DiY LEDs - How to Power Them

midsummer

midsummer

Well-Known Member
nogod_ said:
There is a graph on the datasheets that shows you how efficiency is decreased as you get away from max output.
Click to expand...
I have looked over those data sheets. I have only touched on electrical concepts in my two semesters of physics with calculus in 1978. So please forgive my Noob type questions.

In the data sheet I see the efficiency (percent) vs. load (percent) graph. So if load is impedance or resistance measured in ohms. R=V/I. If I look at the voltage of the Cob (data sheet electrical characteristics graph- Cree product family data sheet CXA 3070) I see that voltage (Vf I assume) increases as the input current increases. The slope of the voltage to amperage curve appears to be less than one. If it was one I would be able to say that load is constant with these COB's (i.e. R is constant as V and I increase proportionally).

It looks as though the voltage drops slower than the amperage on that curve, so much that Vf could almost be considered constant. So that if I drops, and V is constant; R would be increasing. So that if my driver is already highly loaded then decreasing the current would have almost no effect on system Vf and load and efficiency would increase.

So to answer my own original question of "is there any penalty to running the HLG drivers at less that their max current output", I would say "No in fact efficiency increases"

Have i got that right?
 
alesh

alesh

Well-Known Member
midsummer said:
I have looked over those data sheets. I have only touched on electrical concepts in my two semesters of physics with calculus in 1978. So please forgive my Noob type questions.

In the data sheet I see the efficiency (percent) vs. load (percent) graph. So if load is impedance or resistance measured in ohms. R=V/I. If I look at the voltage of the Cob (data sheet electrical characteristics graph- Cree product family data sheet CXA 3070) I see that voltage (Vf I assume) increases as the input current increases. The slope of the voltage to amperage curve appears to be less than one. If it was one I would be able to say that load is constant with these COB's (i.e. R is constant as V and I increase proportionally).

It looks as though the voltage drops slower than the amperage on that curve, so much that Vf could almost be considered constant. So that if I drops, and V is constant; R would be increasing. So that if my driver is already highly loaded then decreasing the current would have almost no effect on system Vf and load and efficiency would increase.

So to answer my own original question of "is there any penalty to running the HLG drivers at less that their max current output", I would say "No in fact efficiency increases"

Have i got that right?
Click to expand...
Forget about R, you're dealing with semiconductors.
"is there any penalty to running the HLG drivers at less that their max current output" - yes, the driver will run slightly less efficient (about 2% decrease when run @50%), while the LEDs will run more efficient.
 
Lighterfighter

Lighterfighter

Well-Known Member
You cant really change the resistance of the led, the controllable factors are amps and votls nusing the driver to make watts/ per led. so ignore the resistance. you would say the more efficient it is, the less resistance it would have. well the less current, the less resistance of the led ( idk why volts arent kicked back as Resistance, but thats why we run ac instead of dc, that I do know ask tesla or supra lol) thus higher efficiency.
 
Lighterfighter

Lighterfighter

Well-Known Member
The curve of the veros lean towards higher current(or amps) being more efficient, and the cxa have the reverse curve.more efficient at lower current, they are most efficient running at lower watts or Amps x volts. So it isnt always the scenario you stated, just on the cxa model.
 
SupraSPL

SupraSPL

Well-Known Member
Dimming does not affect efficiency much but the voltage of the string does, here are the measurements I got from an HLG-185H-C1050A, powered by 120V mains:

driver cold, dimmer at max, with a variety of voltages on the string. 191V is a hard cap
@36.7V - 1.157A - 78.8% -> dimmed to min 33.8V - .481A - 71.6%
@108.5V - 1.155A - 90.9% -> dimmed to min 100.5V - .48A - 88.5%
@144.7V - 1.155A - 92.2% -> dimmed to min 134.4V - .48A - 90.5%
@181.6V - 1.155A - 93.14% -> dimmed to min 168.2V - .48A - 91.33%
@190.3V - 1.155A - 93.2% -> dimmed to min 175.6V - .48A - 92.32%

once the driver warmed up I did some more testing with the string fully loaded (5 CXA3070s + 3 XML2s). These are the results as I dimmed the string
190.1V - 1.155A - 93.6%
187.9V - 1.054A - 93.6%
180.9V - .705A - 93.57%
176.9V - .525A - 93%
176V - .479A - 92.54%
 
midsummer

midsummer

Well-Known Member
alesh said:
Forget about R, you're dealing with semiconductors.
"is there any penalty to running the HLG drivers at less that their max current output" - yes, the driver will run slightly less efficient (about 2% decrease when run @50%), while the LEDs will run more efficient.
Click to expand...
SupraSPL said:
Dimming does not affect efficiency much but the voltage of the string does, here are the measurements I got from an HLG-185H-C1050A, powered by 120V mains:

driver cold, dimmer at max, with a variety of voltages on the string. 191V is a hard cap
@36.7V - 1.157A - 78.8% -> dimmed to min 33.8V - .481A - 71.6%
@108.5V - 1.155A - 90.9% -> dimmed to min 100.5V - .48A - 88.5%
@144.7V - 1.155A - 92.2% -> dimmed to min 134.4V - .48A - 90.5%
@181.6V - 1.155A - 93.14% -> dimmed to min 168.2V - .48A - 91.33%
@190.3V - 1.155A - 93.2% -> dimmed to min 175.6V - .48A - 92.32%

once the driver warmed up I did some more testing with the string fully loaded (5 CXA3070s + 3 XML2s). These are the results as I dimmed the string
190.1V - 1.155A - 93.6%
187.9V - 1.054A - 93.6%
180.9V - .705A - 93.57%
176.9V - .525A - 93%
176V - .479A - 92.54%
Click to expand...
Thanks; I knew I was out of my league. They probably didn't have semiconductors in 78. Thanks guys and have a great day!
 
DonPetro

DonPetro

Well-Known Member
midsummer said:
Thanks; I knew I was out of my league. They probably didn't have semiconductors in 78. Thanks guys and have a great day!
Click to expand...
Don't get discouraged! Study the first few pages of this sticky then cruise through the rest. I certainly didnt take physics with calculus and i was able to build a panel after a few months of steady research. Dont expect it to happen over night. But when you put in the time it can be very rewarding.
 
rolledupdriver

rolledupdriver

Well-Known Member
This is all very good info, thank you, I could use some input on the setup I want to get running. I have a 1.5' x 2.5' grow space I have roughly 20 in of actual canopy space, planning on SOG type grow with good training, ventilation is not an issue. and currently I have five 43w blue/red led spots, with some bag seed growing along nicely, I wanted to know if:

x4 vero 29's
-x1 5000K
-x3 3000k​

paired with four HLG-185H-48, pushing out current at 3.9A, correct my math if I'm wrong but each of those would push out at least 22000 lumens, and give me more then enough light for my space, as I'm aware running it at that amperage would require active cooling I have plenty of spare cpu heatsinks to use for this, paired with a 400cfm intake fan right on top of them, I should be able to keep it cool enough, right? or is it overkill for that little space?
 
rolledupdriver

rolledupdriver

Well-Known Member
Lighterfighter said:
@rolledupdriver is it for flower or veg? 3 is alot for veg in that space id say. I almost would venture to say you cant over due flower, but you can def over due veg.
Click to expand...
Oh theres no way I can veg more then a few weeks, once I have the power I will probably go 12/12 from seed.

SupraSPL said:
Do I understand correctly you have 3.75 sq ft of canopy? If so, 140W would be very good intensity. If you want to get it started cheaply, a pair of Vero29 @ 2.23A would give you 160W and the drivers only $30 ea

http://www.ebay.com/itm/100W-Watt-High-Power-LED-Driver-AC90V-265V-50-60HZ-Waterproof-/181302171749
Click to expand...
Yes that's correct, Oh man that actually makes my day, so cheap, which greatly expedites the money saving phase, I can go straight to action. Thank you I will update with final result
 
Lighterfighter

Lighterfighter

Well-Known Member
ya for sure, supra hit the head on the nail. The point I was trying to make was that your plan may be over kill lol. or do like 4 cxa3070ab at lower amprage for coverage. last option would be wait for cxbs to come out in a month. if u make a thread and layout your thoughts. there are some nice charts I can find you from supra.
 
UKpeanuts

UKpeanuts

Well-Known Member
SupraSPL said:
Dimming does not affect efficiency much but the voltage of the string does, here are the measurements I got from an HLG-185H-C1050A, powered by 120V mains:

driver cold, dimmer at max, with a variety of voltages on the string. 191V is a hard cap
@36.7V - 1.157A - 78.8% -> dimmed to min 33.8V - .481A - 71.6% = 2.9V diff
@108.5V - 1.155A - 90.9% -> dimmed to min 100.5V - .48A - 88.5% = 8V diff
@144.7V - 1.155A - 92.2% -> dimmed to min 134.4V - .48A - 90.5%
@181.6V - 1.155A - 93.14% -> dimmed to min 168.2V - .48A - 91.33%
@190.3V - 1.155A - 93.2% -> dimmed to min 175.6V - .48A - 92.32%

once the driver warmed up I did some more testing with the string fully loaded (5 CXA3070s + 3 XML2s). These are the results as I dimmed the string
190.1V - 1.155A - 93.6%
187.9V - 1.054A - 93.6%
180.9V - .705A - 93.57%
176.9V - .525A - 93%
176V - .479A - 92.54%
Click to expand...
When using the dimming feature of a driver, is the output from the driver PWM?
 
SupraSPL

SupraSPL

Well-Known Member
No PWM, thankfully it is constant current dimming. If it were PWM some people could detect the flicker in your peripheral while moving. Also we would be able to see a steady lux/W while dimming which would be a real punishment to COB efficiency.
 
sforza

sforza

Well-Known Member
This is what the vendor of the CREE cxa3590 is recommending for a driver:


http://www.aliexpress.com/item/12W-LED-Driver-Transformer-free-shipping-1000pcs/557414469.html

Item description:

  • LED driver transformer for cree cxa3590 LED bulbs
  • Colour: silver
  • Material: Aluminum
  • Output Power: 100W
  • Input: 85-265V
  • Output: DC 77V
  • Dimension: 160 x 70 x 45 mm
  • output current:1200ma
  • waterproof:IP68
Is the 1.2 amps going to be pushing the COB too hard? I see .5 or .7 amps being recommended. Would the 1.2 amps cause the COB to burn our faster or are lower amps being recommended because at lower amps the light is more efficient? Is the idea that at at 1.2 amps one is not getting twice as much light as being run at .5 amps, but you are going to get more light out of the COB by running at 1.2 amps?

Is there something special about the output being 77V? If I divide 100 watts by 1.2 amps, and then multiply the result by the 95% efficieny rating, I get around 79. Is that where the 77V comes from?

I think I have seen data sheet for the CXA3590 that shows the COB being run at 77V and it is different than when the COB is run at 36V. Is that correct?

Thanks in advance for any light anyone can shed on this subject. I am learning as I go.
 
You must log in or register to reply here.
Top