1
\$\begingroup\$

Ok, for starters I haven't been able to find a lot of solid details about the LEDs I have here. All I know is that there are 32 of them; 8 LEDs per strip, in series, and 4 strips, each also in series. Through my poking and prodding, I've managed to determine that the LEDs are probably around 3.6-3.7v and 150mA. The power circuit board in the TV was very well labeled and it specified 118V and 260mA at the LED connector, so I have to assume that means that each LED gets 3.6875v and 260mA. Unfortunately, the power board was broken (which is probably why the TV was in the dumpster...)

So, my question is, how the heck can I get 118v and 260mA? My initial thought was to use a resistor and rectified AC voltage from the wall, which would give me around 170v. That wouldn't work though, because $$\frac{(170 - 118)}{260} = 200Ω$$ and $$0.260^2 * 200 = 13.52W$$ Good luck finding a resistor that can handle 13.5W, and that's SUPER inefficient.

I have a basic understanding of electronics, but this kinda stuff confuses me for some reason. I'm learning more every day, I'm just stuck figuring this out. What would you do in this situation?

Edit: If it helps at all, I found one of these in my parts bin:

https://www.fairchildsemi.com/datasheets/FQ/FQA11N90_F109.pdf

900 volt N-channel MOSFET. Could this be useful?

Edit 2: I bought a DC-DC buck converter, 1.5-30V adjustable. Hooked it up to 32V, adjusted it down to 28 volts, then hooked it up to one of the light strips. 3.5V each and it's using around 200mA. Oddly, when I put more of the strips in parallel the current doesn't increase linearly. All 3 strips only draws about 400mA.

\$\endgroup\$
14
  • 3
    \$\begingroup\$ "What would you do in this situation?" I'd put it back in the dumpster. \$\endgroup\$
    – Roger Rowland
    Sep 12, 2016 at 12:43
  • 2
    \$\begingroup\$ Cut the strip in more manageable pieces and wire them in parallel. \$\endgroup\$
    – Arsenal
    Sep 12, 2016 at 12:46
  • 2
    \$\begingroup\$ @Roger Rowland Or experiment with it until you break it. Then put it back in the dumpster. It's more educational and potentially more fun as well. \$\endgroup\$
    – Dampmaskin
    Sep 12, 2016 at 12:47
  • \$\begingroup\$ Are you sure there aren´t any other components on the board - and that they are all in series? A picture would help. 260mA per LED for a backlight doesn´t sound right. \$\endgroup\$
    – F. Bloggs
    Sep 12, 2016 at 12:53
  • \$\begingroup\$ .75W per led, yeah that sounds okay. \$\endgroup\$
    – Passerby
    Sep 12, 2016 at 13:03

4 Answers 4

Reset to default
2
\$\begingroup\$

I expect that the supply has been a switched mode constant current supply. Output 260 mA resulting in a 118 V when all the leds are in series.The open voltage of such a supply is most probably higher. An LM317HV in constant current mode with a supply of 170 V could be a simple solution but does not solve the inefficiency. If a high voltage switching supply is difficult you could try to connect the strings in parallel with 4 LM317 in constant current mode ( maintaining 260 mA) for each string and use a 33 to 35V DC power source for all of the strings. In that case the efficiency would be acceptable.

\$\endgroup\$
2
  • \$\begingroup\$ Ooh, I hadn't looked into the LM317HV. Efficiency isn't really important to me right now, so much as just getting them lit. I seriously considered changing things around so that the strips are in parallel (but the LEDs would still be in series). From my understanding, using an LM317 as a constant current source for currents above 200mA isn't recommended, right? My problem is that I can't test the LEDs to see how they act with various currents, because I don't have an adjustable constant current power supply, only a PC power supply with set voltages. \$\endgroup\$
    – HaLo2FrEeEk
    Sep 12, 2016 at 14:35
  • \$\begingroup\$ Also, just looked up the LM317HV datasheet, it looks like it can only handle a 60v differential, so that would be asking a lot of it to give me 118v from 170. \$\endgroup\$
    – HaLo2FrEeEk
    Sep 12, 2016 at 14:38
0
\$\begingroup\$

Easiest solution would be to cut the strips at their 8 led sections. With 4 sections, that's 29.5 Volts, just right for any common led driver. And even some common wall warts. You could drive them at a lower voltage as well, so a 24V laptop power can be used. Alternatively a boost regulator or a boost led driver can be used with any suitable lower voltage supply.

\$\endgroup\$
1
  • \$\begingroup\$ The strips themselves consist of 8 LEDs in series, but they're detachable from the main "bus board" which itself arranges each of the strips in series. My idea was to modify the bus board to be in parallel, but I found something in my parts bin that might be useful: a 900v N-channel MOSFET, specifically an FQA 11N90. I've updated the original post with a link to the datasheet. \$\endgroup\$
    – HaLo2FrEeEk
    Sep 13, 2016 at 23:47
0
\$\begingroup\$

It would seem that a series buck regulator to drop 60V from a bridge cap source with CC regulated is the most efficient. Adjustable CC.

edit

$$\frac{(170 - 118)}{260} = 200Ω$$ is a linear supply but if you switched a $$1Ω . MOSFET$$ rated for more than the difference of voltage of 52V with a duty cycle of 1/200 , it would now be $$200Ω$$. Then with a series choke to smoothen the current, you now have a Buck regulator except the duty cycle is modulated to match sensed current with a PWM

\$\endgroup\$
4
  • \$\begingroup\$ I updated the original post with a link to the datasheet of a 900v n-channel MOSFET that I found in my parts bin. Could this be useful? \$\endgroup\$
    – HaLo2FrEeEk
    Sep 13, 2016 at 23:52
  • \$\begingroup\$ that's overkill \$\endgroup\$
    – Tony Stewart EE75
    Sep 14, 2016 at 3:00
  • \$\begingroup\$ Isn't 900V just the rated maximum? Doesn't that mean that I could use it for this just fine, because it's rated much higher than the 52v difference? I don't have a large collection of parts of anything, I just moved and had to throw out a lot of stuff, so I mostly just have what I've pulled off of circuit boards, and that mosfet is what I have. Even though it's overkill, would it work? \$\endgroup\$
    – HaLo2FrEeEk
    Sep 15, 2016 at 1:06
  • \$\begingroup\$ if you are only interested in salvaging parts rather than a optimal design, I can't help. In integrated Buck CC regulator is required with a PCB \$\endgroup\$
    – Tony Stewart EE75
    Sep 15, 2016 at 1:41
0
\$\begingroup\$

Just as CCFL backlights had special CCFL all in one IC driver chips, LED backlighting also has constant current LED IC driver modules - I also believe some do PWM and some other smarts. Anyway online shops are selling LED TV backlight testers for about $20. If just one single led dies, and the strips are in series, no backlight. It seem high brightness can shorten LED life, especially if marginal cheaper ones were used in the design. 2nd hand ones may also have a reduced lifespan.

When repairing , consider adding heatsink paste if you have cheaper quality 2/3rd inch wide strips with clear plastic buttons. The glue placement strips mean those raised of the backplate miss thermal cooling. I also added a hole and threaded test wires to light some strips if one failed again. And when open, consider a speaker upgrade. Most have 3 speaker size mount points, meaning a quick screwdriver upgrade.

Recycle Hint. Even junked ones contain perspex plates or diffuses. Useful for picture frames,greenhouses or photographers

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.