New LED DIY Build

Just to get this straight; do you have the XM-L strings AND the XR-E RB strings in one parallel setting and connected to one driver?

If that's the case, you'll have a problem. (or not when you buy some stocks from the fuse and LED company ;)
XM-Ls and XR-E need different amps and volts, XM-L needs roughly 3V @ 1500mA (2.9V@ 700mA) and the XR-E will need 3.5V @ 700mA.
So you will need at least 2 drivers. You can't put different strings in one parallel setting.
 
He said he ballanced the voltage you can do something like

7 XML at 3 volts and 6 XP-E at 3.5 volts that the simple math. What ever voltage the driver puts will be split. For instance if it was too white you could do 8 XM-Ls at 21 / 8 volts. You just have to watch that you don't over/under drive a string by too much.
 
That's exactly what I'm doing (mixing and balancing). From what I understand as long as I don't exceed the maximum current for the LED I will be ok (which means that I'll be underdriving the XM-L's by quite a bit but at 833ma they still put out a respectable 350lm, while the XR-E's should put out 215lm.

I just need to figure out how balanced the strings really are instead of in theory, so I'll need to pick up some new fuses and then get my multi-meter in line to see what's going on for current. I'll try to take some pictures tonight too to show how its setup currently.

I just find it really strange that with all 6 parallel strings that I was able to get to the 1a current needed to blow the fuses on the 3 RB strings. I'm hoping that measuring the RB string with the driver turned all the way down (voltage and current) that I can get a better understanding. My theory is that maybe the RB strings are actually a little higher voltage which is drawing more of the current, I should be able to test for this.

If anyone else has had this type of issue please let me know what you did.

Thanks!
 
So, if I understand correctly; it's no problem to have LEDs in a string, or multiple strings, which need different voltage - like 2.9V and 3.5V - as long they're balanced to run at the same current?
Or am I way off?
 
Ok, here we go:

With the power on, internal pots turned all the way down, here are the measurements.

I get 38 volts on the powersupply which is a lot lower than the stated minimum 44 on the 240 datasheet.

I now have 6 5w 10ohm resistors in-line with each string. Measuring across the resistor I get the following (on the 2000m setting):

String with 12 XR-E and 1 XM-L: 1016
String with 13 XM-L and 2 XR-E: 16

I'm not sure how to calculate the difference in current based on these readings, is this saying that there is 1a on one of the strings and 16ma on the other?
 
Measured the LED voltage across the individual LEDs in two of the strings, one XM-L string and one XR-E string:

XM-L (37.7): 2.52, 2.52, 2.53, 2.53, 2.52, 2.52, 2.51 (XR-E), 2.52, 2.50 (XR-E), 2.51, 2.52, 2.49, 2.51, 2.50, 2.50
XR-E (36.58): 2.78, 2.80, 2.80, 2.78, 2.77, 3.03, 2.66(XM-L), 2.81, 2.79, 2.82, 2.95, 2.77, 2.82

Pretty close as expected, so why the huge difference in current? and why would it be focused on the XR-E dominated strings?

I tried turning the voltage pot up on the powersupply and regardless of how much it was turned up, it didn't change the voltage being drawn either on each string or on the powersupply output. I also tried adjusting the current a minor amount, and then remeasured, and the strings had increased current readings (and the visible light), but I turned it down quickly as I'm not sure if I'm over the power rating for the LEDs (although the fuses are still intact). Am I measuring the current of the strings wrong?
 
Ok, this gets stranger the more I mess with it.

After measuring the voltages across the LEDs, and looking at the data sheet, I sort of reverse the chart to read about what the current is based on the voltage, it looks like they are running way under. So a slowly increase the current using the POT. As I increase the current, I keep measuring across a few of the LEDs just to see how much the voltage is increasing, and strangely the voltage difference between similar LEDs in different strings gets smaller. I increase the current more and same thing, they keep getting closer.

I now have the current turned up to the maximum that the powersupply will put out, and all of the LEDs now look (hard to tell because they are so bright) to be running about the same. When I measure them now the XM-L's are now running at the same forward voltage (around 2.81-2.83) which means that they are running at close to the 800ma that I'm shooting for.

I can't imagine how bright these XM-Ls would be at 3a.

Now I'm going to slowly adjust the voltage to get them as close to my target current (800ma) as I can. Looks like its now working as expected though. Maybe my first attempt was due to a short or a faulty fuse.
 
At low current a small change has a greater affect on voltage. At higher current the reverse is true. Look at the graphs.

Pretty close as expected, so why the huge difference in current? and why would it be focused on the XR-E dominated strings?
Click to expand...
They are different types the XR-E are running closer to the max so a small change in voltage makes a large change in current as compared to the XM-Ls.

I tried turning the voltage pot up on the powersupply and regardless of how much it was turned up, it didn't change the voltage being drawn either on each string or on the powersupply output. I also tried adjusting the current a minor amount, and then remeasured, and the strings had increased current readings (and the visible light), but I turned it down quickly as I'm not sure if I'm over the power rating for the LEDs (although the fuses are still intact). Am I measuring the current of the strings wrong?
Click to expand...
As you figured out the LED were not voltage limited at this point but current limited.

strangely the voltage difference between similar LEDs in different strings gets smaller
Click to expand...
You are getting to a point were large current change make very small voltage changes. Notice the graph gets more vertical.
 
Have a electrical question for anyone who can answer. I planned on 80 fedy leds with a large china PS 200-36 @ 5.2A constant voltage. 8 strings of 10 leds for a total of 80, was the original plan. Each string fused and a 1A 5W resistor. All leds are 3.6Fv and 700 mA. One LED bad so 7 string of 10 and 1 string of 9..will this work still or am I flirting with disaster?
 
Last edited:
Bugcrusher: This is the same basic idea as my parallel setup, I have 1a fuses in-line to prevent a cascading/catastrophic failure, the idea is that if you have one LED burn out that the additional voltage/current will never exceed the 1a limit. You will however lose all of your fuses, as the over-current will then load up on the remaining strings. You also probably won't get close to the 5.2a that the powersupply is rated for and I don't see a datasheet for that powersupply, but if it doesn't have any adjustment you will be stuck with the voltage/current that it produces, which is why I chose the HLG as they have limits for both the voltage and current.

FishMan: From what I can tell with my setup, the setting that really makes a difference is the voltage, if you can control / limit how much voltage is being used then you control the current going through the LEDs. When I set the voltage to 38v (lowest limit on the HLG) even with the current at max it appeared that the LEDs could not get enough voltage to use all of the available current, and as I carefully raised the powersupply voltage the amount of current went up based on the LEDs ability to use it (curves from the datasheets). So my theory (although I don't really want to test it) is that if I lose an LED (fails open) on one string there will be an imbalance in the total voltage in that one string, and the remaining LEDs will split the available voltage across them causing an increase in the current passing through that one string, which -might- result in another LED failing and cascading from there. But in my case with the 1a fuses, if the current of that string exceeds the 1a the fuse will blow and likely cause successive fuse failures due to being over the 1a current across the remaining strings.

The other question I have is why do we want to use expensive power resistors just to take a current reading from the circuit, wouldn't it be easier and more efficient to use a higher rated resistor 1k ohm, that would only dissipate 48ma instead of the 100ohm versions that dissipate 480ma?
 
Last edited:
It was part of a group buy, so stuck with it now. Trial and error has gave my club the parameters for it. As far as the leds Fv = 3.6 to 4.0, 700mA from the website. These are old school leds from what I have been told. Had I known it was a constant voltage source I would not have gotten it. I agree the HLG is the better way to go. Any of them 36 volts or so with 5.2A?
 
Bugcrusher said:
Have a electrical question for anyone who can answer. I planned on 80 fedy leds with a large china PS 200-36 @ 5.2A constant voltage. 8 strings of 10 leds for a total of 80, was the original plan. Each string fused and a 1A 5W resistor. All leds are 3.6Fv and 700 mA. One LED bad so 7 string of 10 and 1 string of 9..will this work still or am I flirting with disaster?
Click to expand...

With a constant voltage driver, you're gonna need lots of fans to keep the LEDs cool .... extremely cool ;)
or you might have to deal with a so-called "thermal runaway"
As LEDs get warmer, they tend to use more amps, using more amps means more heat,
more heat means inturn more amps etc. etc. and a Thermal Runaway has born.

When LEDs get warmer and want to use more amps, a Constant Current driver will automatically puts out less voltage to keep the amps at a specific level,
thats the whole secret;)
.
 
Last edited:
You cannot run with just a constant voltage supply and a fixed resistor. Disaster awaits. If you want to run in this mode you need to look at my thread "on a budget" and follow the instructions in the first post. You CANNOT use a fixed resistor - it has to be adjustable.
 
The other question I have is why do we want to use expensive power resistors just to take a current reading from the circuit, wouldn't it be easier and more efficient to use a higher rated resistor 1k ohm, that would only dissipate 48ma instead of the 100ohm versions that dissipate 480ma?
Click to expand...
Resistor don't dissipate ma they dissipate power which is watts. Watts is current * voltage. Voltage is current * resistance. So power is current square times resistance. So if you had 50 ma going through a resistor the power would be

R = 100 P = 100 * 0.05 *0.05 = 0.25 watts
R = 1000 p 1000 * 0.05 * 0.05 = 2.5 watts
So for reading the current you want a very low ohm resistor 1 or 1/10 ohm. Also not sure why you think a 1k is better than a 100ohm.
 
I've now got the driver running at 47 volts with the 100ohm 5W resistors, and the XR-E's are running at about 3.1v and the XM-L's are running at about 2.9 accross all of the strings. The only problem now is that the resistors are getting so hot they are melting the terminal blocks they are connected to after a couple of minutes. I'm thinking that the voltage and current through these is way to much, might also explain why the fuses aren't blowing as the resistor is eating up lots of the voltage/current.

Here are some pics.

The project box with the connectors and the power supply:
album.php


A crappy picture of them all on: http://www.reefcentral.com/forums/album.php?albumid=4527&pictureid=31219

A picture of the array in test mode:
http://www.reefcentral.com/forums/album.php?albumid=4527&pictureid=31225

I also changed out the 100ohm 5w resistors with some 10ohm 1/4w resistors, and at about 40v the resistors on the XR-E strings start to smoke - my guess is that the imbalance of the strings is causing more current to flow through the XR-E strings.

If someone could suggest a process for testing the imbalance between the XR-E and the XM-L strings I will follow it and post the results. Thanks!
 
Last edited:
No you are grossly over power the resistors and frying them. A 5 watt resistor can burn 5 watts of energy.
Watts = Current * voltage
Voltage = Current * Resistance
So
Watts = Current ^ 2 * Resistance
So at .25 amps (which I expect you are way over)
.25 * .25 * 100 = 6.25 watts of energy being used by the resistor.
If you are running at 1 amp
1 * 1 * 100 = 100 watts so you really need a 3-5 (IIRC kcress comment) over to keep it at a reasonable temperature. So starting looking for some 300-500 watt 100 ohm resistors.

This is why constant voltage source don't work real well. At least this is what I think you have.
 
You must log in or register to reply here.

Similar threads

T
Alternative Lighting for Biocube 29
Replies
0
Views
108
thommyboy12183
T
Anemone
Kessil AP700 FS
Replies
0
Views
592
Anemone
Anemone
R
DIY AI-coded ESP32 based aquarium controller
Replies
0
Views
1K
ronnyking
R
M
Can't find 60"VHO Actinics anymore... LED recommendations
2
Replies
28
Views
2K
ClownBubbaJoe
ClownBubbaJoe
C
Facing LED Woes: Controller Troubles and Compatibility Questions 🤔💡
Replies
3
Views
824
combs.eric.813
C
Back
Top