DiY LEDs - How to Power Them

I suspect the 4000K AD bins may be the most efficient CXA3070s because they are only one bin step down from the 6500K BB but they are huge step down in color temp. So even though they have a lower lumen figure they may actually have a higher photon count (and a higher photosynthetic efficiency as well). In other words if there is a 3000K AB, a 4000K AD, you would expect a 5000K BB?

According to the data sheet, there is no difference in bin availability from 4000K-6500K. The AD is the top 80CRI bin...the BB is the top in 70cri no matter what temp it's in. And the 4000K comes in 70CRI BB(haven't seen it public yet). So I agree that the 4000k's could be the most efficient available. But the BB might pop up someday, and be even more efficient radiometrically, though possibly less photosynthetically than it's AD80cri counterpart.
Also...the data sheet shows the 4000K 70 cir spectrum...but the 80cri is what is available....so I wonder what exactly it will look like comparatively. I will see if I can get some spectro readings of it when I get mine going.

Supra, do you think this would be ok to power a VERO 18?

http://www.amazon.com/gp/product/B00EAXP1ZW/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1

I think you want it to be 30v or so for the vero 18. The vero 29 and cxa series require 38+ I think.

Any body with input on these:
https://www.trcelectronics.com/ecomm/pdf/plp60.pdf

FYI.. for anyone else that was looking for this kinda stuff:

http://www.fullcompass.com/product/414569.html

and https://www.google.com/search?q=IEC+Panel+Flush+Mount,+Male&gws_rd=ssl

It was hell trying to find "what" to plug into the search. Then I opened up a site with work and had a "DOH" moment.

Bueno Time said:

Got those cheap drivers today only 1 week from china, not bad. They are a little different then the ones in the picture, almost the same but the ac wires are blue instead of white and the sticker is different as well as output voltage of 39-63V instead of the 30-65V pictured (according to the sticker havent tested). Oh well, good enough to play with the two V10s I got to test, now just need to assemble them maybe tomorrow.

Click to expand...

Slapped together the string of two V10 4000K to test them. Positioned the emitters apart 4" and 4.375" on center, the distances that would be between them if I choose to run the 35 V10 setup. Good output and coverage in the middle of the emitters even close, ~3-4" away from the COB faces (though they are running ~280-300mA vs 200mA I would be running each at in the 35 V10 setup), seems like it would spread the light out real nice.

I like the color temp just by eye, looks like its got a healthy amount more blue than the CXA 3000K and looks almost pure white with a hint of yellow but mostly neutral white I would say. Going to see later tonight when the flowering tent lights kick on for the night I can plug in the V10s and see how they look side by side to the 3000K CXA so I can tell the difference better.

For now I snapped a couple pics though, nothing special. I will try to see if I can capture the CXA 3000K vs Vero 4000K spectrums tonight in some pics but it might be hard to tell since the Vero 4000K didnt look the same in most the pictures I took as it does in person but the second pic shows pretty good how white the light is.

SupraSPL said:

If you run them in parallel they would each get about 480mA. But they would need to be balanced with (above my knowledge resistor setup) otherwise some COBs could be much brighter than the others and it could be a very large difference. Also the 2.079" has a relatively thin base plate so might not be ideal for large COBs. Since it is a veg light you could use Vero 10s Vero 13s or Cree XML2/XPLs which would give you a better spread and would work better with the thin base plate.

You could use a $3 driver like this for the Veros 10s, it puts out 270-300mA and can run a very long string:
http://www.ebay.com/itm/390696103437?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

Click to expand...

thank you for your help supra. can you give me a parts list of what you would use?i dont want to order the wrong ones

Well I tested a little bit the performance of the drivers I got, paid $5.54 for two of them though the auction said $2.92 each. Drivers say DC 39-63V 300mA +/-5%. Again this was 2 Vero 10 4000K in series on one of the cheap drivers.

Kill-A-Watt #s
Initially PF .63 18.2w .23A 28.7VA
After ~1hr on PF .61 16.5w .21A 26.6VA

Multimeter Readings
COB 1 25.4 vF
COB 2 25.5 vF
Driver Current .3A (300mA) <--put out full rated current! haha

Heres the actual drivers I got, BTW according to the sticker DC + is black and DC- is red but I just didnt think that was right so I wired red as positive and it worked that way first try thankfully.

Not bad performance were the multimeter readings with the driver warmed up (the 16.6W figure) or cool (the 18.2W figure)? Also, was the multimeter in line to read current during the Kill A Watt readings? It adds some vF to the string and can make the driver appear less efficient than it really is.

16.5w warm, 18.2w cool, tested with multimeter after doing the Kill A Watt tests (had removed the KAW and plugged driver power cord direct into wall socket for multimeter readings). Took the multimeter readings after the heatsinks, COBs, and driver had been warm for a while but only tested driver current for a couple 2-3 second bursts to be safe and measured the same .3A both times.

So the drivers were probably somewhere in between warm and cool during the multimeter tests. If we estimate 17W, then it is about 89% efficient, not too shabby I love these drivers. Most of my small drivers max out at 80% efficiency and get as low as 50% when powering single 3V LEDs (for situations where you need very small amounts of indoor lught or in cloning boxes etc). It is a bummer when your LED is putting out 180lm/W with a 50% efficient driver it almost defeats the purpose.

This little driver can give us very cost efficient output $1.50-$2/PAR W with great spread and very good efficiency ~(40-44%) with Veros or it can gives us ridiculous efficiency for about $5/PAR W CXA3590 at over 200lm/W (62%), amazing to me

You can get them bare for $2.50 but I like the case even if it is wrong about positive and negative sometimes LOL.

Very good data thank you BT. Good to know the Vero4000K is noticeably cooler than the CXA 3000K I have been wondering about that. I use the CXA1512 4000K for COB headlights and the color is very good for the road. I wish they had Veros in them though I think they would perform much better than the CXA1512s. But the Veros are too big because of their casing, I wonder if there is a way to remove it.

SupraSPL said:

So the drivers were probably somewhere in between warm and cool during the multimeter tests. If we estimate 17W, then it is about 89% efficient, not too shabby I love these drivers. Most of my small drivers max out at 80% efficiency and get as low as 50% when powering single 3V LEDs (for situations where you need very small amounts of indoor lught or in cloning boxes etc). It is a bummer when your LED is putting out 180lm/W with a 50% efficient driver it almost defeats the purpose.

This little driver can give us very cost efficient output $1.50-$2/PAR W with great spread and very good efficiency ~(40-44%) with Veros or it can gives us ridiculous efficiency for about $5/PAR W CXA3590 at over 200lm/W (62%), amazing to me

You can get them bare for $2.50 but I like the case even if it is wrong about positive and negative sometimes LOL.

Click to expand...

89% efficiency is decent, too bad they arent PF corrected too though.

Which Veros would you get 40-44% efficiency with using these drivers? What bin/spectrum 3590 can you use to get 62% efficiency?

I was referring to the "typical" figures since we are mentioning Veros. The CXA3590 5000K CD bin is the 200lm/W I was referring to. I am using this driver and COB combo for vegging the taller sativa dom ladies.
Here are the updated numbers for the entire Vero series (based on approx Digikey price with a bit of shipping)



Looks like if you ran the Vero29 5000K at 300mA you could get up to 55% for $5/PAR W (for those on the nutty side, like me) and if you ran a pair of them on the driver it would drop to 270mA. The CXA3590 5000K CD runs at about 270mA on this driver and is about 62% efficient typical for about $5/PAR W.

I guess $5/PAR W is actually not that unreasonable especially for a vegging lamp. We are paying $4-$6/PAR W to run our reds and deep reds at 700mA. I paid $7/ PAR W for the Oslon SSL deep red 3T bin, seemed economical at the time but now the Cree photo reds are much more efficient and only cost $4.60/PAR W.

Wow those numbers are a lot better than the last chart you posted for me of the minimum numbers. Is this different because of the updated spreadsheets or are these the typical numbers or both?

Good info for sure. I when you mentioned the 3590 on this driver being really efficient I thought it would make a pretty good veg light if someone had an area with enough height to run it high enough to get a good spread over a decent sized canopy. Would be expensive but real efficient.

Got a couple pics taken of the CXA 3000K vs Vero 4000K, wow a big difference, the Vero 4000K is pure white. Makes a good light for taking pics of the girls.

I got two quick pictures of the Vero 4000K light shining into the tent on the girls before I failed to continue holding both heatsinks with one hand, aiming them between the tent and heatsink bar at the girls while taking pictures with my other hand. I dropped a heatsink and it bungi jumped (minus the bungi) into the tent and as the heatsink took up all the slack of the wires connected to it a sudden jerk on the wiring was felt followed by darkness. One of the wires ripped the solder pad clean off the Vero 10 lol. $5 mistake there, not too bad and glad I got all the info I needed out of the test setup in its short lifespan lol. Anyway on to the pics.

CXA3000K 80CRI


Vero 4000K 80CRI




The carnage, broken COB and a broken off fan leaf/petiole stem/bud combo


But wow I was surprised to see that much of a difference it was very clear the difference in spectrums, Vero 4000K = neutral white, CXA 3000K = warm white, IMO.

A sacrifice in the name of science, thank you sir and RIP Vero 10! Maybe you can jailbreak the Vero from the container and access some other solder pads?

Yes those pics look great under that color temp. I use 4000K-4200K Cree XML2 in my headlamps, great color temp for human vision.

The new Vero datasheets show a lower Vf curve but claim the same output so that helped a bit. Using the "typical" numbers makes a very significant difference for the Vero and not as much for the CXA because of its narrow brightness bins. We won't know if the Vero is living up to the "typical" numbers and we might get more variation if they are simply mixing in poor performers with top performers. Not a very nice approach but it allows them to offer cheaper prices it seems. And finally, there was another significant price drop since I last updated my sheets so that really opened up some design possibilities.