I was reading on (W)LED drivers and I never see resistors with the leds in the app notes. The LED Strings are connected directly to the chip. For example, the O2 Micro one (which some believe is close to what is in the iPad circuitry) OZ9956C http://www.o2micro.com/prods/o2m_int_ltg_brochure.pdf

Is there a way to measure the specifications of LED Strings — at least just enough info to start making a driver? (forward voltage, etc.)

To story here is that I am trying to find a DIY-friendly/cheap solution to drive the backlight portion of various aftermarket LED LCD panels. When I say DIY-friendly, I mean either off-the shelf ~20$ board or not to difficult to build from someone that just got himself an iron from radio shack (no TQFP soldering please).

For example, the following LCD Panel (the iPad screen):
http://www.beyondinfinite.com/lcd/Library/LG-Philips/LP097X02-SLA3.pdf

This one has 6 strings (of 6 leds). We have access to the “cathode” of each strings (and a “common anode”). The datasheet is not very verbose in terms of specification or they are not very clear.

Thanks!

——-

Side question: I recall that some LED driving circuit uses some sort of high-frequency to drive leds (for efficiency reasons)? Did I day-dream this?

Input DC 4.8v – 5.6v
Led 3 x 3W, working voltage 4-5v each. So had to wire in parellel
Prefered current 0.8A-0.9A

Please email me ngwanhoe@yahoo.com

Thanks a lot

Thanks very much for the speedy reply. I will keep digging around and follow up on your suggestions. Incidentally, I realized that my circuit, using .5 uF Cap on the + and 1K ohm / 250V resistor on the backend to deal with inrush, is basically finished for use on a 120V AC main. I am going to add 2 Zener Diodes rated for 110V to regulate spikes and then I think I am all set. I think that for mobile solutions though, your method may work perfectly and the gel cells are a nice idea. Thanks very much.

I’d be leery about the 40V supply being regulated well enough to have enough headroom and would feel better about 30 LEDs per string (36V). Or 20 LEDs per string (24V) would give you more flexibility in power supply selection (28-30V laptop supplies are fairly common) and five strings would get you 100 LEDs (six would get you 120) on two EasyBrights.

If you want to put gel cells into your project, two would give you nominal 24V with at least 22V before they’re too discharged to be useful. 16-17 LEDs would be about 20V per string and six strings would cover your 100 LEDs.

Regardless of the supply and the number of LEDs per string, each string only draws the 20mA of the individual LEDs, so you’re looking at 100-120mA for the whole schmear. On a series pair of 7Ah gel cells, that’s at least 5 usable Ah = 5000mAh ~= 50 hours.

I recently collaborated on an art installation containing one used, two-year-old gel cell powering quite a few LEDs, two IR proximity sensors, and a motor. It’d go about 20 hours of light use to 5 hours of intense motor activity before winding down and needing a charge.

Hopefully some of these numbers give you ideas about avenues to pursue?

I am building an DSI (surface illuminated) display for multitouch tracking. I’ve been building an AC circuit using capacitive reactance to drive the ~100 IR leds that need to run the outside of the surface. Each LED has Vf of 1.2v @ 20mA optimal current. I’ve only just realized the value of constant current drivers but was wondering if/how your solution would work for me? Would I need to hook up a whole bunch of small strings or could this drive the whole array?

These LED’s will likely have to be on for a long time so my guess is that battery power is not useful.

Your thoughts are very much appreciated.

Scott
PS – I know that I can buy a strip with prebuilt driver but A. they are expensive and B. I would like to build the tech from the ground up the first time. Thanks Again.

To echo some other comments, awesome site, great descriptions and I wish I knew half of what you could have discarded as useless knowledge. My interest in electronics is poorly funded in the ole grey matter department.