No announcement yet.

Adjustable backlight?

  • Filter
  • Time
  • Show
Clear All
new posts

  • Adjustable backlight?

    I have a Lilliput transflective screen bought through a vendor on this forum. It was advertised as having a 5-step adjustable LED backlight, but that's not the case. The brightness control simply adjusts the image and not the backlight.

    What I was hoping to do was use the illumination lead running to my radio to adjust the backlight of the LCD.

    I'll have to look at the wiring diagram for my car, but I'm assuming that this lead is simply an adjustable voltage based on the dimmer switch position.

    I really know nothing about building circuts but if there's some sort of guide I'm sure I could cobble something together.

    So is there a way to adjust the brightness of the LEDs based on an input voltage to the LCD driver?


  • #2
    Voltage isn't used to dim LEDs (they are current devices).
    But chances are that your vehicle's dimmer is a PWM type unlike old types that were rheostats (big variable resistors - aka pots/potentiometers - that varied voltage).

    A problem may that your backlight is (probably) powered internally from a regulated (voltage) source, hence it's not as straight forward as using the vehicle's PWM dimmer (eg, typically varies from 12V - 14.5V).

    One option is to insert a PWM circuit - eg, a 555 timer with a pot and a few other components (resistors, diode(s), capacitors) - eg, see recommendations for attiny or arduino newb? Reply #23.

    If you want the backlight to track an existing (PWM) dimmer, maybe a Zenor diode circuit could be used (clamps the maximum voltage), or a fast reacting voltage regulator (that doesn't mind operating at the PWM frequency), etc.
    Maybe even a PIC (a $3 PICAXE 08M2 plus circuitry & program) that measures & duplicates the existing dimmer's duty cycle, or analog dimmer's voltage (LCD driver?), or modifies based on ambient lighting, or overrides based on your adjustment (ie, manual & auto modes)...


    • #3
      I'm not quite sure how it is done in my car.

      I see that the potentiometer used for the dash light dimming is wired to the "Body Integrated Unit", but I can't for the life of me see how it adjusts the brightness. The dash illumination appears to be wired directly between the "Tail and illumination" relay and ground.

      I'd rather tap into the dimmer and not use an ambient light sensor. I hate automatic lighting control.

      Also I don't think the backlighting on my LCD is has it's own connection - it's driven by the main ribon cable. The only other wire is the touch panel. This might not be possible at all.

      Vehicle diagram attached. Confusion heightened.
      Attached Files


      • #4
        Alas the diagram gives no hints - it could be an analog or digital (PWM) dimmer. The pot could be controlling a PWM, or a transistor circuit.

        What vintage and model is your vehicle? Maybe someone will know...?


        • #5
          I'm going to guess that it's a Subaru. Just a guess really, but the pdf lists a wagon or H6 model. H6 could be the 6 cylinder boxer engine.

          Anyway, if it's a newer Subaru it uses PWM for dimming. The older models (2006 and older) varied voltage. 12volts at the dimmest setting and 0volt for the brightest (yeah, it's backwards).
 is a good site for Subie radio pinouts.

          PWM or variable voltage I'd recommend not hooking it up to the unregulated voltage of the car. Just using a resistor is a bad idea too (just been there, got the t-shirt).
          OldSpark's suggestion of using a 555 is a good one. You need a little electronics knowledge to get the components right.

          My suggestion is to use a CAT4101 1amp LED driver. You only need a resistor and capacitor to get regulated current. The chip has an input that accepts PWM (up to 5KHz). You might be able to drive that through a zener diode or better yet a PICaxe microcontroler, again like OldSpark recommends. The datasheets have most all the info you'll need to hook things up.
          These little PECaxe controllers are so cool. A few lines of code replaces so much frustration with an analog circuit.

          Check out this thread of mine.
          I just went through replacing the backlights and controller on my system.
          I'll deffer to OldSpark for actual electronics knowledge. I can only relate my experiences with what I've done.

          Good luck,


          • #6
            Great info davekra! Especially about the "upside-down" Subaru voltages - man, what a trap!

            I like the CAT4101 LED driver, though I see that as intended "per LED string". However, not knowing the parallel and series LED combination in the said backlight, I'll defer to you.
            It will work however until a LED or string fails. After that, the others will get brighter, or maybe burn out if the CAT4101 was used to replace resistors.

            My usual method is to stabilise each string and then PWM the lot.
            The string stabilisation might be a resistor (from a regulated voltage if variation is an issue) after taking any PWM transistor or FET voltage drop into account, else a (2-transistor or LM317 etc) current limiter in which case voltages are irrelevant and resistors not required.

            Current variation usually isn't a major problem unless well exceeding LEDs specs - it's just the normal "higher current, lower life" relationship.
            But since the voltage variation can effect brightness (though not as much as with lamps), a regulated voltage solves a lot - eg, a string resistor fixes the max current per string.
            But even I have used a limited range PWM to enable resistor-less (or smaller resistor) LED strings - eg, if the 20mA LED handles 200mA for a 10% duty cycle & max 1ms, a resistor limits to 200mA and the PWM is adjustable from 0% to 10% only at a minimum frequency of 100Hz (10mS hence 1mS at its "full" 10%) [and LED PWMs are usually at least a few hundred Hertz].


            • #7
              It's a 2005 Outback XT, so I suppose it is using varied voltage.

              I still don't know how it is adjusting the dash lights, the diagrams don't show the lights hooked up to the BIU at all. I'm probably missing something.

              The LCD I have doesn't have a seperate wire for the backlighting so I'm not sure how I'd power it anyway


              • #8
                In my Lilliput 701 the original backlight had a short ribbon cable soldered to the controller, which is not the circuit board you can see in my pics. I had cut the traces and soldered to the board rather than unsoldering the fragile ribbon cable.
                It worked for a while using unregulated 12v and a resistor but eventually one or some of the led's burnt out which increased the current to the rest of them and poof.
                My new setup uses regulated 12v from the M2-atx.

                In my display, and I'd suspect most others, the led's are an array of parallel/series. In mine it has 27 led's in a 3s/9p array. There are no resistors or printed resistors and only two wires so isolating the strings would be somewhat difficult.
                9 strings at 20mah a piece gives 180mah total current. I've set my new driver to provide 165mah. There was very little noticeable difference in brightness over 180 and it keeps the led's under 20mah each.

                I'd initially thought about using the dimmer control for backlight control but decided not to. I wasn't sure what kind of signal was coming out and didn't have the tools to look at it.
                Following ambient light turned out to be a really good option. I could customise the PIC program to get things just right.

                This would be somewhat of an advanced project which will require some delicate soldering and disassembly of your display. The led driver and PIC circuit are pretty simple (I think) and we can help with some of that.
                I think you should choose to read ambient light. Start by building the driver/PIC circuit. If you get that working, then start the display modification.
                Last edited by davekra; 03-02-2012, 09:43 PM.


                • #9
                  I really don't want to rip apart my display - it's a $600 transflective unit and I don't want to risk screwing it up. I appreciate your posts though!

                  I just wish it had the backlighting it was advertised to have


                  • #10
                    That would be a refund in this country, and a possible $10,000 "misleading advertising" fine. I too see an image adjustment as being very different to LED dimming. (EG - LED dimming can significantly alter the overall power consumption.)

                    And not that I am familiar with your display nor modern display construction methods, but controllers were usually separate to the display screen itself. The screen (backlight board) would be only the lighting (flouro or resistors & LEDs. Hence the wires to the screen would contain the brightness source.
                    I doubt that such controllers would be now part of the screen even if size permitted (for various reasons), but I could well be wrong. (I should recall the pics I've seen here and elsewhere, but I should recall if backlight controllers were now "on screen". SHOULD...)

                    If I'm right, then adding PWM should merely be the interception of one wire, not the ripping apart the screen section.
                    That however doesn't mean NOT dismantling (ripping apart) the display itself - I'm treating the display as the whole unit as opposed to the mere screen within in case that's what you mean.
                    But I see it self evident that a display unit will not have its screen backlight (voltage) supplied from and external source. It has whatever power input (eg, 12V) which it regulates internally for the screen (and other circuitry). It may have an additional input to enable dimming of its backlight module, but that input will not be the power used for the backlight.

                    I hope that makes sense, but please someone correct my assumptions.
                    Not that it matters if you don't want to open the display enclosure, but it might if you meant the screen with backlight.
                    Only if replacing fex fluoros with LEDs would a physical screen hack be required. (Even then, some backplanes can be easy to separate.)


                    • #11
                      For my display there are two circuit boards. The one you see in the pics is easily removed. You don't really need to disassemble the actual display panel.
                      There is a thicker ribbon cable that connects to the other circuit board.

                      The second board that you can't see is enclosed in the metal that houses the actual glass and backlight.
                      The glass that is the display is permanently attached to that board with fragile ribbon cables.
                      My backlight was soldered to that board with a fragile ribbom cable. The driver for the backlight is integrated into it.
                      I assume that it's a PWM type of supply. I'm sure you could take that power and modify it to dim the led's. I'm not quite sure how you would modify a signal that's already pwm.

                      Getting to that second board does require opening the actual display panel but, in mine, the board was right on the back and I didn't have to remove the glass.
                      Regardless, it can be nerve racking to do if you've not got a lot of experience opening things up.


                      • #12
                        "I'm not quite sure how you would modify a signal that's already pwm"...

                        It must be PWM for a reason, but that could be to minimise the resistor size as I outlined.

                        Regardless, if it's a circuit that generates the PWM duty cycle based on a resistor, that resistor could be potted so to speak (ie, replaced with a pot). Making a voltage-dependent resistor (to respond to electronic signal rather than a pot) can be done, but that adds complexity. (Is that a transconductance amplifier, or should I be embarrassed again?) And a duty-cycle to voltage converter would probably be the input (to the VDR) if required.

                        Many PWMs are set with an R-C combination, but others could be fixed in some program (PIC, controller, etc) or have some other voltage input (LDR, external, etc). That becomes a firmware hack exercise and is usually possible (even if it takes extreme work which is never economically justified) but for practical purposes, it probably isn't possible or practical. (How I loved computers that loaded their OS or peripheral call routines into RAM. A hackers dream....)

                        But I reckon I'll bow out of this. Yet again I have no detailed familiarity with the hardware concerned (crap!, I have zilch familiarity with it!) - but I intended only so provide some ideas or possibilities.
                        In my case, I'd return the display else threaten "misleading advertising" action. Then again, I have been silly enough to attempt workarounds, and they are probably still in my room(s) of unfinished project LOL!

                        Best of luck.

                        And davekra, thanks for your insights and links. You're the man, ... I mean... Poster.


                        • #13
                          Thanks guys for the responses! This is getting more and more over my head. I have no problem replacing an LCD panel or motherboard in a laptop, but I'm just not willing to experiment on this screen.

                          My screen is a Lilliput. There's the main controller board where the VGA is connected, then there's the LCD panel. Embedded into the back of the LCD panel's plastic backing is a thin circuit board where the main board is connected with a 50 pin ribon cable. That "embedded" board in turn drives the LCD and I do now see a smaller ribon cable for the backlight.

                          My previous Xenarc with a CCFL had an entirely seperate inverter board for the backlight.

                          I was poking around in my parent's Dodge truck - the backlighting was clearly adjustable, the brightness wasn't simply image processing. If my screen worked like that I'd have been happy.