditions, it does not gain as much efficiency as the Cree does at low currents. The Vero gen 7 seems very promising to me in terms of bang for the buck setups.
 
SupraSPL

SupraSPL

Well-Known Member
Interestingly, I saw Bridgelux made a negative comment about running COBs at low currents, they advise not to do it because results are unpredictable see document below. Maybe they are worried about variable color temp. To add insult to injury, they recommend dimming using PWM. PWM dimming means to flicker the light on and off at full power rather than to simply reduce the current. Using PWM you would miss out on any efficiency gains due to the effects of current droop. PWM dimming also causes a flicker that could be a health issue.

Page 17
http://www.bridgelux.com/sites/default/files/resource_media/DS93_RevD Vero 29 Array Data Sheet.pdf

"Bridgelux does not recommend driving high power LEDs at low currents. Doing so may produce unpredictable results. Pulse width modulation (PWM) is recommended for dimming effects"

I am not trying to pick on Bridgelux just busting balls, they are probably marketing these lights toward corporate signage. If I recall CREE had the same attitude regarding running a 150W COB at 25W when @Greengenes707 quizzed them about it. Sorry for the off topic OP, I am holding my tongue regarding the CREE Horticulture Reference Design!
 
Last edited:
C

churchhaze

Well-Known Member
SupraSPL said:
Interestingly, I saw Bridgelux made a negative comment about running COBs at low currents, they advise not to do it because results are unpredictable see document below. Maybe they are worried about variable color temp. To add insult to injury, they recommend dimming using PWM. PWM dimming means to flicker the light on and off at full power rather than to simply reduce the current. Using PWM you would miss out on any efficiency gains due to the effects of current droop. PWM dimming also causes a flicker that could be a health issue.

Page 17
http://www.bridgelux.com/sites/default/files/resource_media/DS93_RevD Vero 29 Array Data Sheet.pdf

"Bridgelux does not recommend driving high power LEDs at low currents. Doing so may produce unpredictable results. Pulse width modulation (PWM) is recommended for dimming effects"

I am not trying to pick on Bridgelux just busting balls, they are probably marketing these lights toward corporate signage. If I recall CREE had the same attitude regarding running a 150W COB at 25W when @Greengenes707 quizzed them about it. Sorry for the off topic OP, I am holding my tongue regarding the CREE Horticulture Reference Design!
Click to expand...
The loss of efficiency is why it makes no sense to me to run plant lights pulsed. Pulsing makes a lot more sense for direct lighting meant for humans. (like LED signs) as the brain can be fooled into thinking even very short pulses are continuous bright light. I'm sure that their suggestion is intended for human lighting.
 
CobKits

CobKits

Well-Known Member
SupraSPL said:
I have been out of the loop a bit but I have worked up the numbers for the large Citizens and Samsung S6 boards, looking for something that might top the efficiency of the CREE CXB3590. So far I have not found it, but I am searching for it so please bring me up to speed if I am missing something.
Click to expand...
recent test in 20" sphere with LI-cor PAR meter. this was done on two different days so there are some temp variations. gonna redo with more chips when i can

day1(cooler - better performance): CXB3590A, 1216, 1818
day2 (warmer- slight performance penalty:CXB3590B (same chip as A), Vero22C, luminus CXM22, Luminus CLM22

upload_2017-5-3_18-16-30.png

as you can see most all of those chips run close at low currents. some better options than CXB3590:
1818 gen 6 ($24) - about the same as CXB from 35-50W, more efficient than CXB above 50W, about 1% less than CXB at 23W
luminus CXM22 ($17) - 2-4% more efficient than CXB at all currents

cree def a strong performer at low currents (esp say relative to vero22) but not sure cost is justified vs say 1818 or luminus CXM22
 
SupraSPL

SupraSPL

Well-Known Member
Good info thank you for that hard work! Now I have to try this CXM22. I will also do a thermal test to see if that backs up my PAR measurements. I love this stuff so I hope you dont mind I have a bunch of questions, I will move over to the other thread though.
 
SupraSPL

SupraSPL

Well-Known Member
When I first attempted to graph the data from my PAR meter measurements I decided to use PPFD/W @ 12" for several reasons. The sensor method I used (Apogee PAR meter open air @ 12") measures PPFD (photons/area/second) rather than PPF (total photon output/second). In order to measure PPF you would have to run the test in a calibrated integrating sphere.

I carefully monitored the dissipation power so I could use PPFD/W to level the playing field. This way we can visualize the performance of many different COBs on one simple graph, in terms of efficiency/efficacy rather than in terms of total output.

The tests were performed in a large room at room temperature with a constant 12" distance between the COB and the sensor, so theoretically you could reproduce and compare results from separate labs and separate test sessions.

"I used adjustable constant current drivers and monitored Vf and If with Fluke and Amprobe multimeters. The dissipation Wattage was divided by the PPFD reading to arrive at the awkward but useful "PPFD/W @ 12""
 
Last edited:
C

churchhaze

Well-Known Member
SupraSPL said:
When I first attempted to graph the data from my PAR meter measurements I decided to use PPFD/W @ 12" for several reasons. The sensor method I used (Apogee PAR meter open air @ 12") measures PPFD (photons/area/second) rather than PPF (total photon output/second). In order to measure PPF you would have to run the test in a calibrated integrating sphere.

I carefully monitored the dissipation power so I could use PPFD/W to level the playing field. This way we can visualize the performance of many different COBs on one simple graph, in terms of efficiency/efficacy rather than in terms of total output.

The tests were performed in a large room at room temperature with a constant 12" distance between the COB and the sensor, so theoretically you could reproduce and compare results from separate labs and separate test sessions.

"I used adjustable constant current drivers and monitored Vf and If with Fluke and Amprobe multimeters. The dissipation Wattage was divided by the PPFD reading to arrive at the awkward but useful "PPFD/W @ 12""
Click to expand...
Everyone measures PPFD in units of PPFD (I have 800 PPFD). I don't understand why, and it drives me nuts... but I see no signs of it stopping.

To cope, I will now smoke 1 marijuana of marijuana.
 
Last edited:
BurgerBro

BurgerBro

Member
CobKits said:
recent test in 20" sphere with LI-cor PAR meter. this was done on two different days so there are some temp variations. gonna redo with more chips when i can

day1(cooler - better performance): CXB3590A, 1216, 1818
day2 (warmer- slight performance penalty:CXB3590B (same chip as A), Vero22C, luminus CXM22, Luminus CLM22

View attachment 3935918

as you can see most all of those chips run close at low currents. some better options than CXB3590:
1818 gen 6 ($24) - about the same as CXB from 35-50W, more efficient than CXB above 50W, about 1% less than CXB at 23W
luminus CXM22 ($17) - 2-4% more efficient than CXB at all currents

cree def a strong performer at low currents (esp say relative to vero22) but not sure cost is justified vs say 1818 or luminus CXM22
Click to expand...
This chart is pure gold! Do you have any more data like that? For instance, the VERO 29 and clu058 are sorely missing from the chart.

Since we last spoke (months ago, life happened), I have been considering going with a BXRC-30G10K0-D-7X build. That is the VERO 29 3000K, 90 CRi, 37.6v COB.
After seeing your chart, I really want to go with the Luminus CXM22 that you keep mentioning. That is a no-brainer. But I don't live in the Americas, and I don't know who sells them in batches of less than 400 in Europe or Asia.

The reason I keep mentioning efficiency (I probably mean efficacy) is that I pay around 30 American cents per kilowatt/hour... That little piece of information probably makes my obsession with energy efficiency more understandable.

So my current idea is 9 BXRC-30G10K0-D-7X COBs on an HLG-240H-C700B. Does that seem sensible?

SupraSPL said:
Sorry for the off topic OP, I am holding my tongue regarding the CREE Horticulture Reference Design!
Click to expand...
Oh, you don't have to apologize, SupraSPL. You are a true legend in my book, and I am always interested in your opinion!
 
CobKits

CobKits

Well-Known Member
BurgerBro said:
This chart is pure gold! Do you have any more data like that? For instance, the VERO 29 and clu058 are sorely missing from the chart.
Click to expand...
correct i never got around to testing larger chips but the sphere data fell right in line with my open-air on-axis testing at 12" so i know where they fall. vero C is right on top of the 1825 gen5

upload_2017-7-10_11-4-15.png
 
You must log in or register to reply here.
Top