DiY LEDs - How to Power Them

alesh

Well-Known Member
@SupraSPL Can you be a little more specific about different Vf's please? I've measured a few dozens single die Cree LEDs and the differences were a few hundreths of Volt. Since in 3070 there's a lot of dies and they have to be exactly same Vf, I can imagine it could add up to around 3-4V?? Just a guess, correct me please.

Do you know of any driver with PWM dimming which actually does PWM dimming? I mean, MW drivers for example, do accept PWM input. However they reduce output according to duty cycle. Yea it's more efficient for human vision application. But IIRC knna stated that true PWM output could be beneficial even with the loses from higher current. And remember that he used Luxeons K2's and then Golden Dragons. These COBs nowadays are much more tolerant in terms of high current and temperature. Perhaps this could be a way to achieve even greater results.
 

SupraSPL

Well-Known Member
Good questions Alesh. The differences in Vf may be more noticeable from batch to batch, but they can be significant even in one batch. I got a batch of CXA3070 Z4s that were generally very high Vf and a batch of Z2s that were generally low Vf. I had 40 CXA3070 ABs hooked up to 40 of the same drivers and I was cutting it so close on the Vf range that I had to have a customer swap a COB to another driver because that particular COB had too high of a Vf for that particular driver. Swapping to a different COB actually solved the problem.

KNNA was opposed to PWM dimming vs constant current dimming. It is true that PWM dimming is a way to save power versus no dimming at all, but it is much less efficient than CC dimming/direct dimming. This is especially true for blue/white COBs which are tough against temp droop but duffer quite a bit from current droop.

I have noticed some car manufacturers have started using PWM dimming in the red LED tail lights/brake lights. I was bummed to see that because not only is the flickering annoying and possibly dangerous, but it demonstrates the engineers failing to maximize the benefit of LED efficiency. On the upside, some red LEDs can deal with the current droop very well while they tend to suffer from temp droop instead.
 
Last edited:

HiloReign

Well-Known Member
Hi Supra. It's great to see you still jamming around here. I wanted to express my appreciation for all your hard work and enthusiasm, so that mindless fools like me can manipulate your work and benefit from it greatly (not monetarily :eyesmoke:).
 
Last edited:

epicfail

Well-Known Member
Do you know of any driver with PWM dimming which actually does PWM dimming?
Mean Well introduced the "PWM Series" recently, It might be what your looking for
  • Wattage - 40W~120W
  • 90~305VAC input, built-in active PFC function
  • Fully isolated plastic case with IP67 level
  • Constant voltage PWM style output
  • Class II design, without FG
  • No load power consumption<0.5W
  • Built-in 2 in 1 dimming function (0~10VDC or PWM signal)
  • Suitable for LED strip lighting
  • 5 years warranty
http://www.meanwell.com/webapp/product/search.aspx?prod=PWM-120
http://www.meanwell.com/webapp/product/search.aspx?prod=PWM-90
http://www.meanwell.com/webapp/product/search.aspx?prod=PWM-60
http://www.meanwell.com/webapp/product/search.aspx?prod=PWM-40
 

alesh

Well-Known Member
Thanks for the link EF, not exactly what I had in mind though.

Good questions Alesh. The differences in Vf may be more noticeable from batch to batch, but they can be significant even in one batch. I got a batch of CXA3070 Z4s that were generally very high Vf and a batch of Z2s that were generally low Vf. I had 40 CXA3070 ABs hooked up to 40 of the same drivers and I was cutting it so close on the Vf range that I had to have a customer swap a COB to another driver because that particular COB had too high of a Vf for that particular driver. Swapping to a different COB actually solved the problem.

KNNA was opposed to PWM dimming vs constant current dimming. It is true that PWM dimming is a way to save power versus no dimming at all, but it is much less efficient than CC dimming/direct dimming. This is especially true for blue/white COBs which are tough against temp droop but duffer quite a bit from current droop.

I have noticed some car manufacturers have started using PWM dimming in the red LED tail lights/brake lights. I was bummed to see that because not only is the flickering annoying and possibly dangerous, but it demonstrates the engineers failing to maximize the benefit of LED efficiency. On the upside, some red LEDs can deal with the current droop very well while they tend to suffer from temp droop instead.
I did a little digging and found what I meant (dated 2006!!). The theory is well explained in the link.
The thing is that I got it confused and knna actually stated:
knna said:
Not, electrical savings on using LEDs not depend of PWM. Its just another hypothetical way of obtain electric savings. As the electric savings involved are large, it worth experimenting deeper in this topic, but in fact until now nobody has been capable of prove it enhances photosynthesis efficiency. [source]
My bad. It still might be interesting phenomenon to test.
 

Tazbud

Well-Known Member
Dimming is still leaving me guessing, any thoughts on using this knob:

Screen Shot 2014-12-12 at 2.58.41 AM.png

To efficiently dim this driver:

Screen Shot 2014-12-12 at 2.53.05 AM.png

(iirc it was suggested for another MW driver here...)

:peace:
 
Last edited:

SupraSPL

Well-Known Member
I chose which color for which pin arbitrarily in this case, but it should work for this job either way.
 
Last edited:

lapin

Active Member
Not sure but looking into it. Probably would not hurt to start a new one with updated info at the beginning though.
apologies if i brought the evil skies to shut that thread down with my latest build. perhaps i should just start a new thread.
 

SupraSPL

Well-Known Member
1.5A driver review. Took about 3 weeks to arrive from China. Very cheap $13 ea. Power factor corrected and put out 1450mA warmed up. Surprisingly high Vf range, up to 47V (lot of COB dissipation from this driver, 68W, drawing 80W) There is one downside though, at 37.5Vf it was only 86-87% efficient. At 47Vf only 85% efficienct. So it is not up to par with our 90-91% efficienct perfectdeal_us drivers that are temporarily sold out.

http://www.ebay.com/itm/381056054738?_trksid=p2059210.m2749.l2649&ssPageName=STRK:MEBIDX:IT
driver.JPG
 

SupraSPL

Well-Known Member
In this case the price for each driver is the same, $13. But for those who have their drivers mounted onto their heatsinks, the extra (3.5) watts per driver from poor driver efficiency becomes pure heat in their heatsink and grow space. In my case I have remote drivers outside the tent, but the driver heat still goes into circulation and heats up my incoming air temp.

For this board, if I had used 86% drivers instead of 91%, the entire build would be 70W hotter, 10% more heat. The electrical savings would be about $3.80/month or $45/year, not counting any increase in air conditioning use or ventilation requirements.
DSC07251a.jpg

Also, those extra watts create extra strain on the wiring and timers, especially if the drivers are not power factor corrected. The 91% efficient build will draw 9.5A and the 86% efficient would draw 10.1A, an increase of 6.1%. Not a huge deal but worth considering.

Another way to look at it, if you used 91% efficient drivers instead of 86%, that would have the same effect on your system efficiency as a bin increase, 6.1% increase in light/W. A bin increase for Cree CXA is about 7%.
 
Last edited:
Top