Anything other than an XM-L for a flood light?

[Mattaus]

Hi all,


I'm new to this whole game so please be patient.


I'll cut to the chase and just ask those in the know: are there any alternatives to the XM-L LEDs for use in flood lights? I realise that XM-L LEDs appear to be the 'best' at the moment, but there are so many LEDs out there floating around at half the price that surely one of them will be a decent 2nd choice?


I'm building a 4 emitter flood light. The brighter the better obviously. I'm using the following driver (http://www.dealextreme.com/p/18v-5w-cree-circuit-board-for-flashlights-16-8mm-5-5mm-26110) to run 2 LEDs in series, with two of these setups in parallel forming one flood light. The setup will then be connected to a deep cycle battery.


If a different driver is required as a result of using a different LED that is fine, but a total current draw up 3A or less is desirable.


If nothing is even close to being suitable I'll go with the XM-Ls.


Any guidance or insight would be greatly appreciated.


Cheers,


- Matt.

 

[calipsoii]

Cree's not the only game in town. Check out the Luminus SST-90. With proper heatsinking they're rated to 2200lm. The die on these LED's is larger than the XM-L (giving it much more flood) and since they've been out for a few years now, they're available in pretty much any tint you could want. Nailbender made me a dropin with an SST-90 3000k color temperature that does about 550 lumens out-the-front driven at 2.8 amps.

 

[Mattaus]

The SST-90 seem expensive ($30+) and at 2.8A I can run several smaller LED's and get over 1000Lumen. If my understanding of how it all works is correct?

How about 4 of these: http://www.dealextreme.com/p/cree-xp-g-r51b-7000k-350lm-emitter-with-16mm-base-44306. Driven at 1.2A they could put out around 380Lumen each.

Or alternatively 6 of these (http://www.dealextreme.com/p/cree-xr-e-q5-emitter-on-premium-star-228lm-at-1a-2394) driven at 1A would give me around 1400 Lumen total?

Either of these 'solutions' appear to still give approximately the same output strength as the 4 XM-Ls but at close to half the cost. What am I missing here?!?!

 

[hellokitty[hk]]

Those XR-E's are Q5 binned, they're about 28% less efficient than R5s and T6s. Also, they are most certainly not flood lights.
Your not driving the XM-L's hard enough; if they were driven at 3A each, there is no way an array of less than 6+ XP-G's are going to be able to keep up.
If you modify DX SKU 20330, you can get 3A output and four XM-L's in series.
Though since your saying 3A or less, then there is no point in using four XM-L's.

 

[Mattaus]

Ok but what about the XP-G's I have listed above? Same story I guess?

Maybe one option would be to run that driver you have listed with 2 XML-T6 in series with the driver providing 2.4A. This should give about 1600 Lumens output apparently...though I should point out this light needs to run on a 12V deep cycle battery. For that to work I'd have to hook a diode (or 2) up in series with the battery to drop the voltage down to the required 8.4V. My understanding of this concept is loose so I may have to ask around to confirm :-/

 

[sn0wBLiND]

XR-E and XP-G would be easier to drive than XM-Ls but I don't think you will save any money by using them.
LCK-led has a sale on T6 flux 1A tint XM-Ls, $6 each.
Also have you considered using bridgelux leds?

 

[Mattaus]

Also have you considered using bridgelux leds?

Totally. What are bridgelux leds lol? Seriously is it a brand or a type?

I signed up with LCK-led and just bought several T6's. Damn sales...so I guess that's that sorted haha! Just need to decide how I best drive them. I'll probably go with 2 in series with the driver hellokitty suggested (with a diode on the battery line to drop the voltage below 8.4V), and if that is not bright enough I can hook 2 sets in series and they'll share the 12V input so won't go over the 8.4V input limit.

If that is even the right way to think about it.

 

[evilc66]

www.bridgelux.com

Great LEDs, with good color rendering, and high quality, although not as efficient as the latest and greatest Crees. But, you can get a single emitter (multi-die) 8800lm cool white LED from them 🙂

 

[Mattaus]

Having just bought the T6's I'll have to keep bridgelux LEDs in mind for next time :-(

 

[videoman]

The DX 26110 driver puts out 1.2A , 2 leds in series will still have 1.2 A going through them, really underdriving the Xm-l's by a wide margin. Like driving a ferrari under 65. You really want lots of output from 4 leds with a floody beam ? Try using 4 XM-L's 2S2P driven by taskled's 6A driver H6CC I beleive, put carclo 70 degree reflectors ( 10170 I think) on them, Made for the MC-E but works fine with the XM-L, and whoa, you're looking at somewhere like 6K luxes to brighten things up a lot. I would drive them at around 75% full throttle for better heat management/efficiency and drawing around 3-4 A from the batt. Nice thing about that driver is that it is dimmible ready.

 

[hellokitty[hk]]

I do not think diodes reduce voltage.
The vin of the driver I mentioned can be increased by replacing the capacitor.

 

[Mattaus]

I shied away from the H6CC driver in my flood light thread already because I couldnt bring myself to spend $32 on one driver. I still feel much the same way despite getting the XM-Ls for $6 each. Surely Task-LED are not the only mod who make high power drivers?

 

[Mattaus]

I do not think diodes reduce voltage.
The vin of the driver I mentioned can be increased by replacing the capacitor.

Yeah I saw that in one of the reviews but I can't quite follow the maths he's used to determine it. It's probably not complicated but I can't understand it the way he's written it :-(

Be good to clarify though because if that is easy enough to do (my sodlering skills are better than average) then that's the ticket for sure.

EDIT: Sorry...the guy's maths just clicked. Replace the capacitor with a 10mF one and the unit will work with any Vin up to 16V. That cap is nice and big as well so it'll be a piece of cake top replace.

 

[Mattaus]

This is possibly a stupid question, but what's the limit on the number of LEDs I can run off one of these drivers? Is it dependent on the voltage of the system? In my mind 🙂duh2🙂 a 12V system would mean I should be able to run 4 XM-L LEDs (3V each) off the one driver while a 9V system could only run 3 and so on. Or is this a fundamentally flawed idea?

:thinking:

The reason I ask is because it would seem doing this would allow me to run 4 XM-Ls at 2.8A at 12V. However, if I can only run 2 XM-Ls off 1 driver, then 4 LEDs would require 2 drivers and thus pull 4.8A at 12V. Obviously less current draw is better...

I'm thinking about this too much today. Might be a good time for a break!

 

[hellokitty[hk]]

The vf is not <3v except at very low levels, so no it won't work.

The reason I ask is because it would seem doing this would allow me to run 4 XM-Ls at 2.8A at 12V. However, if I can only run 2 XM-Ls off 1 driver, then 4 LEDs would require 2 drivers and thus pull 4.8A at 12V. Obviously less current draw is better...

If you're saying that 4 XM-Ls2s2p @4.8a per series circuit (2.8a each) at 12v is equivalent to 4 XM-Ls in series @2.8a and 12v, then you're doing it wrong. It sounds like you're saying that you can get the same output for half the power just by using one driver instead of two, which is not true.
Now firstly, 2.8*2 is 5.6.

If you use two drivers in parallel with your power source, powering two XM-L's in series each, then you'll be drawing 2.8a per driver and 2*vf, so each XM-L will receive 2.8a and will consume ≈ 6.2v - 7v per pair.
If you use one driver and 4 XM-Ls in parallel, then each LED will also be getting 2.8a and the proper vf, but the driver will be outputting ≈12.4v - 14v.
Both scenarios multiply to the same wattage, so now we're not randomly pulling half as much energy for the same output.

That driver is a buck, so you vin will need to be at least a couple volts higher than the output voltage, and it will most likely run more efficiently with a higher vin.

Remember that in both of these calculations we are not factoring in the temperature coefficient of the XM-L, which is -3mv/°C, or resistance in wires ect...careful with battery vdroop too if you're using a battery.

 

[Mattaus]

Now firstly, 2.8*2 is 5.6.

Haha! Fail. 😳

In my defense I have been looking at a few drivers today all with different voltages/currents etc so probably got them mixed up. I can multiply that is for sure...at least in my line of work I'd hope so.

Most of everything else you've written makes total sense. 3 XM-Ls in parallel with 1 driver could be a good compromise - 2.8A at 9.3-10.5 volts gives a 26-30W system and sits nicely between the two solutions. Leaves some room in the battery/other losses etc - though is it enough or too close?

Voltage drop shouldn't be too much of a concern as the deep cycle battery I need to use indicates when it is at 50% capacity and needs a recharge. Not totally sure how the temp coeff changes things but I'd imagine that it's to do with an increase in voltage as the LED gets hotter. I'll google that.

 

[videoman]

A little bit of basic electronics here would help you understand 3 most important formulas. The good ol' ohm's law that V=IR and P = VI . Now you can interpolate them as you will like I=V/R or I=P/V. Also to remember is that when resistances (or leds for that matter) are paralelled, each led gets 50% of the total current applied. ( considering that the leds are the same spec'd resistance values). When leds are series connected, each gets the SAME current through them (again considering all the leds used are theoretically identical). Now, when one is using 4 identical leds of 2 series strings connected in parallel, say you have 3 amps as the feed, that 3 amps gets split in two for 1.5 amps for each series string so that each led gets fed 1.5 amps. If you are using a battery of 12 Volts DC, (car battery, deep cycle, motorcycle battery, or any good 12 Volt dc source that can give out a healthy 4 amps of current without droop ( voltage drop when applied to load), then you can easily string 3 leds in series ( if the leds are around 3 volt rated, usually 3.2-3.4) or 2 leds in series so that the driver ( if it is a buck design) will
be able to buck that 12 volt to the voltage level that the leds require WITHOUT going over their max rated voltage. It does so by having a constant current going through each led as specified by the driver design. If your driver is rated at 6 amps out, then you cannot just put 3 leds in series that only accept a maximum of 1.5 amps each. That will let each led take in 6 amps and in no time will burn them out.If you have a 3 amp driver, even that will burn out the series string UNLESS 1- you put 2 SERIES STRINGS IN PARALLEL to HALVE that 3 amps to 1.5 so that each led will receive the maximum without burning up. If you have a 6 amp driver, now you can put 4 series strings and have each string with 1.5 amps OR take the XM-L led that can handle 3 amps and connect 2 strings with 3 amps each. Reducing the amount of leds has advantages. Fewer leds mean fewer lenses/reflectors, less soldering, less adhesive potting, less overall cost. I heard it before that 2 XP-G's are equal to 1 XL-M , or very close. So let's take an example of using 4 XM-l's ( 2S2P configuration) with a 3 amp driver, you have 1.5 a through each led (underdriving them) Nice and cool running at that, with around 450 lumens each ( I could be a little off on this one) depending on the bin. You could also get that kind of figure using 4 XP-G's instead, but you will be maxing them out. Not a problem . Let's say you have a large enclosure with excellent heatsinking and ou really want to up the lumens.Then you go for the big guns (XM-L's and feed them a steady diet of 3A each with a 6 A driver.) Put 2 series strings of 3 each paralelled. Now look at that figure of around 750 lumens X 6 = 4500. Add reflectors on those and you get what I'm talkin' about.

 

[videoman]

Haha! Fail. 😳

In my defense I have been looking at a few drivers today all with different voltages/currents etc so probably got them mixed up. I can multiply that is for sure...at least in my line of work I'd hope so.

Most of everything else you've written makes total sense. 3 XM-Ls in parallel with 1 driver could be a good compromise - 2.8A at 9.3-10.5 volts gives a 26-30W system and sits nicely between the two solutions. Leaves some room in the battery/other losses etc - though is it enough or too close?

Voltage drop shouldn't be too much of a concern as the deep cycle battery I need to use indicates when it is at 50% capacity and needs a recharge. Not totally sure how the temp coeff changes things but I'd imagine that it's to do with an increase in voltage as the LED gets hotter. I'll google that.

Driving 3 XM-L's in parallel at 2.8A to them really means you are driving each at 933 milliamp if your driver Iout is 2.8A. Not really the best solution as you can just drive 1 let at full pop (2.8 A)for around the same lumens. To drive 3 XM-L's in parallel at 2.8A each is only possible if your driver is 2.8 X 3 = 8.4A. Perhaps you meant 3 leds IN SERIES, now that would make sense.

 

[Mattaus]

Perhaps you meant 3 leds IN SERIES, now that would make sense.

That's exactly what I meant...I just described it very poorly. I'm losing my way it seems lol. Assuming I said 'series', arethe 3 XM-Ls in series with 1 driver an OK bet? Lets imagine I have not done anything dumb (again) this would provide each LED with 2.8A and require 9.3-10.5 volts from the driver.

Maybe I should just stick with 2 LEDs to keep it simple...

For the record I'm still digesting your last post. I have seriously copied it to a word file and am in the process of putting in my own terms (ie// dumbing it down), though this pesky thing called 'work' is slowing my progress.

 

[Starlight]

You really need to look at this thread. You can run 3 XM-L's with 12v input and a cheap driver. http://www.candlepowerforums.com/vb/showthread.php?201392-Poorman-Mutli-Lux-setup-method