1. Home
  2. Electronics Design
  3. LED backlight driver

LED backlight driver

Hi,

A new product I'm designing uses a TFT display with an LED backlight. The LEDs are arranged as 5 series strings of (I think) 9 white LEDs. Each string is spec'd at Vf = 30V @ 20 mA. They also recommend that the currents in each string are matched to within 5% for good brightness uniformity.

The power supply is a wall-wart with battery backup, giving a supply rail of 6 to 9VDC. The brightness needs to be adjustable down to 2 mA per string.

There are several schemes I can think of to drive this, none of them particularly elegant:

  1. Five boost converters (LT3461 or similar), one driving each LED string, feedback pin connected to current-sensing resistor.

  1. One boost converter to provide a regulated 35V rail, then a linear current regulator (op-amp + npn or MOSFET) in the negative end of each string.

  2. Connect all 5 strings in series and drive from a 150V boost converter. This would probably need a transformer rather than an inductor, so we'd be looking at a custom-wound part.

This isn't a particularly cost-sensitive application, but none of the options above look particularly great.

There was a fourth option, similar to (2) but with the strings connected in parallel, then a single current-sense resistor connected to the SMPS feedback pin. I don't physically have the LCD panel yet, but I tried a few strings of Nichia NSPW300's in parallel, with 10R current-sharing resistors, and couldn't get them to balance even within 50%.

Anyone got any better ideas?

Thanks

Rhydian.

Reply to
news
Loading thread data ...

Try more like 100R series resistor on each string, strings in parallel. Visually look at what matching you actually need.. I'm not sure 10-20% is that visible in practice, and I doubt the LED matching is much better.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
Reply to
Spehro Pefhany

OK, thanks, I'll give that a try. The 30V @ 20 mA spec is the only one I have, there is no spec for the Vf temp coefficient so it's hard to predict the worst-case mismatch based on a test of one unit. I'm more concerned about damaging one of the strings if it pulls more than the absolute max current (35 mA), than any visible effects. I should have a sample display to test early next week.

Thanks

Rhydian.

Reply to
news

Not sure ... you could use the 35 V and put a current mirror in there if you need to adjust all the strings simultaneously for you 2 to 20 ma requirement. If you need individual control, an LM317 with a 62 ohm R and

500 ohm pot in series between Vout and Adj will give you about 2.2 to 20.2 ma from the adj pin.

Ed

Reply to
ehsjr

ehsjr wrote in news:zaPgj.5002$O97.4020@trndny01:

the question is;will his LEDs output the same brightness with identical currents? I suspect there will be a significant variation between LEDs.

--
Jim Yanik
jyanik
at
kua.net
Reply to
Jim Yanik

They are probably matched to within some tolerance. A single bin tolerance on LED displays is typically something like +/-20% in mcd.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
Reply to
Spehro Pefhany

That's why the individual control circuit might be preferable. OTOH, he could put a series trimming resistor in each string driven by the mirror.

Or are you referring to different brightness in individual LEDs within a series string of 9 LEDs? Individually controlling 45 LEDs can be done, of course, but it may be impractical for the op.

Ed

Reply to
ehsjr

ehsjr wrote in news:5rZgj.4941$hS.2365@trnddc08:

I suspect that each LED will have a different brightness for the same current. So,if they are in a string,then each LED's brightness will still be different. Matching LED's for brightness may be impractical. It also would be impractical to have a separate drive for each of 45 LEDs. He probably needs some method of diffusing the LED light to even it out.

--
Jim Yanik
jyanik
at
kua.net
Reply to
Jim Yanik

This is a problem for the backlight manufacturer. They're tested and "binned" while they're still bare dice. That's how they can make

7/14-segment displays that are fairly well matched without adjusting the current through each LED.

Ditto.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
Reply to
Spehro Pefhany

On 9 Jan 2008 14:33:09 GMT, in sci.electronics.design, Jim Yanik bloviated:

There are commercial backlights that use embedded LED in a lexan/lucite plate. Plate is *painted* whatever color you wish on backside.

Same idea as a light pipe over a larger surface area.

Reply to
Bingo

The display is a sealed unit; I'm assuming the backlight has whatever optical features are needed to spread the light evenly over the panel. CCFL backlights seem to manage OK with the light source concentrated at one edge. I'll do some tests on the sample unit when it arrives; provided the display *looks* evenly illuminated and none of the LEDs are being over- driven I won't worry too much about meeting the +/-5% spec - it does seem a bit tight.

Thanks

Rhydian.

Reply to
news

^^^^^^ parallel

Reply to
ehsjr

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.