Page 10 of 16 FirstFirst 12345678910111213141516 LastLast
Results 181 to 200 of 302

Thread: LG TV wipeout

  1. #181
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    Looking back through some of the posts I realized a question I didn't answer : IC101 RFmin (and stby) pins to ground is 8Kohms.
    I went methodically lifting pins starting at 1 then checking the voltage on pins4/5.
    Lifting Pins 1,2 and 3 didn't raise rfmin voltage, Lifting Pins 6,7 and 8 dropped rfmin to 0v. Hmmmmm
    Last edited by loopyloo; 12-08-17 at 08:14 AM.



  • #182
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    Looked at ring lights for the camera (how to make one) and concluded it was way cheaper to just order one on Ebay as they come with a variety of adaptors and hot shoe connection etc etc. Sorta takes all the challenge out of angling the desklamp and camera just right to minimise reflections tho

    So Vcc didn't improve or change when those pins were lifted?
    And any variations if the chip is supplied with a good 12V Vcc from an external supply (needs to be around 12V to guarantee start-up)
    Even just testing the chip with a variable supply to see if there's a threshold voltage below which RFmin's 2V reference is off would be valuable info.
    According to the datasheet, the typical Icc current should never be above 10mA in operation so that's a major red flag (the drain on Vcc is killing the supply voltage)

    I'd be powering the chip up from an external variable supply with all the outputs (LVG, HVG, OUT) disconnected, LINE voltage spoofed as I posted earlier, all with no mains connected for safety. Need to make that chip think it's in circuit with all valid inputs to see if it will run in some normal sort of fashion (CRO helps to monitor HVG, LVG outputs)

    Need to try everything under controlled conditions. It wouldn't make sense that 2 new chips could be faulty but so far that's the way it looks :|

  • The Following User Says Thank You to Skepticist For This Useful Post:

    loopyloo (11-08-17)

  • #183
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    Quote Originally Posted by Skepticist View Post
    Looked at ring lights for the camera (how to make one) and concluded it was way cheaper to just order one on Ebay as they come with a variety of adaptors and hot shoe connection etc etc. Sorta takes all the challenge out of angling the desklamp and camera just right to minimise reflections tho

    So Vcc didn't improve or change when those pins were lifted?
    And any variations if the chip is supplied with a good 12V Vcc from an external supply (needs to be around 12V to guarantee start-up)
    Even just testing the chip with a variable supply to see if there's a threshold voltage below which RFmin's 2V reference is off would be valuable info.
    According to the datasheet, the typical Icc current should never be above 10mA in operation so that's a major red flag (the drain on Vcc is killing the supply voltage)

    I'd be powering the chip up from an external variable supply with all the outputs (LVG, HVG, OUT) disconnected, LINE voltage spoofed as I posted earlier, all with no mains connected for safety. Need to make that chip think it's in circuit with all valid inputs to see if it will run in some normal sort of fashion (CRO helps to monitor HVG, LVG outputs)

    Need to try everything under controlled conditions. It wouldn't make sense that 2 new chips could be faulty but so far that's the way it looks :|
    Lifting Pins 1,2 and 3 didn't raise rfmin voltage, Lifting Pins 6,7 and 8 dropped rfmin to 0v. I lifted these pins one at a time .
    OK so done exactly as you said : Bingo ! spot on 2 volts on RFmin.
    So the next step would be to re-apply the mains but leave pins 11, 14 and 15 disconnected and check RFmin, right ?
    What if D103 and/or D121 are dud ? I'll have to make another iron tip so I can check them properly.
    or Do you think C119 could be shot ?

    PS Close inspection with my loupe reveals D121 seems to have lost its markings. Junction test shows D103 and D121 reads 100mV both directions (in circuit)
    Last edited by loopyloo; 12-08-17 at 01:58 PM.

  • #184
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    OMG sounds like progress!
    Stay with the external Vcc supply (solid 12V) until the culprit is found
    Next would be to check the supply current with different combinations of pins 11,14,15 connected (one at a time first) to see which of them causes a leap in current - use that sewing needle to briefly connect a lifted pin to the track just long enough to read the current - have you checked what Icc is with all 3 disconnected and a good RFmin voltage?

    You won't get a real reading on those diodes in circuit due to the total of 69 ohms in parallel with them

    Make sure R120 & R123 (10k) are actually still 10k because the only other thing that could overload those gate drive outputs is a failed mosfet
    Last edited by Skepticist; 12-08-17 at 02:50 PM.

  • #185
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    failed mosfet ??? better not be, they're brand new.
    I'll check them anyway.

    These soldering iron tips are getting fancier.
    This larger smd tip is mounted on an old tip I shaped to do the smaller smd. I'll use it to remove those diodes.


    Last edited by loopyloo; 12-08-17 at 06:28 PM.

  • #186
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    The diodes are only for controlling overswings in the gate drives when the chip operates and I doubt you'd notice anything if they were shorted without a CRO but the 10k resistor (particularly R123) could cause the problems if it's shorted to ground as there'd be only a 47 ohm resistor between LVG and GND in that case, the unthinkable other possibility is that Q102 is broken down between gate & source. You still have 2 spare mosfets in the PFC circuit though.
    Last edited by Skepticist; 12-08-17 at 06:55 PM. Reason: rough typin

  • The Following User Says Thank You to Skepticist For This Useful Post:

    loopyloo (12-08-17)

  • #187
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    I do have 1 spare fet cos I had to buy 5. They test funny with my in circuit fet tester so I'd have to remove them to see.
    I tried shorting the pins to pads but none of them individually lowered the RFmin. Must need either 2 of them connected or all 3 to cause the fault.
    Bit worried about too much soldering and desoldering those pins cos they're so fragile and so are the pads. I don't have enough hands so pick one and I'll solder it down then try the test again.
    R123 is 10k in circuit.

    This prob didn't help either. I discovered this open lead while doing those last measurements, and I know I used this test lead a couple of times before when taking readings.




    The old, fold the wire under and crimp the lot (plastic and all) trick.
    For shit sake China, train your damn workers.

    PS By the time this thing's fixed (if ever) , I might turn it on to find the screen is broken sometimes you just can't see a crack.
    Last edited by loopyloo; 12-08-17 at 08:19 PM.

  • #188
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    What I was suggesting was monitoring the Vcc input current (Icc) while individually connecting those outputs to try and narrow down exactly where to look for the fault. As long as there's no conductivity between gate and source I'd presume the mosfet to be OK. The fault current on Vcc appears to be flowing on the output side of the chip so shouldn't be too hard to find it there (down to 3 probables now), so close
    Was R120 the same (10k) ?

    Had a few test leads similar to that myself, some had such flimsy conductors the Cu fatigued and broke leaving the insulation intact (nice intermittent fault there). Jacks the price up if they were to use solder on them.

  • The Following User Says Thank You to Skepticist For This Useful Post:

    loopyloo (13-08-17)

  • #189
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    Quote Originally Posted by Skepticist View Post
    What I was suggesting was monitoring the Vcc input current (Icc) while individually connecting those outputs to try and narrow down exactly where to look for the fault. As long as there's no conductivity between gate and source I'd presume the mosfet to be OK. The fault current on Vcc appears to be flowing on the output side of the chip so shouldn't be too hard to find it there (down to 3 probables now), so close
    Was R120 the same (10k) ?

    Had a few test leads similar to that myself, some had such flimsy conductors the Cu fatigued and broke leaving the insulation intact (nice intermittent fault there). Jacks the price up if they were to use solder on them.
    Yes I did check Icc but I could check better if I use my other meter. I went off the meter built in to my power supply which only shows switchable to amps or milliamps. Not once during testing did it shift off zero so the current would be less than 500uA since it didn't even average out to display a 0001. Do you think I should try it again ? I mean, are we really looking at only microamps to make a difference ?

    Oh and yes R120 is also 10k in circuit.

    Getting close woo hoo

  • #190
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    Hmmm it's getting elusive again
    Need to re-establish the conditions where there's enough overload on Vcc to kill RFmin's 2V reference then track down what pin(s) draw that excess current - latest tests seem to indicate it's LVG, HVG, OUT or some combination of those.
    It's enough of a fault to draw the normal Vcc supply down below the start-up threshold so should be easy to find by measuring voltage drops across a few resistors and the voltage drop across a low value resistor in series with the Vcc supply (from an external power supply) will allow calculation of that current under fault condition (compared to the negligible current under normal condition when RFmin 2V is OK)
    EG assuming D103, 121 are ok, you should see higher than expected voltage across R118 or R121 if it's going via a mosfet G-S or R120, R123.
    If a diode is short cct it will still be a high enough voltage drop to show something is abnormal and where (or you could remove R119, R122 to be certain)

    There couldn't be any solder bridges around the mosfets could there?

    I hope this makes sense - just need more data to pin it all down a little better and we seem to be down to a couple of pins and about a dozen or less suspects now.

  • The Following User Says Thank You to Skepticist For This Useful Post:

    loopyloo (13-08-17)

  • #191
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    No bridges around the fets.
    Up to here so far : Table of connecting pins to their pads and observing RFmin....x = connected

    Tomorrow I'll remove the 2 fets.


    Last edited by loopyloo; 13-08-17 at 01:12 AM.

  • #192
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    So RFmin is OK at 2V as long at least 1 of the 3 output pins is disconnected, only dead when all 3 are connected?
    Is Icc highest when RFmin is 0V?

    Something I noticed is your replacement IC is a L6599D while the schematic shows the improved version of that chip L6599AD. I skimmed over the datasheets for the 2 versions but can't see any difference in function or ratings and they are interchangeable.
    The only real difference appears to be around Css, with the older chip using a bipolar transistor to discharge the external cap and the improved version using a fet. The DELAY threshold voltages related to Css switching are different old 2V new 3.5V (could be a typo in the document too)

    It could be the FETs so will be interesting to see results with them removed. But how could they have failed is the question - maybe when powering the board up with the old fried IC still in there?
    Last edited by Skepticist; 13-08-17 at 12:33 PM.

  • #193
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    I'm also wondering if I could have bought the wrong FETs.
    These are the one I got : Search ..... ..... for 1241106.....the schematic calls for STF13MN60N and stating them as a critical component. I couldn't find enough comparison info when I chose them so was hoping the specs would match ok

  • #194
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    I don't see any problem with the FET specs at all. Key features like Vds, Id, Rds, Vgs and Vgsth are all the same so it's the same device maybe packaging difference only

  • The Following User Says Thank You to Skepticist For This Useful Post:

    loopyloo (13-08-17)

  • #195
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    Just great ! .... what next.
    Go to check them fets and find out my fet tester is playing up now.
    Looks like I'll be fixing that before anything else.

  • #196
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    Do the previous external supply test with the FETs disconnected to see if the RFmin voltage etc problems are gone with those output pins all connected to their tracks. If the problem still exists the FETs might be innocent.

  • #197
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    Quote Originally Posted by Skepticist View Post
    Do the previous external supply test with the FETs disconnected to see if the RFmin voltage etc problems are gone with those output pins all connected to their tracks. If the problem still exists the FETs might be innocent.
    Yeah that's it. With the fets out and everything else connected 2volts is still there.
    But here's the thing, The fets seem to be ok , they are not short and they read into the megohms. D-S 10meg, G-D and G-S 40meg and climbing (must be my meter).
    There has to be just one stupid little thing getting in the way here.
    Last edited by loopyloo; 13-08-17 at 04:05 PM.

  • #198
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default

    What's the resistance reading between gate-source & drain-source pads on the PCB for each Fet position now? test both ways for drain-source to see if the internal diode is ok
    and R124 (1M) resistance
    and that's just about everything external to the chip covered

    FETs are awkward for testing at the best of times, those ones have threshold Vgs of 3-4V but need more like 10V to be fully 'on'

    I'm getting so familiar with this L6599 I'm getting myself some to have a play with (won't have them for a while yet tho - snail mail)
    Last edited by Skepticist; 13-08-17 at 04:08 PM.

  • #199
    Senior Member
    loopyloo's Avatar
    Join Date
    Feb 2008
    Location
    Mid North Coast NSW . Australia
    Age
    67
    Posts
    2,207
    Thanks
    1,431
    Thanked 469 Times in 313 Posts
    Rep Power
    415
    Reputation
    7297

    Default

    Quote Originally Posted by Skepticist View Post
    What's the resistance reading between gate-source & drain-source pads on the PCB for each Fet position now? test both ways for drain-source to see if the internal diode is ok
    and R124 (1M) resistance
    and that's just about everything external to the chip covered

    FETs are awkward for testing at the best of times, those ones have threshold Vgs of 3-4V but need more like 10V to be fully 'on'

    I'm getting so familiar with this L6599 I'm getting myself some to have a play with (won't have them for a while yet tho - snail mail)
    R124 is 1M.
    For the pads : Q101 G-S 7k ohms, D-S 5Meg ... Q102 : G-S 8k ohms, D-S 1Meg
    Last edited by loopyloo; 13-08-17 at 06:58 PM.

  • #200
    Premium Member
    Skepticist's Avatar
    Join Date
    Apr 2009
    Posts
    1,139
    Thanks
    714
    Thanked 670 Times in 525 Posts
    Rep Power
    474
    Reputation
    12780

    Default



    Just knocked out this little FET tester in Ubuntu Kolourpaint
    Won't test for voltage breakdown but should give a fair go-nogo indication with minimum parts.

  • Page 10 of 16 FirstFirst 12345678910111213141516 LastLast

    Bookmarks

    Posting Permissions

    • You may not post new threads
    • You may not post replies
    • You may not post attachments
    • You may not edit your posts
    •