Need help understanding circuit board parameters

[Entacmaea]

Hi there, just wondering how to read and shop for circuit boards for flashlights- I'm making a dive light. I understand that the circuit board's output current is the mA given (3000mA, 2500 mA, etc) As an example, let's use this circuit board:

http://www.lightmalls.com/super-output-ssc-p7-led-driver-board-low-high-middle-5-5v-15v-2

1) Is the voltage they list on the board (15v) the output voltage, so this would be able to drive about 4 XML 5W LED's whose forward voltage is 3.35v at 3000mA?

2) Some other circuit boards list a range - like this one http://dx.com/p/4-mode-2500mah-led-driver-circuit-board-for-flashlight-dc-5-0-8-4v-106303?item=18
in this case I would be able to drive 2 XML's whose forward voltage is 6.7v?

3) Do circuit boards need to be cooled? In my application it will be easier to remote mount them with the batteries instead of new the LED in the heat sink.

4) Lastly, regarding batteries, in general I would drive example 1 above with four 18650 3.7 volt batteries, and example two with 2 18650 volt batteries? If I wanted longer run time for example 2, the 2 XML's, how do I run it off 4 18650 batteries whose voltage is higher than the forward voltage of the LED's? (2 parallel packs of 2 18650 batteries?)



Thanks for the help- new to this...

 

[Entacmaea]

Anyone?

 

[uk_caver]

(Assuming the description is accurate)

1/2) No. It is described as a 'buck' regulator, which is the term for a voltage-reducing regulator.
The other references in the text (lack of mention of multiple LEDs, calculated runtimes) only suggest it driving a single LED.
That doesn't mean it's guaranteed not to be able to drive 2 LEDs in series from a suitable power source (voltage higher than the combined Vf of the LEDs), but expecting it to do that at all, let alone reliably, might be expecting too much.

3) It could be worth at least mounting a driver on a small metal plate with suitable epoxy (ideally after working out which are the heat-generating bits and getting them close to the plate, but electrically isolated from it).

4) Since it's a buck driver, you'd need more cell voltage than the LED drew - running from 2x18650 for 1 LED would be fine. 4x18650 would seem too high for the recommended maximum input voltage.
But that might be irrelevant if you really are looking for a driver designed for multiple XMLs in series.

 

[Entacmaea]

Thanks so much for the reply- UK caver! So, here is another driver whose voltage is in the range of two XML's (5-8.4volts) and current is max 2500mA, which would be great for the 3000mA XML's- would this work to drive two of them?

http://dx.com/p/4-mode-2500mah-led-driver-circuit-board-for-flashlight-dc-5-0-8-4v-106303

Combined fV of the two XML's is 6.7 at max, so this would have around 1 volt of head room once the 2x18650 batteries drop from 8.4V fresh off the charger to 7.4-ish.

I probably plan to create a pack of at least 6 (may 8) 18650's, 2 in series for voltage and 4 in parallel for capacity.

Thanks again for your help- been posting here for a couple weeks, and you are the first person to respond :)

 

[uk_caver]

To be honest, I'd be wary of any driver which didn't explicitly state it was capable of driving multiple series LEDs.

Given the typical nature of the circuits, if a buck driver was intended to drive one LED, asking it to drive more than one might well work in the short term, but since the circuit would be driving a load taking twice the power, there would be something like twice the normal heat generated in the driver, and twice the average current running through some devices in the circuit.

A driver might be able to cope with that, or it might not, but the lack of immediate failure wouldn't necessarily give you any idea how reliable a driver would be in the case of driving a 2-LED load.

Possibly making best efforts to keep the driver cooled would help the chances of success, but I'd be wary of saying anything that seemed like encouragement to push drivers beyond what it is claimed they can do, especially cheap drivers where components may well have been chosen with limited capacity to do more than the designer intended.

 

[langham]

http://dx.com/p/5v-12v-9w-3-cree-3-mode-circuit-board-for-flashlights-16-7mm-5-6mm-25516 This board will not drive the led's like you want, but it is the type you need. Notice the 9W rating, and that it is designed for multiple led applications. I assume what you are trying to do is make a dive light that has an external battery pack and board, so all you would need to do is ensure that you have some sort of material that is good for heat transfer around the driver. The mosfets are normally the components that overload in these applications, there is a thread that is devoted to dive lights and they have done builds like yours before. It is difficult to find what you are looking for, but you should look for some of the lights that are built with 3*XML and driven at 3A each to see what you need. You could always have an individual driver for each led and a single battery pack with the wiring combined in the same wire shield. I personally would try and find a dependable light that is made by a good manufacturer because you kind of need your light to perform down there right? The cooling shouldn't be a problem since you are under water.

 

[langham]

http://dx.com/p/power-led-driver-co...88?rt=1&p=2&m=2&r=3&k=1&t=1&s=140060&u=156788
http://dx.com/p/t6-3-led-lamp-cap-9-22v-100483
I got this when I looked up what you listed, I only looked at DX, but I think that these are the kinds of drivers you should look at. If there is an issue with heat sinking you should look up potting the led driver. This is how you keep drivers cool and there is a wealth of information here on CPF about the process, it can be expensive though. To address what to look at normally if there is a range that is for input voltage and the output voltage should be a constant value.

 

[uk_caver]

Personally I like to have redundancy in lights where possible, though that's maybe not a huge issue if a light is an extra, large one, rather than the only one someone has (most of the divers I know tend to use lots of lights, since they dive in the dark).

 

[Justin Case]

Most likely the voltage ranges advertised are the INPUT voltage ranges that the drivers can handle. Thus, something listed as 5.0V to 8.0V can basically take battery combinations such as 2x123A, 2xLi-ion, 4xNiMH, etc. A 3xLi-ion combination would be too much and you'd risk frying some component on the driver.

As others have said, if the driver doesn't specifically claim it can handle multiple LEDs in series, I wouldn't try it. My guess is that these cheap, high-powered Chinese driver can barely handle the power draw and heat for a single LED. At 3A drive to a typical LED like an XM-L, the LED will draw about 10W of power. Yes, I see that the 19mm diameter lightmalls driver is claimed to be 90% to 95% efficient, but I don't believe it. I would suspect more like 80% efficiency. That means the driver needs to draw 12.5W from the batteries to deliver 10W to the LED. That's 2.5W of waste heat. If you ran two XM-Ls in series, that would be about 20W of power and the driver would generate about 5W of waste heat.

2.5W of waste heat is normally fairly manageable. 5W is a lot. The issue is that these cheap drivers basically have no thermal management in their design. The components are simply soldered onto a PC board. The only way you can draw heat from key components like the diode, the inductor, and the buck IC is to attach a heat sink to the case of these components. That is a very poor thermal pathway since the case has poor thermal conductivity. Quality drivers such as those from TaskLED and The Sandwich Shoppe use approaches such as silver-filled thermal vias to contact thermal pads built into the components that they select. Then you connect those thermal vias to a heat sink. That is far more effective to pull heat from the board components.

On top of the driver waste heat, you will also have LED waste heat. LEDs are probably only 20%-30% efficient. Even at 30%, a single XM-L can waste 7W as waste heat. Two XM-Ls waste 14W. In a small light, that waste heat is going to go everywhere, including into the driver and its heat sink. The whole works is going to get hot.

For a dive light, maybe your environment can compensate for this waste heat generation. But I'd certainly make sure I tested the light for reliability before trusting it.

If you will depend on the dive light for critical issues, I would spend the extra money to get a TaskLED or Sandwich Shoppe driver. The soldering job on that lightmalls driver doesn't inspire a lot of confidence for me.