LED lighting on a budget!

[AcroSteve]

Nice work.

If you look around you will find that 24VDC is a fairly common size that can be had for very little money. We'll work with that but most any voltage will do the trick.

With all this concern about efficiency of the LED system, have you looked into the efficiency of the power supplies themselves?

I would imagine that there is a great range of differences. I have a industrial one at home(I think it is a Lambda) that certainly looks nice and impressive. I would assume it would be more efficient than some wall wart type ell cheapo, but I just wonder.

Newark does post some efficiency numbers for some, but not all. Any particular brand you know is very good?

 

[AcroSteve]

D) But what about efficiency, isn't this setup likely to be very inefficient?

We are looking at 190W to provide 147W of LED light. 147W/190W => 77%
BuckPucks land around 79% with this same design.
Mean Well efficiency is about 82%

If you want to get crazy about efficiency you adjust your supply voltage so that you need even less voltage across our power resistors. If you do this correctly you can beat the Mean Wells by a few percent, but I don't think it's required and it will reduce your system flexibility a little.

Where did the 190W come from? Is that the actual wattage used buy the power supply on the 120v circuit? Or is that 190w just on the 24v side of the power supply?



Back to the efficiency question on the 24v power supply. Would the efficiency of a given power supply be linear over the output range? Or does there tend to be a sweet spot most of the time?

 

[Danielrvb]

I am working on doing this build right now, and am a little confused one part actually putting into practice, and just want to confirm it.

I dont seem to understand the volt measurements I am getting after the resistor. If I hook up power supply straight to resistor and measure volts with nothing attached, it gives me the same as if there was no resistor.

Then when I am hooking LED's up in line, it changes the volts reading to 8ish. ??????

I have measure the amps going over, and that is what I have been using. I am setting them up to run between .7 to .8 amps. I feel that range should be good with out trying to get to picky.

My main question is to make sure I will be ok with just measuring amps of the string to set my resistor? as I am not getting any logical reading from volts, or I am confused at least.

 

[Danielrvb]

I am running strings of 5. Using an old laptop power supply.. 19v. 3.16 amp.. plan is 3 strings of 5 running at .7ish amps. I have XPG's and RB XRe's..
I do plan to run the RB's a little higher to try and catch up so to speak with the XPG's brightness.

 

[kcress]

AcroSteve; That 190W includes the PS efficiency. That would be what was sucking from the outlet teat.

As for 'linearity over range'. Well, um, ah. It's hard to say, because it depends on all sorts of aspects of each supply's design. The various topology types result in various efficiency curves. Some get more and more efficient as you load them more and more. Some get less efficient! And some have a sweet spot - somewhere. For the most part the difference is negligible to your power bill.

Danielrvb; This is great! Glad you're going for it!

About your readings. You need to realize that the voltage drop across a resistor is entirely dependent on the CURRENT through it.

If no current is flowing thru a resistor there will be NO voltage drop at all. So when you measure from one supply lead to the other supply lead you will get the supply voltage. Now if you measure from the other supply lead to a resistor hooked to the second supply lead(side away from supply), you will again measure the entire supply voltage(it will be the same).. That is the correct result. That's what I understand you to be doing and seeing.

Now once you actually close the loop and have current flowing thru the resistor you will then have a voltage drop. Several drops actually.


I want to make sure you are not over-driving your LEDs.

1) Turn off the power.
2) Using the ohms setting. Measure the resistor's resistance. Get it as accurate as you can. Use the lowest resistor setting on your meter.
3) Turn on the string. Measure the VOLTAGE (change your meter settings)across the resistor only - meter lead on each side of resistor.

4) Calculate the current which is flowing through your string.
Take the voltage you just measured and divide by the resistance you just measured. The number you get from that division is the string's current. What is it? I'm interested!

 

[Danielrvb]

ok, I am having a hard time following supply lead to second supply lead.

second part, if I understand correctly now.. your saying this.

Powersupply positve --> Resistor --> leds --> power supply negative


So to measure voltage drop, I would completely hook up LED's. The put on measure like this

Powersupply positive -->**touch meter here** Resistor -->**touch meter here** LEDS --> Powersupply negative.


IE keep everything hooked up, and essentialy jump over resistor and it would show the voltage being dropped?

as for resistance. I did measure that, and I was at 2.2 I think, or about that if I remember now. of course I only measured one, and still fiddled with it after.

I then measure amps going through line. I made a break inline actually from the LEDS to the PS negative end. then connected that gap with multimeter.

I adjusted to where on my RB's I was at .95ish amps, and on the CW's I was .75ish amps.

So I hope this makes sense, and I hope I am understanding you correctly.

 

[Danielrvb]

of course now I am a little worried. Most of my hooking up was very breif just adjusting. I then set up my CW's and let them run for maybe a few minutes. if that, not even a minute probably. (as one string was not coming on, I later realized while moving it around a solder came off) So I kept fiddling to figure that out, and I am not patient, so it would not have been very long.

Anyway, then I turned them back on for a second, and noticed only 4 leds were on for the two strings coming on. For some reason the first LED in each string stopped working? also unplugged then, got meter, and measured line and saw it was like 1.2 amps going through line and quickly shut line off. ( I am assuming the increase was do to the one less LED running?) Anyway, Would this mean I fried the first LED in each string?? Sorta seems odd to me that it would have only taken one out?

 

[kcress]

Ouch!

Your meter description above is correct!

Yes, the added current is because you are shorting the LEDs thru melt-down. You are over driving them or over driving them for the heat removal of your heatsink setup.

This is why you need to measure your resistor.
Turn ON the system, measure the voltage across the resistor.
Turn OFF the system.
Do the math to confirm you aren't over driving the string. If you are over, adjust the resistor so it's longer, (higher resistance), and repeat the measurement cycle until the current is where you should have it. Doing it this way eliminates having to break the string each time to use an ammeter. But - both ways work.

What fuses are you using?

 

[salty joe]

When a resistor is used, does system efficiency necessarily go down?

If one LED burns out will the others be overdriven? Is that what a fuse is for?

When an LED burns out, is it an all or nothing thing? In other words, do they either work or not work? Or would slightly over driving them cause cumulative damage?

I think I read somewhere that LEDs are designed to work most efficiently at the hairy edge of overheating. Any truth to that?

 

[Danielrvb]

Well, I measure the volt drop. I show a drop of 3.47. (measuring over the resistor). For some reason the Ohms is not working, when I try to measure it does not give me a reading. It is a cheap multimeter, and I will take the resistors to my brothers later tonight. He should have a good one.

Also, I noticed it is the last LED in my two strings that have stopped working. I dont really get why they would have been burn out? Heatsink should not have been an issue. I can post pictures of what I am using. However, I continually felt it, and it barely got any noticeable difference. (and even that may have been me making it up it was so small)

And I dont know if you can do it without me being able to measure the ohms. But everything seems interchange able to me. So with only one part of the equation missing, should be able to see others right?

anyway, the power supply, with no load shows 19.2v's. so with a 3.47 drop, it should leave 15.73 going through my string. with 5 LED's per string each LED should be getting 3.146v. This is on the XP-G's which would be in there acceptable range??

So if my thinking is accurate with this. My thought is.. ***?

 

[kcress]

When a resistor is used, does system efficiency necessarily go down?

No.

If one LED burns out will the others be overdriven? Is that what a fuse is for?

Yes. And Yes.

When an LED burns out, is it an all or nothing thing? In other words, do they either work or not work? Or would slightly over driving them cause cumulative damage?

It's all over the map as various things occur during over driving. Like asking, "does someone always get broken arm in a car wreck?", too many variables.

I think I read somewhere that LEDs are designed to work most efficiently at the hairy edge of overheating. Any truth to that?

No.

 

[kcress]

Well, I measure the volt drop. I show a drop of 3.47. (measuring over the resistor). For some reason the Ohms is not working, when I try to measure it does not give me a reading. It is a cheap multimeter, and I will take the resistors to my brothers later tonight. He should have a good one.

If you have tried to measure the ohms with the power on you could well have blown the meter's shunt out. It would be better to bring the meter to your build and not just carry the resistors to somewhere else. This whole adjustment is iterative, that is adjust - check adjust - check

Also, I noticed it is the last LED in my two strings that have stopped working. I dont really get why they would have been burn out? Heatsink should not have been an issue. I can post pictures of what I am using. However, I continually felt it, and it barely got any noticeable difference. (and even that may have been me making it up it was so small)

They all have to have the same current running thru them. If you are pushing things I could see end ones that maybe have less air flow would tend to be the ones that failed first depending on your housing, etc. I suspect you are well over on your current. One problem with HPLEDs is that they are so bright that as a human you can't see much if any difference between high drive and over drive levels. With an incandescent bulb it was easy for us to see what happens as we over drive them. They get noticeably brighter and change towards blue. With LEDs there is no color change with over drive.

And I dont know if you can do it without me being able to measure the ohms. But everything seems interchange able to me. So with only one part of the equation missing, should be able to see others right?

umm.. depends. I can't answer this without more detail.

anyway, the power supply, with no load shows 19.2v's. so with a 3.47 drop, it should leave 15.73 going through my string. with 5 LED's per string each LED should be getting 3.146v. This is on the XP-G's which would be in there acceptable range??

There can be NO assumptions about current using your numbers. That could result in huge over drive. You must have either the current measurement or the voltage across your resistor and the exact resistor value. Yes your voltages look to be adding up well but again NO current value can be derived.

So if my thinking is accurate with this. My thought is.. ***?

Your thought should be.. I need a functioning ohm meter on site so I can do the adjustments needed - correctly.

I see you have also avoided answering my fuse question which leads me to believe you have no fuses. That will likely result in many if not all the LEDs being damaged now, or later, with the added possibility of a fire.:hmm3:

 

[Danielrvb]

Ok, tested them both. The resistance on my brothers multi was actually reading .1 with just his wires connect to each other.

but here are the results I got for my 6 resistors (first three are running XPG's, second three are XPE RB's)

1- 4.0 ohm 3.5 v
2- 4.1 ohm 3.5 v
3- 4.5 ohm 3.4 v

4- 2.4 ohm 2.5 v
5- 2.4 ohm 2.3 v
6- 2.4 ohm 2.5 v


and btw, I just forgot about the fuse quesiton. No, currently I dont have fuses hooked up. I have not been turning them on for long, just testing. And I just got fuses, but dont have anyway to hook them in yet.

 

[kcress]

1- 3.5/4.0 = 875mA

4- 2.5/2.4 = 1042mA

 

[Danielrvb]

See, this is where I am confused. That goes along with the string amperage I measure previously as well. Now, I guess this is a little high on the XP-E's. And actually this does make sense, as when I turn on all strings together, 3 out of the 15 are dimmer. I assume this would be because my power supply cant provide enough amps to run all of them that high.

But the weird part is its actually the XP-G's that I had two go out on. and I actually planned on running them lower to sort of equal out the color to not over power the blue of less lumen capability.

There specs say there max is 1500ma even though they dont recommend it. This shows they should be running at half of their capability.

Am I missing something?? and maybe pictures will help. I will work on getting some posted when I get home, or wake up.


There is only one other thing I can think of, and maybe you can shed some light. I have noticed with a couple of the LED's, they have been grounded (best word I can think of to explain) to the heat sink. When I am testing with just two AA's, I have had a few that when I touch power to heat sink, and other to the LED, it lights up.

Is it maybe more power is going to a single LED or to because they got grounded out to heat sink?

 

[kcress]

I have had a few that when I touch power to heat sink, and other to the LED, it lights up.

Is it maybe more power is going to a single LED or to because they got grounded out to heat sink?

You betcha! Wiring problems could certainly cause your problems.

 

[ebe]

I have a 20volt 4500ma power supply. Appropriate for a DIY build?

 

[AcroSteve]

That would be 4.5 amps.

Given the example here, that would drive 6 strings of 6 LED's - 36 Total.

Is that enough for your needs?

 

[AcroSteve]

Ha!

Double post

 

[TheFishMan65]

20 volts may only drive 5, check the current vs voltage curve to see where you plan to drive them. XPG are probably Ok but XRe at 3.7 volts would be too much