Announcement

Collapse
No announcement yet.

Xenarc CCFL brightness reduced after modification - please help me debug

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Xenarc CCFL brightness reduced after modification - please help me debug

    recently I extended the high voltage wires from the inverter to the CCFL on my Xenarc 700TSV. I'm still alive after the experience and so is my Xenarc. I used hi-flex PVC insulated test lead cable. The extended cable run is about 8 inches long. There's no crackling, or arcing that I can see going on but my fear is that somehow the voltage is being lost along the run.

    My problem is the CCFL brightness now seems to be low. I like to be as precise as possible about things but I have no way of measuring how low it actually is. My Xenarc has been in bits for best part of 3 months so I can't remember how bright it should be. I know these screens are not exceptional in this respect (hence the upgrade project), but this just seems too dim.

    Using the OSD menu I have checked that the backlight setting is on 99. Indeed reducing it reduces the backlight intensity even further so the control mechanism seems to be working.

    I've read about people saying that the Xenarc benefits from being supplied from cleaner power supply than may be available from a car, so I'm supplying it from the output of a 12V linear regulator. It says in the manual that the screen's minimum DC input voltage is 11V. I've just measured the input on my screen and it falls to 10.2V when everything's on. I think this is to do with the regulator's internal loss as my bench power supply is 12V. Technically this is out of spec so maybe that's my problem. This is something I'm going to investigate and I'll probably try bypassing the regulator on the bench to see if that extra 1.8V makes any difference to the brightness. Anyone have any anecdotal evidence that backlight brightness on a Xenarc is a function of input voltage in the range 10-13V?

    I suppose I really need to be able to measure the output voltage from the inverter, but I suspect that's not easy with basic test equipment. Maybe there's some kind of resistor network I could apply to it to scale the voltage range down so that it's measurable with a standard digital meter? That then begs the question:- what voltage am I expecting to see out of the inverter?
    If that method works then I guess I could then tap off the voltage at the CCFL end of the extension and see if there's been a voltage drop alon the wires. BTW are CCFLs supplied with AC or DC? and if it's AC, what sort of wave is it?

    The other thing I wondered was whether there is any way I can probe the (low voltage) 4-pin input connector to the inverter board to test for expected voltages. I don't know the pinout of this - maybe someone could assist. I assume it's got to be a supply rail, a ground, a brightness analogue/PWM input and something else? I just wondered if it was this side of the inverter where the problem lies.

    Any help greatly appreciated.
    James

  • #2
    the wires from inverter to buld may not be lengthened on a ccfl tube
    time for the sig:

    '00 VW golf TDI upped to 130 HP - 18" RS 6 wheels - liliput 7" tft in indash housing - epia 10K - SB audigy - caliber 1 fahrad CAP - RF punch amp - hollywood 1.6K mono amp - boston acoustics pro - RF 10" DVD sub

    Comment


    • #3
      Agreed. Those wires need to be as short as possbile! They're very high resistance wires...
      Current Setup

      Nexus 7 Tablet
      Built in head unit
      Brutus 1500Watt Amp
      Two Boston 12" subs

      2006 Infiniti M45 Sport ;) My 4 door Camaro... lol

      Comment


      • #4
        okay, I think I can help.

        1.) 10.2v will result in reduced brightness. 11v is the minumum, but even that is low. you need to have 13-14v for ideal brightness.

        2.) lengthening the wires is fine, as long as you use good enough wire. you seem to have used the proper wire, as yours is rated at 1500v. I do ask, though, why didn't you lengthen the low voltage side?

        @NRG: the wires pass a few milliamps. Ohm's law:

        V=I*r

        even if the wires were 10 ohms, which would be terrible:

        10*.005=.05v

        Hardly an impact when you're talking 600-800 VAC

        3.) If you have a fluke DMM or similar, you can test. It should be around 1000v starting, and 700v running. You can apply a resistor network if you're careful, but chances are you won't maintain 1000v of dielectric insulation.

        4.) 12v, ground, ground, 3-5v PWM for brightness.

        good luck

        Comment


        • #5
          greenman100:-

          thanks so much for your input on this, I've definitely got things to go at now.

          Unfortunately I can't lengthen the low voltage side because of space constraints around the LCD module. I physically need to get the inverter away from the module. It's for a homebrew Xenarc to motorised in-dash conversion so there's just no room on the sliding plate to put the inverter.

          the pic makes it clearer...

          Comment


          • #6
            I have personally extended wires about 8", but I also replaced the inverter while I was at it.

            Let me know how it goes for you. If need be, I could swap inverters with you, and give you my cabling

            Comment


            • #7
              Well I went out and bought a decent bench supply with variable voltage to see what effect varying the Xenarc input voltage has on the backlight brightness.

              I found that:-
              • The Xenarc doesn't regulate the supply voltage to the inverter, it just passes it the supply voltage directly.
              • CCFL brightness only started falling-off below 11V (this agrees with spec) and was essentially constant up to 15V
              • Rapid changes in input voltage (say from 11-13V) would cause momentary increase in CCFL brightness. I conclude from this that the inverter has a slightly noticable feedback loop that corrects for input voltage fluctuations. This highlights the point made in the 'powering your LCD monitor' sticky that you may see brightness 'blips' as your engine revs (& hence supply voltage) change.



              For the second graph I monitored the brightness voltage pin whilst adjusting the Xenarc's On Screen Display control for backlight (an integer from 0 - 99). The actual curve isn't as linear as that but you get the point.

              Remember I did this to debug my dim screen so I'm sort of hoping someone might notice an anomaly with my results. I'd particularly like to validate the transfer characteristic of the brightness control voltage.

              Unfortunately for me though I didn't solve my reduced brightness problem so I'm going to rip the whole thing apart and re-locate the inverter close to the LCD to see if it's my wiring

              James

              Comment


              • #8
                hey good work with the plotting the inverter stuff.

                The control voltage range will be 0V for full and 5V for dimmest as you have seen.

                I would not recommend testing the inverter with anything rated for less than 1500V just in case. The Vstart should not be over 1000V but you never know. You could always try and test the votlage once the tube is running as then it should be about 600V but as we dont know what it shoudl be this is pretty pointless.

                the tube brightness is controlled by a PWM output so you would need to check this anyway.

                I would say it's probably the way you have joined the wires. You should get some proper high voltage inverter wire and properly insulate any joins. Also be careful how you route the cables as metal can cause losses. up to 6-8" should be fine to be honest. I know this may seem long but i was amazed when i was told it's fine.

                It is interesting to see the xenarc doesnt regualte the input voltage. The inverter should be run at 12v. Any more and it is shortening the life of the CCFL.

                Comment


                • #9
                  I just noticed you are in the UK from the RS link. you could see if you can get some proper CCFL cable or CCFL extensions with proper connectors on each end from visualux in cornwall.

                  Comment


                  • #10
                    Originally posted by jwexqm
                    greenman100:-

                    thanks so much for your input on this, I've definitely got things to go at now.

                    Unfortunately I can't lengthen the low voltage side because of space constraints around the LCD module. I physically need to get the inverter away from the module. It's for a homebrew Xenarc to motorised in-dash conversion so there's just no room on the sliding plate to put the inverter.

                    the pic makes it clearer...
                    I have had the same problem when I did my indash screen.
                    You have a private message.

                    It seems that you have a motorized mechanism.
                    From wich unit have you take it?
                    personnaly I have taken it from this unit:


                    New prototype : 2DIN AUDI OEM-LIKE CARPC based on SYMPHONY II
                    Motorized InDash Screen, Core 2 Duo 2.2Ghz, SSD

                    Home Made In-Dash 7" Lilliput V1
                    Home Made In-Dash 7" Lilliput V2 = LM-Lilliput Prototype

                    First project in Peugeot 306

                    Comment


                    • #11
                      Right I've just found out why my high voltage wire extension is causing the CCFL brightness to be reduced. Here's what I did...
                      • Rip the screen apart and remove the extended segment of the high voltage cable.
                      • Switch on and observe that brightness is back to normal (good, I suppose).
                      • Reintroduce an equivalent length of extended high voltage wire on the bench
                      • Switch on and observe a slight decrease in brightness but nowhere near as bad as when the wire was routed through the in-dash enclosure
                      • While switched on, prod the extended cable around with a plastic pen and observe that the brightness is reduced when the cable comes close to the grounded chassis. ('aha' says Scouse)
                      • Observe a marked decrease in brightness when ~4cm of the extended cable is taped parallel to the grounded chassis
                      • Then I get really creative and take a 2cm x 1cm piece of of aluminium foil and wrap it round a part of the extended cable and then offer a ground lead up to the foil...
                        I didn't even get to touch it the spark jumped about 1mm and rebooted the Xenarc.
                      • added one wrap of insulation tape around the wires and repeated the foil test.
                        no spark this time but CCFL still momentarily went off when I touched the foil
                      So my conclusion from all this is that it's a capacitive effects that are causing my brightness problems. I'm not too hot with capacitor theory, going to have to dig out my course notes , but I know they pass AC pretty well which would explain this completely.

                      I'm now thinking my cable is sh1t!

                      So I started looking at the PC case modding community as they use CCFLs to pimp their rigs. Saw this and wondered if the cable and sleeving they've used in this product will be any better than my test lead wire. Too late, just ordered it. I'll be calling Visualux too tomorrow. Words of encouragement please.

                      James

                      Comment


                      • #12
                        it's nto that it's capacitative, it's that the cable's insulation is breaking down. you need better cabling. Perhaps cabling from a cheap computer case CCFL?

                        Comment


                        • #13
                          and the problem with your extension is that it has the wrong ends

                          Comment


                          • #14
                            Dude. As I said before - visualux if you are in the UK (mouser or digikey if you are not).

                            they will have the proper wire and proper connectors for you to make an extension with.

                            We got this very same problem with the backlight upgrades from the metal chassis. SO much so it actually ended up blowing an LCD (and Jerry) up (sorry jerry, had to laugh)

                            Comment


                            • #15
                              Originally posted by Scouse Monkey
                              Dude. As I said before - visualux if you are in the UK (mouser or digikey if you are not).

                              they will have the proper wire and proper connectors for you to make an extension with.

                              We got this very same problem with the backlight upgrades from the metal chassis. SO much so it actually ended up blowing an LCD (and Jerry) up (sorry jerry, had to laugh)
                              Scouse,

                              I've seen the "backlight upgrade" thread is closed...Have you open an other thread?
                              New prototype : 2DIN AUDI OEM-LIKE CARPC based on SYMPHONY II
                              Motorized InDash Screen, Core 2 Duo 2.2Ghz, SSD

                              Home Made In-Dash 7" Lilliput V1
                              Home Made In-Dash 7" Lilliput V2 = LM-Lilliput Prototype

                              First project in Peugeot 306

                              Comment

                              Working...
                              X