Announcement

Collapse
No announcement yet.

Using a resistor to fake a Inverter into thinking a CCFL is good

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

    #21
    Re: Using a resistor to fake a Inverter into thinking a CCFL is good

    Nice approach! Care to share a picture of the inverter?

    Reason I ask is because I've done something similar on a Royer oscillator design in a Dell E193FPc LCD. That design had two inverter transformers and each transformer was controlled separately by a PWM controller. Unfortunately, one of the inverters kept killing the driver transistors, so I decided to disable it. Did that by bridging the op-amp feedback pins on the PWM controller for the faulty inverter. Then removed the dead driver transistors. The monitor has been working like that for about a year now.

    Right now, though, I have a different monitor that uses a single transformer which has only two pins on the high voltage section - one for each pair of CCFLs. I'm still trying to find a way to trick the feedback on the inverter. Seems impossible with this design. Hence why I am interested to see pictures of your inverter.

    Comment


      #22
      Re: Using a resistor to fake a Inverter into thinking a CCFL is good

      Sorry, I already returned the TV to the customer and I didn’t take pictures, but this was few days ago and still remember the main details:

      The Inverter board (it was an old Beko NR20CDLE5) had a single controller IC (OZ964GN, check datasheet), plus 2 pairs of SOP8 dual enhancement mosfets. Each pair drives 3 transformers with a lamp, making a total of 6 CCFL lamps. I just followed a pair of tracks going from each lamp connector back to the central circuit. Then I noticed 6 groups of identical SMD components. I assumed they concentrated feedback from each lamp to a single feedback signal for the main IC. Just bridged between the first components attached to those lines.

      Check page 10 of OZ964GN datasheet (Reference application). Those two lines going back to pin 9 (current feedback) and pin 2 (voltage feedback). That’s the key.

      So far, no complaints from the customer (two days ago).

      By the way, anyone knows why they design inverters to behave this way? I mean, I understand that the inverter has to protect itself from overheat of risk of fire, but what’s the point of preventing it from working when one lamp is disconnected?
      Last edited by Chungalin; 07-05-2014, 06:40 AM.

      Comment


        #23
        Re: Using a resistor to fake a Inverter into thinking a CCFL is good

        Originally posted by Chungalin View Post
        By the way, anyone knows why they design inverters to behave this way? I mean, I understand that the inverter has to protect itself from overheat of risk of fire, but what's the point of preventing it from working when one lamp is disconnected?
        I'm far from an expert, but I=V/R, so if resistance of load decreases the current will increase in the relevant part of the circuit, hence increased heating effect?
        Dell E7450 | i5-5300U | 16GB DDR3 | 256GB SSD

        Comment


          #24
          Re: Using a resistor to fake a Inverter into thinking a CCFL is good

          Originally posted by Chungalin View Post
          Sorry, I already returned the TV to the customer and I didn’t take pictures, but this was few days ago and still remember the main details:

          The Inverter board (it was an old Beko NR20CDLE5) had a single controller IC (OZ964GN, check datasheet), plus 2 pairs of SOP8 dual enhancement mosfets. Each pair drives 3 transformers with a lamp, making a total of 6 CCFL lamps. I just followed a pair of tracks going from each lamp connector back to the central circuit. Then I noticed 6 groups of identical SMD components. I assumed they concentrated feedback from each lamp to a single feedback signal for the main IC. Just bridged between the first components attached to those lines.

          Check page 10 of OZ964GN datasheet (Reference application). Those two lines going back to pin 9 (current feedback) and pin 2 (voltage feedback). That’s the key.
          No need to be sorry. This is very useful info, actually and I really appreciate it .
          The monitor I have has OZ960 controller, which appears to have a nearly identical reference application circuit. The only thing I still can't get my head around to is how they are driving 4 CCLFs with just a single transformer with a single HV winding that appears to have its both ends connected to the CCFLs and no connection to ground.

          Anyways, I still might try what you did. Maybe hopefully it will work for my design too. Although I am suspecting the HV winding on my inverter has shorted turns.

          Originally posted by Chungalin View Post
          By the way, anyone knows why they design inverters to behave this way? I mean, I understand that the inverter has to protect itself from overheat of risk of fire, but what’s the point of preventing it from working when one lamp is disconnected?
          If one lamp is disconnected, you can have a much higher voltage on the other lamp, and I'm guessing that can cause the HV transformer winding to burn out or damage the other CCFL. Similarly, if there is a short on the HV line due to a faulty CCFL, the inverter will be able to detect it and cut power out to prevent from any other components going bad. But that's just all my guess.

          I have seen an open-loop inverter design (i.e. no protection on the HV lines) in a very old LCD monitor. The inverter never gave any problems. Had to take down that monitor, though - the CCLFS were so old that their brightness output was inadequate to see anything on the screen. If I am not mistaken, that monitor served for about 10 years of nearly 24/7 operation. Quite impressive, IMO.
          Last edited by momaka; 07-05-2014, 06:09 PM.

          Comment


            #25
            Re: Using a resistor to fake a Inverter into thinking a CCFL is good

            Hi, my Samsung 226BW recently acquired the "2 seconds to black screen" syndrome and after exhaustive analysis, I concluded that one of the four CCFL lamps was drawing too little current to satisfy the detection circuit (SEM2005 chip)

            Once I realized that pin 3 of this chip was upset because the input voltage there was at 1.18 volts instead of the minimum of 1.3 volts, all I had to do was determine which of the 4 cathodes belonging to the 2 diode pairs D201 and D202 was the origin of the problem.

            It turned out that D201 had good inputs to it (1.85 volts), but one or both of the inputs to D202 was bad (0.72 volts). Since there were 4 mostly awkward combinations in which the 4 CCFL cable plugs could be connected to the 4 sockets, I tried all the combinations and when lamp number 4 was attached to socket number 4, D202 went low, and when lamp number 4 was attached to socket number 1, D201 went low. No other of the 3 lamps was responsible for pulling any diode to low. Conclusion, lamp number 4 is either drawing no current or too little to satisfy the SEM2005 chip, so the chip shuts everything down after 2 seconds.

            I should note that to do this testing, I needed to keep the inverter running by overriding the protection circuit. With the scaler board disconnected from the inverter board, I soldered the + lead from a 3 volt DC battery pack to pin 9 of CN02 (9 line cable connector on inverter board), and soldered the - lead of the same battery pack to pin 4 of the same connector. That fooled the SEM2005 chip into believing that the missing scaler board was requesting the backlight to come on. Then I soldered a wire from pin 1 of the SEM2005 chip to ground via an 82K resistor. That disabled the protection circuit and the backlight remained on indefinitely, allowing me ample time to do my measurements.

            I could not find any videos on the internet showing someone removing a CCFL from this or a similar Samsung monitor, and I had read horror stories about problems arising during disassembly and reassembly of the backlight panel, so I decided that I would prefer to simply run the monitor with 3 working CCFLs.

            I could have kept running the monitor with the protection circuit disabled, but I would prefer to have the benefit of the features of the SEM2005 in case there would be a catastrophic failure of the lamps in the future, with a risk of overheating or worse.

            So I measured the peak power being drawn by the monitor at the 120 VAC input to be 3 watts idling and 21 watts with the 3 good lamps and 1 sick lamp running.

            So any substitute resistor would have to accept 6 watts of heat (21 overall watts - 3 idle watts divided by 3 good lamps).

            To estimate the lamp current I noted that a good lamp will cause the cathode of D201 to be at 1.85 volts DC in my case. This half-wave DC voltage is developed across C214 and is due to high frequency current flowing through the CCFL and into series resistor R221 (910 ohms). To generate 1.85 peak volts across a 910 ohm resistor, the peak current must be about 2 ma. The rms current would be about 1.5 ma.

            So if my substitute resistor is to heat up to 6 watts due to 1.5 ma. of current, it has to have a value of 2.7 meg ohms.


            The attached circuit is for a different but similar Samsung monitor with 4 lamps. I could not find the diagram for the 226BW monitor.
            Attached Files
            Last edited by montreal; 10-16-2017, 11:19 AM.

            Comment


              #26
              Re: Using a resistor to fake a Inverter into thinking a CCFL is good

              6W at 1.5mA, that means the E = 4000V which could not be.
              The firing Voltage of the lamps is around 1200~1500V and the running Voltage is about 700~1000V.
              Looking at you board, you see the Voltage chart shows lamps running Voltage of 650V AC 1.5mA for each connector CN301~304.
              What is the model number of the LCD panel? You can find the spec sheet for the panel to verify if it matches the board spec or not.
              Pictures provided by shojimmy. BN44-00124E as shown in the schematic is your board, right?
              http://www.shopjimmy.com/samsung-bn4...t-inverter.htm
              BTW, CCFL has dynamic resistance:
              https://cdn.badcaps-static.com/pdfs/...ad05cafd7d.pdf
              https://www.microsemi.com/document-p...sign-reference
              Attached Files
              Last edited by budm; 10-16-2017, 09:54 PM.
              Never stop learning
              Basic LCD TV and Monitor troubleshooting guides.
              http://www.badcaps.net/forum/showthr...956#post305956

              Voltage Regulator (LDO) testing:
              http://www.badcaps.net/forum/showthr...999#post300999

              Inverter testing using old CFL:
              http://www.badcaps.net/forum/showthr...er+testing+cfl

              Tear down pictures : Hit the ">" Show Albums and stories" on the left side
              http://s807.photobucket.com/user/budm/library/

              TV Factory reset codes listing:
              http://www.badcaps.net/forum/showthread.php?t=24809

              Comment


                #27
                Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                Originally posted by budm View Post
                6W at 1.5mA, that means the E = 4000V which could not be.
                What is the model number of the LCD panel?

                You can find the spec sheet for the panel to verify if it matches the board spec or not.

                BN44-00124E as shown in the schematic is your board, right?
                P=I x I x R

                or

                R = P / (I x I)

                R = 6 watts / (0.0015 x 0.0015) = 2,727,273 ohms

                V - I x R = 0.0015 x 2727273 =4091 volts, so you are correct.

                The average current through R221 is probably much less than 1.5 ma. and likewise the average voltage across the CCFL, however to get the 1.85 volts DC held across C214 which I was reading today, the peak current through R221 had to have been 2 ma, meaning 1.5 ma rms, and a peak lamp voltage of 4000 volts.

                Once D201 is reversed biased off, the voltage across C214 doesn't really have a resistive path to ground, so the capacitor will maintain the highest voltage it receives, even if that peak lasts for microseconds. I believe that peak lamp voltage had to have been 4000 VAC and I'm guessing that the peak was repeated hundreds of times a second in sync with the PWM frequency of the inverter operating with a duty cycle close to, but never 100%.

                Even though the inverter frequency is 50 khz, I am measuring a DC voltage across C214 because D201 gets rid of the AC, and my digital multimeter can give me an accurate measurement.

                However we need to lower the dummy power resistance to a point where the average current through it, whatever that might be, heats the resistor by 6 watts, so that the normal draw of the missing CCFL is simulated.

                The schematic shown in my previous post does not match exactly my board. My board is BN44-00182B.

                I'll post the panel model number tomorrow.

                Thanks for your comments budm....
                Attached Files
                Last edited by montreal; 10-16-2017, 11:22 PM.

                Comment


                  #28
                  Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                  Originally posted by budm View Post
                  What is the model number of the LCD panel?
                  Here is a photo of the back of my panel.

                  Attached is the spec for my M220EW01 panel.

                  I appreciate better what you were saying about the dynamic resistance of the CCFL.

                  For the SEM2005 chip to believe that a healthy CCFL is in place when in fact there is a substitute fixed resistor, then the resistor will have to mimic the characteristics of the CCFL if detection circuit is to function properly and the transformer is to see the same load as a real CCFL.

                  According to the spec (Note 6), the panel wattage is 23.5 watts, or 5.875 watts per bulb. The standard current ISCFL is 7 ma., so the VCFL is 839 volts. The average resistance must be 120000 oms, which is 20 times lower than the resistance I originally proposed.

                  Using this resistor, the voltage across C214 should be about 9 volts, which will shutdown diode D201 and allow pin 3 of chip SEM2005 to drain to ground via R213 only, and not drain to ground via R216 and R221.

                  Therefore the voltage at pin 3 (likely about 1.85 volts) will convince the chip that there is no open lamp situation and the panel should operate.

                  Thanks for your help.
                  Attached Files

                  Comment


                    #29
                    Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                    Your schematic of BN44-00124E does not match your board, you board has two inverter transformerS, the schematic has one inverter transformer and it matches the pictures of the board I posted so the part designators in the SCH do not match your board BN44-00182B.

                    Another thing about using the resistor is that the break down Voltage of the resistor has to be higher than the applied Voltage.
                    Last edited by budm; 10-17-2017, 08:53 PM.
                    Never stop learning
                    Basic LCD TV and Monitor troubleshooting guides.
                    http://www.badcaps.net/forum/showthr...956#post305956

                    Voltage Regulator (LDO) testing:
                    http://www.badcaps.net/forum/showthr...999#post300999

                    Inverter testing using old CFL:
                    http://www.badcaps.net/forum/showthr...er+testing+cfl

                    Tear down pictures : Hit the ">" Show Albums and stories" on the left side
                    http://s807.photobucket.com/user/budm/library/

                    TV Factory reset codes listing:
                    http://www.badcaps.net/forum/showthread.php?t=24809

                    Comment


                      #30
                      Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                      Since each transformer of your board drives two lamp, then you may be able to join the two lamp circuit detection together to fool the circuit in thinking it has to lamps connected.
                      Or use T5 4W mounted on the inside of the monitor.
                      https://www.badcaps.net/forum/showth...&highlight=cfl
                      Never stop learning
                      Basic LCD TV and Monitor troubleshooting guides.
                      http://www.badcaps.net/forum/showthr...956#post305956

                      Voltage Regulator (LDO) testing:
                      http://www.badcaps.net/forum/showthr...999#post300999

                      Inverter testing using old CFL:
                      http://www.badcaps.net/forum/showthr...er+testing+cfl

                      Tear down pictures : Hit the ">" Show Albums and stories" on the left side
                      http://s807.photobucket.com/user/budm/library/

                      TV Factory reset codes listing:
                      http://www.badcaps.net/forum/showthread.php?t=24809

                      Comment


                        #31
                        Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                        Originally posted by budm View Post
                        Since each transformer of your board drives two lamp, then you may be able to join the two lamp circuit detection together to fool the circuit in thinking it has two lamps connected.
                        Or use T5 4W mounted on the inside of the monitor.
                        https://www.badcaps.net/forum/showth...&highlight=cfl
                        Thanks for your idea, but I don't have enough room to hide a T5 4W.

                        It would be very easy to prevent the detection circuit attached to pin 3 from sensing that lamp number 4 is drawing less current than desired. I just have to remove R219 and that way, one leg of D202 would float.

                        I know that my lamp number 4 draws less current than it should, but I don't think it is drawing no current at all because one time in fifty, when I turn on the monitor, the panel does not shut down after 2 seconds, but stays on indefinitely.

                        I would still be forced to leave lamp number 4 connected so that the transformer sees a partial load instead of no load at all.

                        But this thread is all about using a fixed resistor to substitute for a disconnected CCFL lamp. So why wouldn't a 120K ohm 10 watt resistor work?

                        47 years ago I used to build electronic strobe lights for entertainment purposes. Most of my productions ran on 800 vdc. But I had a client who asked me to build a device designed around the airport runway strobe light, the Sylvania R4336. That flashtube required 2000 vdc. I could not find a 4000 volt diode to build a full wave bridge, so I put four 1000 volt diodes in series, and to make sure that the high voltage from the step-up transformer was proportioned evenly across the 4 diodes in series, I put a high value resistor in parallel with each diode so that the voltage would be evenly proportioned and not exceed the diode's 1000 volt reverse breakdown voltage. I don't remember using special resistors, maybe just carbon or wirewound.

                        I'll start with the resistor replacing the disconnected lamp number 4 and if I find that the monitor screen looks too pale at the top compared to the bottom, then I'll try your fudged circuit modification with the marginal lamp still connected.


                        Originally posted by budm View Post
                        Your schematic of BN44-00124E does not match your board, your board has two inverter transformers, the schematic has one inverter transformer and it matches the pictures of the board I posted so the part designators in the SCH do not match your board BN44-00182B.
                        You are correct about my board not being identical to your board, but they have so much in common and there are no schematics for mine anywhere on the web, except for a page which covers the 120vac entry circuit, a photo of which I show below. The BN number in the diagram is the same as the BN number on the bottom of my board, but the IP number on the top of my board is not the same as the IP number in the photo.
                        Attached Files
                        Last edited by montreal; 10-17-2017, 10:18 PM.

                        Comment


                          #32
                          Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                          Putting 1000V diode in series is typical way of making HV diode. I.E HV diode in microwave oven.
                          You can use resistors without any problem if you read post 17. It is the resistor Voltage breakdown that you have to watch out for, so you can use resistor in series/parallel connection to form the required resistance/Wattage and you probably really do not need to run at 6W, you can experiment with resistor value just enough to satisfy the protection circuit. 5~6W is a lot of heat wasted for not really doing any works.
                          Putting resistors in series to form resistor network with applied high Voltage is commonly done, it is the safe practice, you see that a lot in high Voltage sensing circuit.
                          About resistor breakdown Voltage: They make them for reason.
                          http://www.high-voltage-resistors.co...iAAEgIzFfD_BwE
                          https://cdn.badcaps-static.com/pdfs/...d2f003a910.pdf
                          I already have that diagram, you board is revision B, that one page diagram is for revision L SMPS section.
                          Most Samsung power supply board has BN44-XXXXXx.
                          BTW, you probably did not have the backlights level at the maximum level when you made the measurement?
                          Last edited by budm; 10-17-2017, 11:25 PM.
                          Never stop learning
                          Basic LCD TV and Monitor troubleshooting guides.
                          http://www.badcaps.net/forum/showthr...956#post305956

                          Voltage Regulator (LDO) testing:
                          http://www.badcaps.net/forum/showthr...999#post300999

                          Inverter testing using old CFL:
                          http://www.badcaps.net/forum/showthr...er+testing+cfl

                          Tear down pictures : Hit the ">" Show Albums and stories" on the left side
                          http://s807.photobucket.com/user/budm/library/

                          TV Factory reset codes listing:
                          http://www.badcaps.net/forum/showthread.php?t=24809

                          Comment


                            #33
                            Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                            Good news and bad news.

                            First the good news: After looking at a few Youtube videos showing how panels identical or similar to mine are disassembled, I partially disassembled mine. After I turned off all the background noise in my workshop, I could hear some arcing noise when the CCFLs fired up. As I had removed my grounding resistor to pin 1, the protection circuit shutdown the lamps after 2 seconds. I decided to completely disassemble the whole panel which exposed the top and bottom lamp assemblies. Each assembly contained two lamps. I then plugged both assemblies directly into the inverter board along with the pushbutton harness. Then I turned on the inverter and I could see that 3 of 4 lamps were lighting up and for the forth lamp, (number 4), there was a black carbon coating on the glass next to the rubber boot.

                            Now the bad news: taking a q-tip soaked in alcohol, I tried to rub off the carbon and the glass cracked. Additional, lamp number 3 would come on temporarily, but I no longer had light coming from the full length of the lamp, only the first half. So I really have 2 and a half lamps working, and that is not enough.

                            What I need is a new 19" long assembly of a pair of lamps. I'm hoping to find a second hand assembly from a scraped monitor, as the cost of ordering on-line a new pair of assemblies exceeds what second-hand monitors are selling for locally.

                            This has been a tremendous learning experience, I only regret that I won't be building a bypass resistor to fool the chip into believing that the bad lamp is good.

                            Thanks for your encouragement.

                            Comment


                              #34
                              Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                              Originally posted by montreal View Post
                              Now the bad news: taking a q-tip soaked in alcohol, I tried to rub off the carbon and the glass cracked. Additional, lamp number 3 would come on temporarily, but I no longer had light coming from the full length of the lamp, only the first half. So I really have 2 and a half lamps working, and that is not enough.

                              What I need is a new 19" long assembly of a pair of lamps. I'm hoping to find a second hand assembly from a scraped monitor, as the cost of ordering on-line a new pair of assemblies exceeds what second-hand monitors are selling for locally.
                              Well, as cheaper alternatives, you could completely try to disable one half of the inverter and run the monitor only on 2 lamps. I've done that to a Dell 19" here and it works fine. Though I will admit the contrast at the bottom of the monitor (where the two disabled lamps are) does look pretty bad. Of course, to avoid that, you could just get rid of the non-working and half-working lamps, then take one of the good lamps from the other assembly and have it in place of those two bad lamps. That should give you one working lamp on the top and one on the bottom, so the contrast should be equally good/bad throughout the monitor.

                              A different alternative is one of those LED kits on eBay and elsewhere online. I've never tried them, but I've been curious to do that. Member stj here has used them and says they work well.
                              Last edited by momaka; 10-19-2017, 05:33 PM.

                              Comment


                                #35
                                Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                                Originally posted by momaka View Post
                                A different alternative is one of those LED kits on eBay and elsewhere online. I've never tried them, but I've been curious to do that. Member stj here has used them and says they work well.
                                I found that my lamps are very thin, as two are crammed into 6 millimeters of width within the metal "U" shaped reflector. The bulbs are very difficult to separate from the metal case without cutting the low voltage wires. They feel as fragile as a thick piece of spaghetti.

                                I called a few stores which specialize in recycled computers and none keep broken panels around for spare parts.

                                Today I ordered a LED conversion kit from China via Ebay for under $12 US with free shipping (4-5 week delivery). I've looked at some of the videos and this looks very doable. I know that my Samsung 226WB has an output signal for analog dimming which I believe is the only kind of control that the LED driver will accept. I may have to use some fixed resistors to set the analog voltage level that gives me the best brightness. The less brightness I ask for, the less the chances of shortening the lifespan of the LEDs which probably won't like being confined to a narrow channel.

                                http://www.ebay.ca/itm/272479776238

                                https://www.youtube.com/watch?v=7w805Rw9Mpk

                                Do a Google search for the PDF corresponding to chip "LED backlight driver" DF6113.

                                I'll post my final results.

                                Thanks for your comments.
                                Last edited by montreal; 10-19-2017, 06:01 PM.

                                Comment


                                  #36
                                  Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                                  Originally posted by montreal View Post
                                  I found that my lamps are very thin, as two are crammed into 6 millimeters of width within the metal "U" shaped reflector. The bulbs are very difficult to separate from the metal case without cutting the low voltage wires. They feel as fragile as a thick piece of spaghetti.
                                  Thick piece of spaghetti - I like that description!

                                  I know what you mean, though. I had the same ordeal in a 24" LCD. Same 3 mm thick tubes, but with the only difference that mine were almost 22" long - so even more fragile. But with careful prying and cutting the glue/silicone that held them, I was able to get both of them out in once piece. And for the new ones, I put too much insulation around their leads, so don't ask me how I crammed those back in.

                                  Originally posted by montreal View Post
                                  Today I ordered a LED conversion kit from China via Ebay for under $12 US with free shipping (4-5 week delivery). I've looked at some of the videos and this looks very doable.
                                  Yup, it is. Now you will have an LED monitor.

                                  The only thing that I keep asking myself when thinking if I should do those LED conversion is, how long are the LEDs going to last? I know cheap LEDs don't last too long, so I guess I am waiting on other people here to do the repair and hopefully report back after a few years if the LED lamps still work well. After all, it's a pain to replace the lamps in a monitor (though a good learning experience, nevertheless). So if the LED lamps are going to last only 2-3 years, that's almost not worthwhile.
                                  Last edited by momaka; 10-19-2017, 06:16 PM.

                                  Comment


                                    #37
                                    Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                                    Originally posted by momaka View Post
                                    So if the LED lamps are going to last only 2-3 years, that's almost not worthwhile.
                                    I suspect that each LED chip draws 60 milliwatts. And from the advertisement photo, there appears to be about 2 per inch, so for my 19 inches, that's about 2.5 watts. This compares to 2 CCFL lamps in the same space generating about 12 watts. The metal U shaped reflector will pick up some of the heat from the LED strip and it is touching the metal rear shell of the panel, so I might coat the places where metal touches metal with heat sink grease to improve the thermal conductivity.

                                    The advertisement claims that the two LED strips generate 700 nits of light, which is way more than what I should need, so I can back off the intensity and lower the heat generation.

                                    Another issue that has been raised is the degree that the the supply voltage drops the further along the LED strip. One popular solution is to have the positive wire from the power supply at one end of the LED strip and the negative wire re-soldered to the opposite end of the strip, as a way of getting an identical voltage drop across each LED chip. The down side of this idea is that a wire has to be routed back to the entrance. Such a wire exists for CCFL bulbs, but because the current is much smaller, a very thin wire works well as it fits into the recess provided in the metal reflector. For LEDs, this wire would have to be thicker.

                                    But if I end up turning down the brightness a lot, then the voltage loss along the rails of the LED strip may not warrant re-routing the wiring.

                                    Another issue I notice is that when the reflector is re-assembled to the sandwich composed of a thick glass sheet and 3 thin plastic films, the films don't fit under the lips of the metal reflector. I suggest that the U channel reflector have a tool (wooden block) passed down the cavity so that the edges would be bent open slightly in order to accommodate the full thickness of the elements. Once all is in place, the clipping together of the front and back outer metal shells should press the deformed channel back into shape.


                                    I've seen two videos where the panel is reassembled. The method I like the best (as is the case with the video I referenced above) is where the rear metal shell is laid flat on the work table, and all the elements are inserted from above in the correct order until the front metal shell is finally clicked into position.
                                    Last edited by montreal; 10-19-2017, 07:42 PM.

                                    Comment


                                      #38
                                      Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                                      After 20 days, the LED kit arrived from China.

                                      Initially I hooked up the LED driver board (GYD-9E with its DF6113 chip) to a 10 volt battery and applied 3 volts DC to the enable wire. The LED turned on but the current drawn from the battery was only 25 ma. I added another 5 volts DC in series with the 10 volt battery to give me 15 volts and the LEDs went brighter and the current draw was now up at 0.4 amps, so the lights were now drawing 6 watts.

                                      There was no sticky tape on the back of the LED strips, so I used some 2 sided carpet tape to hold the strips in place within the metal U shaped reflectors.

                                      On my Samsung monitor inverter board I found a source of 16 volts DC (officially it is supposed to be 13.5 volts) which I connected to the LED driver wiring harness. The enable wire for the driver board I connected to pin 1 of the ribbon cable that runs between the Samsung scaler board and the inverter board.

                                      Reassembling the panel went more easily than I had expected. Any specs of dust on the diffuser sheet I wiped off with a special cloth.

                                      Unfortunately, there is now a new problem with the LCD top sheet (see photo).

                                      The LCD top sheet has many flexible film ribbon connectors all across the top of the panel for horizontal addressing and four flexible film ribbons down the panel's side for vertical addressing. Each of the four ribbons on the side corresponds to one quarter of the height of the screen.


                                      All the ribbons at the top connect electrically to the circuit board at the top, however the four ribbons along the side are just squeezed under the outer metal frame.

                                      While there is no hard electrical connection between the four side ribbons and the chassis ground, I assume that there is a capacitive connection that operates due to the high switching frequency of the panel.

                                      As you can see from my photo, only the top quarter zone is working.

                                      I partially disassembled the panel and played with the four side ribbons one at a time and I could see that I was triggering different quarter zones to go on and off, so I conclude that my problem has something to due with the way these side ribbons are physically positioned during panel assembly.

                                      So while my experiment of converting from CCFL to LED was a success for under $15, the failure of my reassembled LCD panel to work properly brings me back to square one.

                                      I recommend this upgrade kit for monitors which allow the old CCFL tube assembles to be pulled out of their cavity without having to pull apart the different layers of the panel sandwich.

                                      When a person has to take apart the panel sandwich, one exposes to problems of dust and damage to the flexible ribbon cables, as was my case.
                                      Attached Files
                                      Last edited by montreal; 11-09-2017, 11:36 AM.

                                      Comment


                                        #39
                                        Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                                        Good day folks, sorry for bumping such an old thread, but when I read through some of the posts, I couldn't help but notice the OZ964 being mentioned and it just so happens that I too am trying to hack an inverter to run on a monitor it's not designed for, so hopefully someone will pitch in...I wrote an individual post about this but apparently it's a bit too ambitious and people didn't have the patience to join this little project. It's about an HP ZR24w monitor which has no inverter but it's otherwise fully functional, so I tried adapting an inverter to get it going.

                                        I first tried it with an inverter which originally drove more lamps than this monitor had, so 4 outputs were left floating...despite my best efforts I COULD NOT get that one to work. I then tried it with another inverter out of a Samsung TV which happened to match the number of outputs perfectly (12). There's 6 U-shaped lamps in this monitor, so multiply that by 2 ends each and you get 12 "outputs". I messed around with this setup as well, but the closest I got was getting it to go on-off-on-off...again, messed around with that OZ IC but on top of not making any progress, I also broke it in the process...the thing just blew a fuse somehow and I couldn't get it to run again, so I abandoned it for the moment...I am however going to persevere with this and I'll hopefully fix that inverter so I can at least get it back to doing what it was doing initially, since now it's totally dead. Maybe some help will come from the chaps who said something about an inverter with an OZ and additional comparator ICs who are obviously more vexed than me. From what I read on some local forums, inverters like these with OZs and additional ICs are hard to trick. I'm not sure if it doesn't run as a direct replacement because the lamps in the 24 incher are smaller and draw too much current, or not enough current...the TV it came out of was perfect as far as backlight was concerned, so I know this second inverter was functional before I trashed it Cheers guys.
                                        Wattevah...

                                        Comment


                                          #40
                                          Re: Using a resistor to fake a Inverter into thinking a CCFL is good

                                          I also successfully used your resistor trick on an Asus monitor today. One of the lamps was flickering, so I disabled it with two 5.1k resistors in parallel. I started off with a higher value but they got very hot because as I figured out later on, CCFL supplies are constant current supplies, thus they require more voltage for the same current with a higher value resistor (load), so I dropped it lower and lower until I felt the resistors weren't overheating anymore...I probably could've gone even lower, but it seemed quite happy like this so I left it, plus I didn't have any smaller resistors anyway Cheers.
                                          Wattevah...

                                          Comment

                                          Working...
                                          X