DiY LEDs - How to Power Them

Greengenes707

Well-Known Member
Got a UK quote for HLG 240H C 2100b, long delay and high price tho ..
but looks like it's coming in :)
View attachment 3464855
So 3 CXB 3590 3500K CD + 1 HLG 240 (225W) beats 4 CXB3070 AD + HLG 185 on just about all fronts, except spread ...
More watts / better efficiency, for +- equal prices.
Just something that I keep noticing, not just specifically here...the 3070 3500K BB would be the one to compare with, not the 3K AD. Apples to apples.

Things will vary, supplier to supplier, google to google. But in general from what I see out there and know from my own buying...based on my calcs with the equal corresponding 3070(3500K BB)...the 3070 would be cheaper(overall and per PAR watt), higher efficiency, and better spread than the 3590 in that situation.

EDIT: I see your edit now that I posted...I like that 1750ma...pct is saying the 3590 will run ~35v there though. Do you have a link to the spec sheet...google is failing me.
 
Last edited:

SupraSPL

Well-Known Member
Not sure if they are available to the public or where, but the comparison would depend on price. The 3590 is about 5% higher efficiency. In my mind that means they are equal in terms of bin. For example, if we got a bin increase to BD 3070 someday, then we could expect an increase to DB 3590.

Another way to think of it, at 50W the dies inside the 3590 are running softer than the dies in a 3070 running at 50W. They are same exact dies in terms of performance, but the 3590 has more of them. So the reason the 3590 gets higher efficiency is mostly due to reduced current droop.
CXB 3500K.png
 
Last edited:

Greengenes707

Well-Known Member
Not sure if they are available to the public or where, but the comparison would depend on price. The 3590 is about 5% higher efficiency. In my mind that means they are equal in terms of bin. For example, if we got a bin increase to BD 3070 someday, then we could expect an increase to DB 3590.

Another way to think of it, at 50W the dies inside the 3590 are running softer than the dies in a 3070 running at 50W. They are same exact dies in terms of performance, but the 3590 has more of them. So the reason the 3590 gets higher efficiency is mostly due to reduced current droop.
View attachment 3464990
Yep...there are just more dies, ran softer per die.
And think how much shit we all used to talk about chips not being driven high enough...now clearly flawed thinking, hello COBs. We should have been using the cheapest possible and retardedly underdriving them, and then packing them so tight...that we would actually of had very solid lights and penetration would have been there despite low power per die.
Life I guess.

SIDE NOTE: The PCT and your charts(data sheet based) are off by a good bit...more discrepancies imagine that. I know we are "ok" with it...just pointing out it is there again and bugs me. I'm going to try to track down a sphere to figure this out...no guarantees I can find one to use.
 

SupraSPL

Well-Known Member
Ya unfortunately the CXB3590 data sheet gives us less droop info for running soft so I had to extrapolate the 23W data from the CXA3590 sheet. On top of that, I am not sure exactly how Tj translates to Tc for each given current so there is a fudge factor there as well. I have more confidence in the the older data sheets that showed separate charts for current droop based on Tj and temp droop based on Tj.
 

Bhookus

Member
What was the space getting lit? I looked back 4 pages and didn't see us...

The difference in amps is just how hard the leds are running, less output/ efficient, or more output/less efficient (more heat per watt)

Voltage is just capacity, so find your amps you want to run at (supras awesome spreadsheets for the given bin) and then find an hlg that fits the total voltage of leds driven at whichever current. Try to fill the capacity of each driver (more efficient)

Edit: The 3070 and 3590 will put out different watts at the same current
Yeah, it was awhile ago.
As far as the space, it's 3 panels over 12'x40" on a light mover.
I see my mistake after looking at supras spreadsheet. But it looks like other people's setups shouldn't work either. I must be missing something, I'll get back at it and see if I can find another mistake I'm making.
I've got 2 cxa3590 running per panel now and I like the look.
Any thoughts on adding uv stars or extra red? It would be in my favor to add some and use up any extra volts?
 
Last edited:

Scotch089

Well-Known Member
12 feet wide?? I may have mistaken the original post and thought you meant 12" deep idk. I don't have experience with light movers but if you see dark spots. Light them up. My initial thought is more light. But again I'm limited here. Better to ask supra greengenes or alesh. 4 3070s 4 3590s in a 12 x3.5' area...... doesn't sound like enough to me. I'm sorry if I confused you early on I'm no expert but that may have just been misreading or assuming a typo.

As for reds I'm a fan, but not at all a neccisity, and positivity has had great success implementing uvb, always wanted to but never gotten around to it. For reds I'm liking the look of a 5:1 watt ratio of white:red, which is close to what the sgs ran. (I'm ~100w of white with 20w red, sgs was 120 and 30)
 

Bhookus

Member
12 feet wide?? I may have mistaken the original post and thought you meant 12" deep idk. I don't have experience with light movers but if you see dark spots. Light them up. My initial thought is more light. But again I'm limited here. Better to ask supra greengenes or alesh. 4 3070s 4 3590s in a 12 x3.5' area...... doesn't sound like enough to me. I'm sorry if I confused you early on I'm no expert but that may have just been misreading or assuming a typo.

As for reds I'm a fan, but not at all a neccisity, and positivity has had great success implementing uvb, always wanted to but never gotten around to it. For reds I'm liking the look of a 5:1 watt ratio of white:red, which is close to what the sgs ran. (I'm ~100w of white with 20w red, sgs was 120 and 30)
My bad- 3 panels of 8 cobs each over 12ft by 3.5ft
The light mover is moving each panel about >30". It's got great coverage, no dark spots, just a solid path of light. All the shadows moving around opening up the canopy.
Please excuse my ignorance - who/what is sgs?
Does the extra red improve frost only or are their other benefits? Do I need to stay away from going over 680nm for the best frost?
 

Bhookus

Member
(The sgs160 by area 51)

Blue tends to give more frost then red. Similar to uvb, not to put a direct comparison to the two- I've never ran uvb. But that'd why back in the day of RB ufos ppl were noticing "leds give more potency!"

A really good read on colors... I wish there were more of THESE kind of threads.

https://www.rollitup.org/t/photosynthesis-under-solid-state-light-setting-the-standards.833449/
I'm an salty dog from the aquarium industry. I have a ton of actinic flourescent kicking around. Should I fire up some of those on this run until I can get some added to led panels?
 

Sevren

Well-Known Member
Quick math check for those fluent. Spent most of the morning/afternoon hunting down all the sheets for correct figures, all but one, relying on Cree PCT for the CXB BB 4000k

Ran at 1400mA:
Assuming the LER for the 4k is 323

CXA AB 3000k - 52.22W x 4 = 208.88W Efficiency = 43.31%
CXB AD 3500k - 49.83W x 2 = 99.66W Efficiency = 48.02%
CXB BB 4000k - 49.315W x 1 = 49.315W Efficiency = 153.9 / 323 = 47.64%
total Dissipation Watt - 357.855

3000k - 208.88 x .4331 = 90.4659 PARwatts
3500k - 99.66 x .4802 = 47.8567 PARwatts
4000k - 49.315 x .4764 = 23.4936 PARwatts
total PARwatts - 161.8162

161.8162 / 8 = 20.227 PAR FT2

161.8162 X 4.88 = 789.663 PPF
 

Bhookus

Member
I was looking at rapidled and they have the semiled total uv. Anybody have any experience with these? It states a max of 1000mA, is this a hard cap or could they run @1050mA?
 

PurpleBuz

Well-Known Member
I have a bunch of the semileds, they work, check out the 4 in one chips they have, gives ya a wider wavelength range. so far I'm running them at 550 mA just because I have a bunch of cheap drivers I'm recycling from failed Chinese units

should be able to run at 1050 if you have enough cooling, if your worried get a dimmable ballast and run just a bit under max. (like HLG-185H-C1050A)
 

Bhookus

Member
I have a bunch of the semileds, they work, check out the 4 in one chips they have, gives ya a wider wavelength range. so far I'm running them at 550 mA just because I have a bunch of cheap drivers I'm recycling from failed Chinese units

should be able to run at 1050 if you have enough cooling, if your worried get a dimmable ballast and run just a bit under max. (like HLG-185H-C1050A)
I already have those drivers. I goofed up my math on my build, so I have room to use up.
Thanks for the info. The 4in1 is what I was looking at. I thought I could adjust down, but needed a little confirmation. I
 

Kempo

Active Member
Hi, i have a problem, i plugged my 4 alpines on a netbook power supply in parallel (all blacks wires together and same for red). Its seem like i burnt them :(. I dont understand

 

littlejacob

Well-Known Member
Bonjour
Maybe you put positive on negative and vice versa, and i believe they do not have protection against it.Or maybe something else! ?.
But wait for someone who have more experience
Have a great day ★
 

poxa

Member
Hi guys,

thanks all so much and Supra particularly, for the thread and all the information provided, impressive.
I'm a newbie about the subject trying to figure out from scratch what should I do for a (2' x 2') setup.

I was considering before starting if it is possible to mix different COBs in a single rail and have all them driven - for example - by a single HLG-185H-C700B (driver must dimmable by a 100μΩ PM).

(2 x CXB3070) + (2 x CXB3590)
or
(2 x Vero29) + (2 x CXB3590)

If some specific information is available - please, point me out without mercy and forgive the noob in advance!

Thanks again!
 
Last edited:
Top