DiY LEDs - How to Power Them

MrTwist1 · Oct 10, 2016

basseye said:
Thanks,so wheres the best place to buy cree cobs in the usa?

Good advice from @sixstring - I run pretty much exactly this setup in 4sqft and it works great. I would go to Cobkits.com for the Crees. good luck

BobCajun · Oct 11, 2016

SupraSPL said:
There are many ways to design a DIY LED but here are the details of the wiring for constant current drivers. If you have no experience working with 120V wiring you should understand that it can be dangerous under certain conditions and you should study and gain a thorough understanding of the dangers before attempting your project or ask for help from an electrician. The most obvious thing is never work on live wires. Once your build is complete, a GFCI can help protect you from electrical shock in our wet growing environments and are available for $7 on eBay. Check your AC circuit for hot spots which is a sign of an overloaded circuit or a poor connection. This can occur in timers and in the connections of extension cords.
==================================
Recommended tools:
Multimeter with 10A current measuring capability
Power meter - KillAWatt or EnSupra (no affiliation)
Wire stripper
Wire crimper
Slide Connectors
Heat Shrink Tubing

=========
The AC side

The AC power input has a hot and a neutral. This is important when wiring an LED driver. If you mix them up, the LEDs can glow a bit during lights out because the timer will cut the neutral instead of the hot and the LEDs may respond to ghost voltages and emf. In most homes the smaller side of the outlet is the hot wire (black) and the larger side is the neutral (white) so it is helpful to use a power cord that forces the plug into the correct polarity. If you install a switch make sure you switch the hot wire or use a double pole single throw switch. The driver should have labeling so you know which is the hot (Line) and which is the neutral. You can make the AC connections using molex connectors although that can get prohibitively expensive if you are making dozens or even hundreds of connections.

=========
The DC side

-Do not apply AC power to a driver that is not connected to an LED string
-Do not switch an LED driver on its DC side, but rather on the hot wire of the AC side.

-Constant current circuits operate with the LEDs connected in series.
-On the DC side of the driver there is a positive and a negative wire. Red is the positive and black or white is the negative.
-The negative wire of the driver goes to the negative side of the LED. (I find this unintuitive in a series connection)
-The positive of an led connects to the negative of the next LED in the series until the whole circuit makes a loop.
-If wired backward the circuit will not function and there is a possibility of damaging the LEDs.

-The driver will have a certain voltage range that it can operate within and this range may not be printed on the driver.
-The efficiency of the driver will vary depending the voltage of the LED string.
-To estimate the total voltage of the string, add the total vF or each LED in the string. If the voltage is of the LED string is too low the driver may flash. If it is too high the driver may flash or it may drastically reduce its current output.
-Each LED will operate it its own voltage but they will all operate at the exact same current.
-As each LED warms up its voltage will drop slightly.
-If you decrease the voltage load on the driver, its current output will rise and vice versa.
-As the driver warms up its current will drop (unless regulated). LED drivers are typically 75-90% efficient so they will heat up as they operate.
-As the current drops the LED voltage will drop slightly.
-Each driver may operate at a slightly different current +/-5% even if they are the same model.
-Adding a fuse that is lower than the maximum current of the LEDs may help protect the LED string in the event of a driver malfunction.
-It is critical that all connections and soldering points are secure. Flickering or arcing in the circuit can damage the driver or worse can destroy the entire LED string.

View attachment 3003741

For the DC connections you can use molex, wire nuts (not recommended), splices, closed end crimp ons or you can solder the connections and heat shrink them. In my case I use .25" bare (non insulated) crimp on slide connectors (quick disconnects) and once I have verified that the connection is very solid, I cover with heat shrink tubing leaving no metal exposed. They are cheap, secure, force the correct polarity, can be color coded and it makes your drivers and LED strings easily swappable.

View attachment 3003738

==========================
Choosing a Driver - Drive Current

LEDs suffer from current droop which means that the harder you run them (higher current), they dissipate more power but become significantly less efficient. They also suffer from temperature droop. As the temperature rises they become less efficient and if run hot they will suffer significant lumen depreciation over time. Because we use our lights for long hours each day, efficiency becomes an even more important factor. Therefore it is recommended to run LEDs relatively soft and use more of them. This increases the up front cost for the LEDs but decreases the cost of drivers, heatsinks and electricity. The value point will vary depending on electrical costs and keep in mind that we expect LEDs to continue to improve in the coming years.

=========================
Testing the String - Do this when the driver and heatsink are already warmed up and stabilized.

Once your driver and LED string is wired it is recommended that you verify the current (amperage) of the string. Very occasionally drivers malfunction and drive at a much lower or much higher current than specified. To check current your multimeter must be in series with the LED string (anywhere in the string on the DC side). Make sure your multimeter probes are plugged into the correct slot for testing current (amps). Make sure you do not apply power to the driver until the multimeter connections are secure in the LED string (slide connectors are helpful here). Record this number (example 700mA = .7A)

If you are curious about the efficiency of the driver, connect your multimeter in parallel in the LED string to measure voltage. The voltage measurement has to be taken from before the first LED and after the last LED in the string. The easiest place to do this is at the DC side driver connections. Make sure to move your test leads to the correct slots on your multimeter for measuring voltage or else you will short circuit the string. (Yes I have done this).

Once you have your volts and amps you can calculate dissipation wattage of the string. It is very simple volts X amps = watts. Next check your input wattage from the KillAWatt or EnSupra. Now divide your dissipation wattage by your input wattage and you get the driver efficiency percentage. Keep in mind that if you measure input wattage while your multimeter is in series with the string you will get a slightly incorrect figure. My multimeter adds .2-.3W.

What does it mean when you put the multimeter voltage tester across the output leads while the light is on and you get no reading at all, no matter what place in the voltage section of the meter you set it on? How can that even happen? The voltage meter works on batteries or power supplies but not on the LED driver, which is a Cree driver 240w max. The meter is Innova 3300. I put it on DCV 500, nothing, 200, nothing. Black lead in the middle hole and red lead in the right hole. Apparently 6 2530s was too much for it, because it burnt one out and flashed. Now I reduced it to 5. The driver is rated at 220w max but it also says 240W absolute max output. I guess 6 of them worked out to be over that. I neglected to notice in the data sheet that the voltage of the cobs is actually a little over 37 when they haven't heated up yet, meaning when first turned on, which is when they burn out. Unlike the 3590s, these ones run high on voltage.

Actually I just checked it again, with the dim leads shorted so it was at 10% power to be safe. On 500v nothing, on 200 it said 0.1. Seems unlikely that it would be 0.1v. BTW, I see now from this thread that you only can run 5 cobs on a 240 driver. 6 seemed possible from the math though. It actually worked for quite a while, a few weeks, but then it pooped out. Fortunately, the CXB high bay came with 12 cobs. They wired them in pairs of paralleled cobs so 6 pairs. However, that split the current so each cob only pulled 525 ma and at that lower current they pulled less voltage, about 35.25. Guess that's why it worked with 12 pairs but not with 6 in series. So far 2 got burnt out completely and a few more are about half out, but I still have enough good ones for 5 with a spare or two. Kind of lost count but I do have 5 good ones hooked up now and hopefully they'll hold up. Thank goodness for the spares though.

BobCajun · Oct 12, 2016

Okay I got the meter to work this time. Set on 500v it showed 185-188, just what it should be, drawing about 37v per cob at 1050 ma. I guess I wasn't making solid enough contact before or didn't hold it on long enough or something. So I actually only needed 5 of the 12 cobs. Glad they made it that way though, spare cobs came in very handy. I could actually turn the other high bay into a 466w one by adding appropriate drivers and maybe a stronger fan if necessary.

8/10 · Oct 13, 2016

Question:

I've got a meanwell ldd-600h. It's a dc/dc driver and I'm unsure about the power input. Should I give it more V's in than I want to get out? and what about the amps?

sanjuan · Oct 13, 2016

https://www.cdiweb.com/datasheets/meanwell/LDD-H-spec.pdf

8/10 · Oct 13, 2016

sanjuan said:
https://www.cdiweb.com/datasheets/meanwell/LDD-H-spec.pdf

I've already looked at that, but I didn't find an answer to my question.

MrTwist1 · Oct 13, 2016

8/10 said:
Question:

I've got a meanwell ldd-600h. It's a dc/dc driver and I'm unsure about the power input. Should I give it more V's in than I want to get out? and what about the amps?

Yeah u need to give it about 3 more volts than you want out. It will only draw around half an amp.

BobCajun · Oct 13, 2016

8/10 said:
I've already looked at that, but I didn't find an answer to my question.

According to that data sheet, it's a step-down converter, meaning it reduces the voltage. It says the input range is 9-56v and output is 2-52v. So supposing your LEDs draw 40v you would need to input a higher voltage than that, anything from 41v to 56v, preferably as close as possible. The current input appears to be a little less than the current output you want. Like according to the table, for 1000 ma output you would input 900 ma. At least this is my interpretation of the data sheet. Whether it works out to be cheaper or better than just buying a full AC to DC driver, I don't know. Maybe if you could get used DC power supplies for cheap.

8/10 · Oct 13, 2016

Okay, got it, thanks. I was thinking of powering this with an old laptop psu, which is 18v 5A. In addition to the driver, I also plan to wire in a dc timer, which runs on a minimum of 5V (I need to check how much power it actually draws before wiring all this). So the timer will control power from the psu to the driver. The leds draw 6-7V, I think. Like this: AC - psu (18V 5A) - timer - dc driver - leds.

It's going to be a far red "flower initiator" which turns on for 10 minutes after lights out.

BobCajun · Oct 13, 2016

8/10 said:
Okay, got it, thanks. I was thinking of powering this with an old laptop psu, which is 18v 5A. In addition to the driver, I also plan to wire in a dc timer, which runs on a minimum of 5V (I need to check how much power it actually draws before wiring all this). So the timer will control power from the psu to the driver. The leds draw 6-7V, I think. Like this: AC - psu (18V 5A) - timer - dc driver - leds.

It's going to be a far red "flower initiator" which turns on for 10 minutes after lights out.

Yeah it should work I think. Probably doesn't matter if the input current is higher than what the data sheet table says is required, as long as it's not under. Your driver puts out 600 ma. So as long as your LEDs can take that you should be fine.

BTW, I don't think the end of day far red thing actually works. I read an article where they tried various end of day spectrums and they said that the far red only made the stems longer, had no effect on flowering. It's actually just theoretical. Granted it was Chrysanthemums that they were growing, but it is a short day plant. It's conceivable that it works on Cannabis but I couldn't find any articles that actually proved it to be true. But if you want stretchier plants you will certainly get that. I tried it before and still had to wait for 10 weeks as I recall. In fact I went 75 days that time. It may be true that you can flower with 14 hours of light, but there are many strains that can flower with that day length anyway. Might not have had anything to do with the far red. They probably just never tried using 14 hours before.

8/10 · Oct 13, 2016

BobCajun said:
BTW, I don't think the end of day far red thing actually works. I read an article where they tried various end of day spectrums and they said that the far red only made the stems longer, had no effect on flowering. It's actually just theoretical. Granted it was Chrysanthemums that they were growing, but it is a short day plant. It's conceivable that it works on Cannabis but I couldn't find any articles that actually proved it to be true. But if you want stretchier plants you will certainly get that.

I'm not a fanboy of the far red theory, by any means. But I've got the time to play around with theories, so it's no loss for me. I had far reds last grow too, but one of the leds got fried halfway through the grow. Did get some unconfirmed results tho, so I'm running it again.

BobCajun · Oct 13, 2016

8/10 said:
I'm not a fanboy of the far red theory, by any means. But I've got the time to play around with theories, so it's no loss for me. I had far reds last grow too, but one of the leds got fried halfway through the grow. Did get some unconfirmed results tho, so I'm running it again.

It may actually be better to run it during the normal whole light period. Supposedly far red can increase photosynthesis that way, the Emerson Effect I think it's called. BTW, I didn't use 14 hours when I used the far red. I used regular 12/12. I wanted to see if it would speed up ripening. It didn't. If EOD far red did have an effect on flowering I should have seen ripe plants probably a week earlier than normal.

8/10 · Oct 14, 2016

BobCajun said:
It may actually be better to run it during the normal whole light period.

Yes, I've read a lot of the "far red thread". Seems like attempts at the emerson effect result mostly in stretchy plants, like you say.

BobCajun · Oct 14, 2016

8/10 said:
Yes, I've read a lot of the "far red thread". Seems like attempts at the emerson effect result mostly in stretchy plants, like you say.

Yeah but don't confuse the Emerson effect with the end of day far red effect. Two different things. The EOD would likely produce more stretching because it's pure far red, whereas Emerson is far red AND the rest of the normal spectrum at the same time.

8/10 · Oct 14, 2016

BobCajun said:
don't confuse the Emerson effect with the end of day far red effect

I know. I had EOD far red in my last grow and didn't notice a lot of stretching. It was implemented after the growth phase, tho.

BobCajun · Oct 14, 2016

8/10 said:
I know. I had EOD far red in my last grow and didn't notice a lot of stretching. It was implemented after the growth phase, tho.

It's usually only about 1/8th more stem length, so you could miss it. It's not crazy stretching just more than without the FR.

ledfan2 · Oct 15, 2016

sanjuan
sanjuan
Well-Known Member

Hi ledfan2,
If you use Ideal brand COB holders, they are only rated for 250V. I have the HLG120H-C350 and it will drive 12 36V COBs in series but that would require soldering wire to the solder pads on the ceramic square. (Unless there is another holder I don't know about.)

Edit: for slightly more money, I should have bought the HLG185H-C700 and used a parallel-series circuit.

Edit2: Ideal holders have two hole connectors for easy paralleling.

sanjuan,i didnt understand realy...

when you make the cobs on paralell strings on a HLG-185H-C700 so it have the half ampere.from 0.700 down to 0.350.that would be fine with me.

but the HLG-120H-C350 have a output 430 volt so i can put 12x36 volt cobs on it(maybe 13) in seriell.i m right that the HLG-185H-C700 with maximum output 286 volt can only drive 8 cobs?

or im wrong and:the volts halfed too? im sure not!
or im wrong with the max. output 286 volt and this going related higher when the ampere go down?

ledfan2 · Oct 15, 2016

and still i have this question:what is better for dimming cause i have read that the pwm take led on of 100 times in a second to half the light and that take lifetime from led?so what is the better way to dimm respectively what is the lowest current losses way to dimm? the driver typ a have a potentiometer,typ b have 3 in 1 dimming function-1-10vdc,10v pwm signal and resistance.

if i choose a resistance so the power i reduce is lost at the resistance.i pay for the power that comes not to the cobs,just only heating the resistance!?

the potentiometer i didn t really understand,just it likes for me a little bit like a resistance?

the pwm take the cobs on and off many times in a second,so i read that reduce the led lifetime.some says not much,other says its relevant!

And first time i read about the cree cobs her i wondering about the high kelvin 3500,the old shool hps with 2100-2700 kelvin is so many years people said the best.

with pwm dimming should the kelvin from a led not change,but with other dimming it should changing down.maybe if its true so is the pwm twice time not so good--reduce life time but not the kelvin(to better spectrum)

so in moment i would not take the poti,resistance,pwm...only left is 1-10vdc....so wat is the best dimming in context of lowest power loss???

some one knows about the story that pwm change not the colour temps? but reduce lifetime? i think thats interesting questions! because it concern efficency and the best kelvin.an if that true so should more dimming change the the light colour more red,lower kelvin...

(again,sorry the bad english!)

BobCajun · Oct 15, 2016

ledfan2 said:
sanjuan
sanjuan
Well-Known Member

Hi ledfan2,
If you use Ideal brand COB holders, they are only rated for 250V. I have the HLG120H-C350 and it will drive 12 36V COBs in series but that would require soldering wire to the solder pads on the ceramic square. (Unless there is another holder I don't know about.)

Edit: for slightly more money, I should have bought the HLG185H-C700 and used a parallel-series circuit.

Edit2: Ideal holders have two hole connectors for easy paralleling.

sanjuan,i didnt understand realy...

when you make the cobs on paralell strings on a HLG-185H-C700 so it have the half ampere.from 0.700 down to 0.350.that would be fine with me.

but the HLG-120H-C350 have a output 430 volt so i can put 12x36 volt cobs on it(maybe 13) in seriell.i m right that the HLG-185H-C700 with maximum output 286 volt can only drive 8 cobs?

or im wrong and:the volts halfed too? im sure not!
or im wrong with the max. output 286 volt and this going related higher when the ampere go down?

The thing only puts out 150w. 12 cobs would be 10.25w each. Might as well just buy a bunch of LED light bulbs. Cost a lot less than $40 each. Granted less efficient but not THAT much less. You can get 8.5w bulbs for about $3 each.

BobCajun · Oct 15, 2016

You'd be paying $40 each for cobs putting out about 10w each. Kind of an expensive way to go isn't it? You can get 8.5w led bulbs for about $3 each. It's your money. I can answer that paralleling doesn't half the voltage though, just the current. I'll leave the rest to others.

One other thing though. PWM does not make the LEDs have shorter life. It's actually the opposite. Their average temperature is lower because the current is off or down to 10% during the off phases of the PWM so they last longer.