Minimalistic multichip DIY LED build
- Thread starter maglofster
- Start date
Lassef
Member
I really do not know. If the driver use PWM out to the Chip is neither. If the driver do not use PWM out to the LED - its normally the current because if you go below the FV it does not work. With the Maxwell drivers its a question - you can use PWM, 1-10 V and a potentiometer in to the driver - what it use out to the chip I do not know. It could be that it will regulate the current and then it will never go down to zerro without switching the driver of.
Sincerely Lasse
That is what I was wondering. On the MeanWell HLG series as you said will accept all three inputs as control. Makes sense you would no drop the voltage to the leds. Must lower current output. I read that some guys are using setscrews of some type to lock/prevent adjustment of their controller into the driver, so as not to overdrive their leds.
Lassef
Member
Last edited:
tahiriqbal
Sialkot
I think 007bond is talking about the internal pots on some of these Meanwell drivers. These pots will let you adjust your current from 1.3A to 900mA etc, kind of adjust once and forget about it unless you change your mind.
Tahir
Tahir
That could be what I was recalling. I know on MeanWell HLG series ending in A has internal adjustment on output voltage and constant current. When it ending in B the constant current is adjustable through an output cable. Ending in C again is controlled internal. I guess what I was wondering is, on a B series driver, when you adjust the drivers output current down does the output voltage follow it down. I'm trying to figure out what the wattage output you can get from the leds.
Last edited:
Lassef
Member
The current to the LED and the actual forward voltage its dependin on each other. Look here and scroll down to the end of page 4. There yuo have the realitionsship between the current and the voltage for Cree XL-M. Its the same principle for the multichips. Here you also can see why its better to use a driver with a constant current than one with constant voltage. If do not have the constant voltag exactly there you want to have it (variations between chip, load and so on) a small change in the voltage give a huge change of the current. As an example in this case. At 3 V the current is around 1100 mA. At 2,8 V it is 400 mA and at 3,2 V it is 2100. And this is an ideal XL-M LED - it is variation between diffrent indviduals also. This is the general factor why it is better to use a constant current for power LED.
To calculate your FV for 2,22 A we need the manufactor data sheet
Sincerely Lasse
To calculate your FV for 2,22 A we need the manufactor data sheet
Sincerely Lasse
Last edited:
Mike31154
New member
Not entirely convinced one way is better than the other. Perhaps we're getting caught up in semantics here with regard to constant current & constant voltage. Really what most of us want out of a LED is sufficient light of the desired spectrum and as a bonus, automatic dimming. Dimming is one of the best reasons for switching to LEDs since MH & Fluorescent technologies don't really offer anything effective in that regard other than adding or removing lamps/tubes.
Anyhow back to the point I'd like to make, constant current vs constant voltage. We know the maximum current the LEDs we have in our hot little hands can handle from the manufacturers' specs. So perhaps we should focus on that rather than what is 'constant'. After all, to dim the LED we are not feeding it a constant anything, we are varying the current/voltage by reducing & increasing them. The voltage/current relationship curve is irrelevant provided we stay within both those limits. How current or voltage are controlled really shouldn't make a big difference provided we ensure the parameters stay within the LED's specification. There are so many different flavours of drivers out there with constant as well as variable options for both current & voltage it's making my head spin!
Granted that in the event of a LED shorting out in a series string, the constant current driver will protect the remaining ones, but usually a burn out results in an open circuit & the whole string will go out regardless, leaving you with a dark string & having to isolate which LED cooked. And yes indeed, you better make sure your Vf does not exceed the LED's spec, or it's game over in a hurry. I just find all the calculations of voltage drops to find out what constant current driver I need for series string of x number of x watt LEDs is too much of a hassle for me. This is why I chose a stable 12 VDC power supply with manual dimmers to control my parallel 10 watt LED build over a series configuration with multiple smaller LEDs in a string. With the power supply & dimmers keeping the Vf seen by all of my LEDs at or below their maximum, things should hum along nicely with little danger of cooking anything.
My fixture's been up & running for 3 weeks with no problems. I still haven't done any current measurements on the individual LEDs (must do that at some point) & whatever minor variations of current between the 27 individual 10 watt LEDs exist, they have not caused any apparent problems. I've chosen to keep the power supply at 10 VDC (below the max Vf for any of the LEDs) & lower the voltage further using 3 dimmers. I prefer this approach rather than to overdrive any LED trying to get a few more lumens out of it, event with a constant current driver, you're risking a shorter life of the LED. If one or more of the LEDs in my setup burn out or short, there should be minimal impact on the remainder since the Vf will still be lower than max & I'll be able to replace the bad LED fairly easily well before there's a danger to the others.
Lasse, your link the the Cree document was a great source of info. What I found most interesting there were the charts relating to how temperature affects LED performance. Quite drastic once things get a little hot. Kind of makes me glad I chose individual heatsinks with fan for each LED. So far I haven't run any of my 10 watt LEDs at their max Vf, but any time I've touched a heatsink or mounting rail, they've been cool as cucumbers. Now if I could just figure out a way to change my manual dimming to something automatic.......
fppf
New member
LED's are DIODE's, the forward voltage drop is NOT a controlled item in the mfgr process. Its on the spec sheet to give you an idea of what to expect.
LED's MUST be driven and controlled by the current, NOT voltage, PERIOD.
If you control them with voltage tiny changes in voltage will yield huge changes in current. The LED temperature also greatly effects the forward voltage. This is why they MUST be current driven. Ideally the forward voltage of an LED should remain constant with changes in current. But there is a small amount of resistance in the junction which causes the slight change in Vf with changes in If.
You can actually make a crude temperature sensor by supplying a constant current to a diode and measuring the voltage drop across it.
So again, do NOT drive LEDs with a voltage source, you will get burned. No one supplies an LED driver that uses a voltage source, they all go through the effort of making a current source, that should tell you something.
Lassef
Member
In Sweden we have a saying that says: There are many ways to skin a cat.
It is evident that if you have a total control of the voltages and if it is so close to VF as possible and limit the current through a resistor or potentiometer it will work with constant voltage. Common low power diodes are usually operated in this manner.
However, with high power LED, you want to drive as safely as possible and it is current that kills. At the same time, you want so much light as possible from your LED. If a LED is maximized at 900 mA, it is clear that there is a lot easier if you have a driver where you can say that it shall deliver 800 mA and voltages vary according to the conditions. When the voltage exceeds FV, current flows through it and the LED lights up. The heat changes the VF but since the driver shall maintain 800 mA - it changes the voltage to the new VF and the LED light is still on. VF is defined as the voltage required for a LED (diode) shall carry current and give light (LED).
With drivers who gives a constant current life will be easier and safer for us who are not electronics experts when we build our light fixtures.
I basically agree with fppf if you change the word must to should be.
Sincerely Lasse
It is evident that if you have a total control of the voltages and if it is so close to VF as possible and limit the current through a resistor or potentiometer it will work with constant voltage. Common low power diodes are usually operated in this manner.
However, with high power LED, you want to drive as safely as possible and it is current that kills. At the same time, you want so much light as possible from your LED. If a LED is maximized at 900 mA, it is clear that there is a lot easier if you have a driver where you can say that it shall deliver 800 mA and voltages vary according to the conditions. When the voltage exceeds FV, current flows through it and the LED lights up. The heat changes the VF but since the driver shall maintain 800 mA - it changes the voltage to the new VF and the LED light is still on. VF is defined as the voltage required for a LED (diode) shall carry current and give light (LED).
With drivers who gives a constant current life will be easier and safer for us who are not electronics experts when we build our light fixtures.
I basically agree with fppf if you change the word must to should be.
Sincerely Lasse
Mike31154
New member
OK, my bad for soliciting very strong opinions on how to drive a LED. It very much depends on the entire set up & whether you're doing series, parallel or a combination thereof and also the type & power of LED you're using. I'm by no means an electronic expert either, but do have a pretty good understanding of how DC circuitry works. All I know is, I have a 12 volt supply throttled back to 10 volts feeding 27 ten watt LEDs in parallel, all with specified forward voltages between 9 & 12 volts, depending on colour & of course manufacturing variations. I control the voltage (& thereby current) to each of the 3 colours with 3 potentiometers that have some electronics inside the box. I let each LED take care of whatever current it wants to run at, at whatever voltage the dimmer is feeding that group of parallel LEDs. So yes, I'm taking a risk at not measuring the current at each LED & trying to fine tune & balance the whole affair. At 4 dollars or so to replace one of the 10 watt LEDs I'm really not that concerned. It's been working for over 3 weeks with no adverse effects & the LEDs are all staying very cool with the heatsink/fan combos. If folks are more comfortable getting 5 to 10 Meanwell drivers and doing the current control thing, fine with me. One power supply & 3 dimmers is a lot simpler to me.
fppf
New member
I'm an electronics engineer.
Your playing with fire here, pony up and use the correct drivers.
Vf also changes over time as the LEDs age.
If you want to use a voltage source switch back to incandescent light bulbs.
Or, if you really feel so strongly that you are correct and the rest of the world that does this for a living is wrong, just provide your address for billing so anyone that takes your advice and smokes a $100 LED knows where to send the bill.
Your playing with fire here, pony up and use the correct drivers.
Vf also changes over time as the LEDs age.
If you want to use a voltage source switch back to incandescent light bulbs.
Or, if you really feel so strongly that you are correct and the rest of the world that does this for a living is wrong, just provide your address for billing so anyone that takes your advice and smokes a $100 LED knows where to send the bill.
Mike31154
New member
Constant voltage LED driver.
http://www.alibaba.com/product-gs/489526510/5W_constant_voltage_led_driver.html?s=p
fppf
New member
Find me an LED mfgr spec sheet that suggests constant voltage drive.
Just because some yahoo in China labels an voltage supply as an "LED Driver" does not make it correct.
That could also be designed for 12V LEDs that have built in current drivers, like the bulb types used in cars.
Just because some yahoo in China labels an voltage supply as an "LED Driver" does not make it correct.
That could also be designed for 12V LEDs that have built in current drivers, like the bulb types used in cars.
Last edited:
Mike31154
New member
I don't believe I suggested anyone try what I'm doing with a $100 LED and also mentioned in my post that I am taking a bit of a risk with the parallel set up. But I haven't gone into this totally blind either and am not suggesting my way is the right way. It's also not totally wrong. I took some sample current measurements using different resistor values in series with the power supply early in the build process to compare some of the LEDs. Most of them were within 100ma of each other at a given voltage/resistance value. Is it not reasonable to assume that if a manufacturer lists the Vf range of an LED to be between 9 to 12 volts, that if that particular type of LED is provided with a voltage within that range (or below), it will operate for it's full expected life without issue?
I mentioned several times in my posts that I have been underdriving these LEDs by ensuring the voltage is always well below the maximum of 12 volts listed as the spec for the 10 watt LEDs. With the potentiometers set to 8 volts, they are still bright enough to provide the lighting my system needs at the moment and can't see where I'm taking a major risk here. If I can't take the 9-12 Vf manufacturer's spec at face value, why would I believe their current spec? If I'm that far off base, yes I'll most likely find out the hard way down the road.
I mentioned several times in my posts that I have been underdriving these LEDs by ensuring the voltage is always well below the maximum of 12 volts listed as the spec for the 10 watt LEDs. With the potentiometers set to 8 volts, they are still bright enough to provide the lighting my system needs at the moment and can't see where I'm taking a major risk here. If I can't take the 9-12 Vf manufacturer's spec at face value, why would I believe their current spec? If I'm that far off base, yes I'll most likely find out the hard way down the road.
fppf
New member
No, it is totally wrong.
The forward voltage range is the general voltage drop to expect at an operating current so you can design your current driver to have enough voltage capability to be able to drive an LED through the error range of Vf from mfgring variations.
Your not understanding the forward voltage spec at all.
You could get a forward voltage of 10V at its rated current in one lot, then the next lot it could take twice the rated current to get 10V
The forward voltage range is the general voltage drop to expect at an operating current so you can design your current driver to have enough voltage capability to be able to drive an LED through the error range of Vf from mfgring variations.
Your not understanding the forward voltage spec at all.
You could get a forward voltage of 10V at its rated current in one lot, then the next lot it could take twice the rated current to get 10V
tahiriqbal
Sialkot
As said above constant voltage drivers to drive high power LEDs should be avoided (PERIOD).
Tahir
Tahir
epicentyr
New member
I am looking to drive five 20W Epistar multichips off a HLG-100-54D. Is this the correct constant current ballast that will be dimmible via the 0-10v dimming signal from my Apex controller? I will be buying 4 ballasts to run a total of twenty 20w LEDs.
Please help. I would really like to get these ballasts ordered today so I do not use another week at that office trying to figure out this LED thing.
TIA Derek
Please help. I would really like to get these ballasts ordered today so I do not use another week at that office trying to figure out this LED thing.
TIA Derek
Lassef
Member
I'm sorry, I can not answer yes or no to that question. However, I think it is HLG-100-54B you should use. The data sheet is not given any constant current region but it should be one. There are however indicated that if you use 1-10 V for control of dimming -> 1V = 10%. Its seems that you not can dim lower than 10%
My experience with 20-watt LED is that they at a current of 1800 mA yields about 10.5 as FV (Royal Blue 455 nm and 445 nm). 5 * 10.5 = 52.5 which is less than 54 V. In addition, you will use the 1750 mA.
It probably work, but I'm not entirely sure but if you decide to try - use the HLG-100-54B
Sincerely Lasse
Edit: Here is another option
Last edited:
