Hi all,
How well does the measured Vf of a given LED correlate with the current through it?
I ask as I am able to measure the Vf of the LED in my new headlamp easily but am unable to remove the driver and splice leads in to measure the current.
The LEDs in question are XMLs and XRE from the very useful guide on this site and I am wondering if accurate measurements of the Vf enables you to provide you with an reasonably accurate idea of the If?
What's the power source? You may be able to get close by measuring the input current instead.
LED forward voltage drops as the LED gets hotter (and the LED simultaneously becomes less efficient in terms of lumens/watt). Vf also varies from sample to sample even at the same temperature. Process improvements lead to lower forward voltage in newer production runs of the same LED. Simply measuring Vf is not be a great way to determine current.
That said, here’s match’s XM-L testing: Results: Testing XM-L, MC-E, SST-50, and XP-G emitters **Updated**
match’s XM-L direct-copper testing: Results: Direct bonded XM-L emitter to Copper
and match’s XM-L2 testing, including copper mounted: Emitter Test Results pt4: XML2, XP-E2, XP-G2, and Nichia219! (along with sinkpad tests)
djozz also did some testing which corroborates match’s results: Crash-testing a XM-L2 and a XP-G2 on copper Sinkpads
It’s important to note that all of the above testing is done with big chunky heatsinks and that some of it is done on copper. Both things make a big difference within seconds of turning the LED on.
What’s your goal once you figure out LED current?
Various, from AAA NiMH to a 8800mAh Li-ion battery pack.
The light is one of those 1000 lumen bicycle lights and uses plugs and sockets to connect the light to the large and remote battery pack. Measuring current draw at the batteries would involve either cutting and the re-splicing the cable or making up a male to female adaptor with some facility to go inbetween the existing connection.
I don’t really want to be cutting and shutting as I don’t want to adversely affect the reliabilty of the unit once I’ve finished tinkering.
It also ouccured to me that dirver losses might or might not be a greater variable than fluctuations in the correlation between Vf and If for a given series of LEDs but as I don’t know much about drivers or LEDs I had to ask… 
Thank you for the links, those are very ahem illuminating threads!
To answer your last question first I must confess to being a little disappointed in my new acquisition: http://www.maplin.co.uk/p/refrakta-1000-lumen-rechargeable-bike-light-a15ln
I had read about the fabledd XML LED and it’s prodigous output and the light has left me a little underwhelmed if I’m completely frank. There seems to be little difference between the high and medium modes which seems to indicate that there is some current bottleneck in the circuit somewhere and preliminary Vf measurements seem to confirm this.
I am measuring 3.12V on high, 3.00V on Medium and 2.8V on low. From the XML datasheet this seems to equate to a If of about 1.5A, 1A and 500mA respectively. Again extrapolating from the data sheet it seems that the light produces somewhere around 500 Lumens on high, 400 on medium and 200 on Low which again seems to agree with the general pattern.
I have noticed that most other lights of this type run a 2S2P configuration for about 8.4V into the dirver. My light has a 4P battery pack for 3.7V nominal and I am wondering whether it is current losses in the ~ 0.75M of cabling and plug/socket junction or whether the driver itself is not up to the task.
To reiterate, I’m not looking to derive an exact figure for the currrent, just wondering whether the Vf of the LED can give an idea of the state of the drive current in emprical terms.
Sounds like a driver with poorly thought out mode spacing. It's nearly impossible to see much difference between 100% and 60%, it needs to be more like 100-33-2 to look 'right' when cycling thru the modes.
That’s interesting, I do know that the eye’s response to light is logarithmic so that makes perfect sense.
It seems that if the drive current of high was a full 3A, the medium figure of 1A would follow the 100-33 ratio.
Does it sound like the driver is not providing enough current on high?
Are lights like these regularly underdriven for thermal reasons or something?
Stock lights driven at 3A are rather rare, that's pushing the limit for a LED mounted on an aluminum star. Most likely it's just crappy mode spacing.
Also if it's using a 4.2v pack the voltage will have a huge effect on the light output, it'll only have limited runtime before output starts falling because of the Vin-vs-Vf thing. To do it properly it needs a battery pack that even fully depleted will be higher than the LED's Vf at whatever drive current. Otherwise it'll be disappointing.
Vin at full load with a freshly charged battery pack ( 4.22V no load) is about 3.7V. At that voltage the Vf is about 3.12V.
From what I’ve read a Li-ion cell is discharged at around 3.4V which is higher than the Vf of an XML driven at 3A ( around 3.3V from the literature and the tests above).
How much of a voltage drop in the driver would be reasonable?
Can you give a link or more info on the exact model light you're working with? The 4.2v pack instead of the more typical 8.4v is a bit of a red flag that they've probably cut corners and cheaped out everywhere possible, their primary concern was to get somebody to buy it.
This one mate: http://www.maplin.co.uk/p/refrakta-1000-lumen-rechargeable-bike-light-a15ln
Looks like a re-badged clone of something.
I will try and get photos of the driver if that helps?
Oh. Well.
You do realize that their definition of words doesn't mean the same as ours, and they're counting on that as a selling tool, right? In these things "1000 lumens" means "the LED used in this is theoretically capable of a maximum of 1000 lumens when driven at the max current and not counting any losses from a reflector or glass lens"... the output in actual use with the actual driver and actual power source is mentioned nowhere. How convenient for them, eh?
I'd say your estimation of 1.5-1.6A on high is probably close, it seems to fit with everything else known about the light. It's not broken, it's just not very good. :(
I do, the thrust of the question was to establish by what degree their claims are exaggerated so as to see whether or not it was worth trying to modify it.
From Match’s power run above it seems that the difference between 2.5 - 3A in terms of output is hardly worth bothering with, if the difference was approximately that much I would leave it be and just use the thing.
As things stand although it puts out a lot of light in spill and out to about 30-40 yards or so it is lacking in throw for my purposes ( hunting ) and I am wondering if driving it properly would add the extra 20-30 yards I need from it?
It appears that because it uses effectively a single Li-ion cell I should be able to swap in a 7135 driver ( from my researches, happy to be corrected here 
)and possibly even run it off a single 18650 cell on the headband.
At this point I am not sure whether it is the driver, the battery or resistive losses holding the light back. I can test the last two with different power sources and things but am not sure how to go about assesing the driver, espceially as it seems like ti would be difficult to get out without breaking. 
If you want to do it just for fun, go for it, but if you want to fix this one to same money buying one that does what you want, it'll be cheaper just to buy a good one. There are drivers for this style light available but their listed specs cannot be trusted, they don't know what they are selling. You'll have to get them and figure it out for yourself. (some are listed as buck drivers, capable of running a single LED but are actually boost drivers, and will burn up if you try to use them as the specs say, etc.)
http://dx.com/p/21-65mm-8-4v-1500ma-circuit-board-for-bike-light-sku-94182-82510-82734-more-166085
http://dx.com/p/smart-led-driver-board-for-bicycle-flashlight-blue-236160
There is a certain element of doing it just for fun, I hope I came to the right place!
Thank you for the links to those drivers, they all seem to be for more than one cell though.
Assuming I can engineer some sort of battery box for a single 18650 cell, I would prefer to use that as two on the back of the head seems like it might be too heavy, the cable run would be shortened to 25% of original and my voltage drop issues disappear?
Thanks again for all the help on this.
If it's using a 1.5A driver and it's putting out 1.5A, lower resistance in the cabling/connectors isn't going to make much difference. But that's the catch, we don't know what the driver's output is, or what it should be.
Quite, I don’t know what the driver is rated for or indeed capable of.
I would imagine, or rather hope, that the manufacturer would have put something at least capable of 3A in there and so am thinking it might be worth investigating resistive loss before replacing it. At 3-4v I think it takes very little resistance to limit the current in the circuit. From my calcs a total resistive path of 2.4 Ohms would limit the current to about 1.5A if the LED was plugged directly into the battery.
Could 1M of what is probably something like 26 SWG wire, a plug and socket of this design: http://luckylarry.co.uk/wp-content/uploads/2009/06/arduino-9v-battery-plug.jpg add up to around that much?
I am sorry for the silly questions, this is all very new to me.
3A driver in a light of questionable quality is very very unlikely.
Is the long cable hard-wired to the light? If it is, you'll have to disassemble it and solder shorter/fatter wires direct to where the original cable attaches on the inside, then see if it makes a difference in the output.
There is a 12” length hard soldered into the light and about about 2 1/2’ of to the battery with the plug and socket illustrated above in between.
I don’t have any other lithium cells at the moment apart from this pack and I would need to take it apart to be able to use shorter leads.
Note with displeasure what you say about the driver…
