Repair of Samsung 55\" TV - backlight problem...

Well, I had a major setback trying to diagnose the repair. Last night I
set it on the dining room table, face down, and this morning I did the
flashlight and audio test again before even taking the back cover off,
just to confirm the video and audio were still working. But after about 30
minutes, the TV started working again. All the backlights came on, and the
picture is perfect. This is the behavior the owner reported - it might
get backlight at any time, or it might not.

I measured 267V between BD9101 and J858, so that\'s the benchmark in case it
dies again.

So I don\'t know how you diagnose anything when it\'s working properly. We
tried tapping on every part and connector on the power board, but not even
a flicker in the lights.

Both sides of the power board look perfect - not a hint of a bad solder, no
domed capacitors, no brown marks. The connector to the LEDs (CLN802) has 7
black wires and one blue wire, and a legend that doesn\'t seem to match the
connector.

In any case, I will let it cool down, and see how it powers up. I guess I
could remove and test all the electrolytics, but there are a lot of them.
Same for the diodes. But it seems it could be anything. Major bummer that
it\'s working. :)
 
On Saturday, July 23, 2022 at 9:27:02 AM UTC-4, Peabody wrote:
ohg...@gmail.com says...
In *this* model, they\'re all in series and it\'s a one
channel LED system - one LED or if one interconnect
opens, no light.
On the ShopJimmy page for this model, it shows the LED
package as seven strips of eight LEDs, plus seven strips of
five LEDs. That\'s 91 LEDs. If they\'re all in series, then
3V LEDs would need 273V. So perhaps the strips are wired in
some parallel configuration?

I have an HU board here in front of me that I grabbed the standby IC out of for another job, and I have the LED voltage marked as 325/375V on the BD9101 ferrite jumper (with respect to cold ground. Use the mounting screws for ground). I have seen a bunch of TVs that use a series/parallel arrangement, but I can\'t recall Samsung ever doing that.



I\'m starting on the diagnostics this morning with a neighbor
who I discovered used to work for TI back in the day.

One of the ShopJimmy videos say flatly that if you pass the
flashlight test, with audio, then a dark screen is either
the power board or the LEDs. I guess that\'s not absolutely
always true, but seems logical to me. I\'m hoping for the
power supply to be the problem. I\'m not sure I would be
willing to tackle the LEDs. I\'ve watched those videos, and
it\'s a major undertaking, with lots of opportunities to break
things.

I\'m not a betting man, but put a gun to my head and I\'m going with bad LEDs, bad connections on the interconnects (or both), but inside the display for sure.
 
On Saturday, July 23, 2022 at 12:33:56 PM UTC-4, Peabody wrote:
Well, I had a major setback trying to diagnose the repair. Last night I
set it on the dining room table, face down, and this morning I did the
flashlight and audio test again before even taking the back cover off,
just to confirm the video and audio were still working. But after about 30
minutes, the TV started working again. All the backlights came on, and the
picture is perfect. This is the behavior the owner reported - it might
get backlight at any time, or it might not.

I measured 267V between BD9101 and J858, so that\'s the benchmark in case it
dies again.

Keep a meter on it all the time. If it has a min/max recording feature, connect a DVD player and leave it on the title screen so the TV doesn\'t dynamically adjust the back light to scene content or time out and shut off. Check the max voltage every once in a while. If your baseline drifts more than a tenth or so with no change of picture content, it indicates a failing LED (they do intermitt both towards open and leaky) or a bad interconnect between the A/B sections of the LED strips or where the LED strips plug into the main feed strip.


> So I don\'t know how you diagnose anything when it\'s working properly.

Sometimes we take the display out and they come to life by themselves. It happens. One thing to do is put a meter across every connection and see if voltage appears across any connection (should be zero volts). Any voltage across an interconnect means a bad connection. If there is any voltage drifting even with the LEDs appearing to be running properly, we know we can find it.

Sometimes we have to hard wire all the interconnects - if just one is bad, another will follow eventually. It\'s not intuitive but poor connections *don\'t* show much if any change in brightness even when tapped on as the LED drive controller is extremely fast and compensates immediately for any change in load impedance. Only when the connection is bad enough to cause the controller IC to detect either an overvoltage or undercurrent condition does the LED array shut down.
 
In article <20220723-163353.893.0@Peabody.ssl-us.astraweb.com>,
waybackNO584SPAM44@yahoo.com says...
Well, I had a major setback trying to diagnose the repair. Last night I
set it on the dining room table, face down, and this morning I did the
flashlight and audio test again before even taking the back cover off,
just to confirm the video and audio were still working. But after about 30
minutes, the TV started working again. All the backlights came on, and the
picture is perfect. This is the behavior the owner reported - it might
get backlight at any time, or it might not.

I measured 267V between BD9101 and J858, so that\'s the benchmark in case it
dies again.

So I don\'t know how you diagnose anything when it\'s working properly. We
tried tapping on every part and connector on the power board, but not even
a flicker in the lights.

Both sides of the power board look perfect - not a hint of a bad solder, no
domed capacitors, no brown marks. The connector to the LEDs (CLN802) has 7
black wires and one blue wire, and a legend that doesn\'t seem to match the
connector.

In any case, I will let it cool down, and see how it powers up. I guess I
could remove and test all the electrolytics, but there are a lot of them.
Same for the diodes. But it seems it could be anything. Major bummer that
it\'s working. :)

If it does not power up, you may try heating the capacitors one at a
time. I have seen that work on equipment that comes on after the set is
powered up for a while if the capacitor is bad.
 
On Sat, 23 Jul 2022 11:33:53 -0500, Peabody
<waybackNO584SPAM44@yahoo.com> wrote:

So I don\'t know how you diagnose anything when it\'s working properly. We
tried tapping on every part and connector on the power board, but not even
a flicker in the lights.

Intermittents are a PITA. I\'ll spare you my guess of the day. The
way I used to troubleshoot intermittents is heating with a hot air gun
and cooling with freeze spray in a can. Whatever connection is
intermittent, is going to move when it gets hot or cold, either making
the connections, or as in this case, breaking the connection. Don\'t
use to much cooling or heating. Dropping the temperature below the
dew point will condense water all over the board. Heating the PCB too
much will melt plastic parts. If you do condense water on the PCB,
remove power from the TV, apply a little heat, and wait until the
water evaporates before trying again.

You missed a golden opportunity when you discovered that:
I measured 267V between BD9101 and J858, so that\'s the
benchmark in case it dies again.

The LED\'s are connected in series. That means you should see a
voltage equal to 267V divided by the number of strips, across each
strip.
<https://www.shopjimmy.com/samsung-bn96-34251a-bn96-34252a-led-backlight-strips-14/#mz-expanded-view-1444464375404>
Assuming they\'re all in series, that would be:
267 / 14 = 19V
Walk your DVM across each strip and see which one has the full 267V
across it. That\'s the defective (open) LED strip.

--
Jeff Liebermann jeffl@cruzio.com
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558
 
On Sat, 23 Jul 2022 14:53:50 -0700, Jeff Liebermann <jeffl@cruzio.com>
wrote:

On Sat, 23 Jul 2022 11:33:53 -0500, Peabody
waybackNO584SPAM44@yahoo.com> wrote:

So I don\'t know how you diagnose anything when it\'s working properly. We
tried tapping on every part and connector on the power board, but not even
a flicker in the lights.

Intermittents are a PITA. I\'ll spare you my guess of the day. The
way I used to troubleshoot intermittents is heating with a hot air gun
and cooling with freeze spray in a can. Whatever connection is
intermittent, is going to move when it gets hot or cold, either making
the connections, or as in this case, breaking the connection. Don\'t
use to much cooling or heating. Dropping the temperature below the
dew point will condense water all over the board. Heating the PCB too
much will melt plastic parts. If you do condense water on the PCB,
remove power from the TV, apply a little heat, and wait until the
water evaporates before trying again.

You missed a golden opportunity when you discovered that:
I measured 267V between BD9101 and J858, so that\'s the
benchmark in case it dies again.

The LED\'s are connected in series. That means you should see a
voltage equal to 267V divided by the number of strips, across each
strip.
https://www.shopjimmy.com/samsung-bn96-34251a-bn96-34252a-led-backlight-strips-14/#mz-expanded-view-1444464375404
Assuming they\'re all in series, that would be:
267 / 14 = 19V
Walk your DVM across each strip and see which one has the full 267V
across it. That\'s the defective (open) LED strip.

Never mind. It looks like you will need to disassemble the LCD panel
in order to obtain access to the LED connections. Once everything is
disassembled, you might as well replace all the LED strips instead of
trying to find the one with an intermittent connection.

I just watched the LED replacement video at:
<https://youtu.be/y5d_JkOEBSg>
I can\'t deduce which LED strip is bad, but my guess(tm) is it\'s one of
the two pin connectors on the LED strips. See video starting at:
<https://youtu.be/y5d_JkOEBSg?t=472>
My free advice is to put the TV back together and hope it continues to
function normally. This doesn\'t appear to be a paying repair, so you
might be able to convince the owner to do it this way which minimizes
the risk of breaking the panel. After all, the previous repair lasted
3 years. If it fails in the future, order the parts and follow the
video. I couldn\'t (and wouldn\'t) do that to a paying customer who
expects a repair and a warranty on the labor.

BTW, you were right about your suspicion that the problem wasn\'t the
main board. The repair shop probably had the intermittent back
lighting magically fix itself, and just added the board replacement to
the invoice to make it look like they actually did something useful.

--
Jeff Liebermann jeffl@cruzio.com
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558
 
Jeff Liebermann says...

The LED\'s are connected in series. That means you
should see a voltage equal to 267V divided by the number
of strips, across each strip. <https://www.shopjimmy.com
/samsung-bn96-34251a-bn96-34252a-led-backlight-strips-
14/#mz-expanded-view-1444464375404> Assuming they\'re all
in series, that would be: 267 / 14 = 19V Walk your DVM
across each strip and see which one has the full 267V
across it. That\'s the defective (open) LED strip.

The LED\'s are connected in series. That means you
should see a voltage equal to 267V divided by the number
of strips, across each strip.

<https://www.shopjimmy.com/samsung-bn96-34251a-bn96-34252a-led-backlight-
strips-14/#mz-expanded-view-1444464375404>

Assuming they\'re all in series, that would be:
267 / 14 = 19V
Walk your DVM across each strip and see which one has
the full 267V across it. That\'s the defective (open)
LED strip.

As you said later, I can\'t get to them without taking
everything apart. But I\'m confused about what the LED
circuit is:

First, the ShopJimmy LED package has 7 strips of 8 LEDs,
plus 7 strips of 5 LEDs. Is this an all-Samsung LED package
so you might not need to use all of them, or are there 91
LEDs in my TV? If not 91, how many?

Second, there appear to be three chains of LEDs in the TV,
connected in series, but the interconnects between the three
chains are on the power supply board via the connector:


Voltage vs Chassis Ground

Connector Pin Black Mixed Light

1+ (also BD9101) 240 283 284

1-, 2+ 153 168 182

2-, 3+ 96 90 113

3- (also J858) 42 15 45


I got these readings by connecting my computer via HDMI.
\"Black\" means a black screen. Light is an editor that\'s
almost all whaite space, and Mixed is a video of Glenn
Gould playing the piano. When in Black, the LEDs are dark -
no light visible coming through the chassis. So the main
board clearly has some way of modifying the backlight
brightness.

But while Black looks like the LEDs are off, there obviously
must be some current getting through.

So how many LEDs are in each of the three circuits? And do
these numbers tell me anything about what might be wrong?
The screen looks completely normal on the front, so I assume
the readings above are normal.

I\'ve decided that I\'m not going to replace the LED strips
unless I can prove that\'s the problem. I don\'t want to go
through all that, and spend the money, only to end up with
it still not working. I think proof would be high voltage
applied to the LED string, but no light. And I can\'t get
that unless it goes back to not working. Short of that, I
can check the electrolytics on the power board, and replace
any that look bad. And maybe try the coolant thing. However,
the owner confirmed that once the backlight comes on, it never
goes off on its own. It continues to work until they turn off
the TV. So I would have to apply coolant while it\'s off, and
hope it won\'t turn on.

I have not checked to see if there\'s any PWM going on in the
LED supply. But I\'m not sure it matters.
 
On Thursday, July 21, 2022 at 12:01:15 AM UTC-4, Peabody wrote:
Friends of mine have a Samsung 55\" 4K TV, model UN55HU6840 (about 2014).
They say they can hear audio, but the screen is black. I\'d like to take a
shot at fixing it for them since they don\'t really have the money to buy
a new one. The encouraging thing is that until recently this has been an
intermittent problem that\'s just gradually gotten worse, and now is
permanent. But I\'m encouraged to think it\'s not burned out LEDs, else it
would never have been intermittent.

I just told you everything I know about repairing TVs, but am generally
experienced in electronics as a hobbyist. Based on some Youtube videos
it seems I should take off the back cover and check all the connections.
Then I would go to the power board and check the voltages on the lines
going to the backlight LEDs, and if low, start checking diodes and
electrolytics, and of course look for bad solder joints.

Does this sound like a reasonable approach? Is there a typical cause of
this symptom?

I\'ve had no luck finding a schematic. Does anyone know where I would
find that?

Well, any guidance would be appreciated.
____________________
And once it is repaired, do not operate the set with the backlight set any higher than 12(I believe Samsung backlight
scale is 1-20). Use a higher Brightness(black level) setting to see detail in darker parts of the image
 
Ok I think I\'ve figured it out. One strip of LEDs across
the panel consists of one 8-LED section plus one 5-LED
section. So 7 strips times 13 LEDs is 91 LEDs. The
voltages show there are one three-strip (39-LED) and two
two-strip (26-LED) lines that are individually wired back to
the power board, but are simply jumpered together there.
This confirms what you guys said - in effect all the LEDs
are in series, so any problem with one LED, or one
connection, and they all go out.

I think this is the most relevant video on LED replacements.
It\'s not exactly the same model, but very similar:

https://www.youtube.com/watch?v=oI52B4ba2Dg

It doesn\'t look too bad, except I don\'t have the suction
cups. But as of this morning, ShopJimmy is out of stock on
these, with no indication of a restock date. I can get them
on Amazon for $100, or I can get knockoffs on Ebay for as
little as $40. I don\'t know if the Ebay Chinese ones are
any good. And I don\'t know if the Amazons are different
from the Chinese ones on Ebay.

https://www.ebay.com/itm/125141436851

I did try disconnecting the main board, and the LEDs did
indeed light up fully, with about 281V across the LEDs. But
unfortunately, it\'s still working, so there\'s still no way
to tell where the problem is. It could be the main board
(unlikely, but still possible), the power board, or the
LEDs. And I just can\'t narrow it down unless it goes into
failure mode again.

So I\'m going to continue testing it for a couple days, but
if it doesn\'t fail again, I\'ll give it back to the owner.
After all, with all the messing around I did, I may have
fixed whatever was wrong, and it may go another five years.

I really appreciate everyone\'s help with this. I just wish
we had reached more definitive result.
 
On Monday, July 25, 2022 at 12:11:46 PM UTC-4, Peabody wrote:
Ok I think I\'ve figured it out. One strip of LEDs across
the panel consists of one 8-LED section plus one 5-LED
section. So 7 strips times 13 LEDs is 91 LEDs. The
voltages show there are one three-strip (39-LED) and two
two-strip (26-LED) lines that are individually wired back to
the power board, but are simply jumpered together there.
This confirms what you guys said - in effect all the LEDs
are in series, so any problem with one LED, or one
connection, and they all go out.

I think this is the most relevant video on LED replacements.
It\'s not exactly the same model, but very similar:

https://www.youtube.com/watch?v=oI52B4ba2Dg

It doesn\'t look too bad, except I don\'t have the suction
cups. But as of this morning, ShopJimmy is out of stock on
these, with no indication of a restock date. I can get them
on Amazon for $100, or I can get knockoffs on Ebay for as
little as $40. I don\'t know if the Ebay Chinese ones are
any good. And I don\'t know if the Amazons are different
from the Chinese ones on Ebay.

https://www.ebay.com/itm/125141436851

I did try disconnecting the main board, and the LEDs did
indeed light up fully, with about 281V across the LEDs. But
unfortunately, it\'s still working, so there\'s still no way
to tell where the problem is. It could be the main board
(unlikely, but still possible), the power board, or the
LEDs. And I just can\'t narrow it down unless it goes into
failure mode again.

So I\'m going to continue testing it for a couple days, but
if it doesn\'t fail again, I\'ll give it back to the owner.
After all, with all the messing around I did, I may have
fixed whatever was wrong, and it may go another five years.

I really appreciate everyone\'s help with this. I just wish
we had reached more definitive result.
 
On Monday, July 25, 2022 at 12:11:46 PM UTC-4, Peabody wrote:

I did try disconnecting the main board, and the LEDs did
indeed light up fully, with about 281V across the LEDs. But
unfortunately, it\'s still working,

Don\'t run it long with the main disconnected when testing - it will run the LEDs at higher current than full backlight adjustment will even drive them.. Good for a quick test but don\'t leave it running that way.

Put the TV on a static input (DVD in pause) and watch that voltage across the LED array. If it drifts more than a couple of tenths either way, either an LED is starting to break down or there\'s a resistive contact in the interconnects. Sometimes one or both will settle down for a while before acting up again.

There is another possibility however. Most Samsung LEDs are built with a zener inside each die. Ordinarily, the zener does nothing but if the LED fails open, the entire voltage of the array will appear across the zener causing it to conduct and (hopefully) short hard effectively jumping out the bad LED. If it does, the problem will fix itself until another wayward LED crops up or if an interconnect gets resistive.
 
ohg...@gmail.com says...

There is another possibility however. Most Samsung LEDs
are built with a zener inside each die. Ordinarily, the
zener does nothing but if the LED fails open, the entire
voltage of the array will appear across the zener
causing it to conduct and (hopefully) short hard
effectively jumping out the bad LED. If it does, the
problem will fix itself until another wayward LED crops
up or if an interconnect gets resistive.

If I connect my computer to the TV via HDMI, and display a
competely white (all FFs) BMP image full screen, would I be
able to detect an LED that\'s not lighting up? Would there
be a dark area on the screen, or is the failure of one LED
not enough to cause a detectable darker area on the screen?
When I try this, I don\'t see anything unusual. The display
is a little darker in the corners, but I think that\'s normal.
 
On Monday, July 25, 2022 at 1:41:31 PM UTC-4, Peabody wrote:
ohg...@gmail.com says...
There is another possibility however. Most Samsung LEDs
are built with a zener inside each die. Ordinarily, the
zener does nothing but if the LED fails open, the entire
voltage of the array will appear across the zener
causing it to conduct and (hopefully) short hard
effectively jumping out the bad LED. If it does, the
problem will fix itself until another wayward LED crops
up or if an interconnect gets resistive.
If I connect my computer to the TV via HDMI, and display a
competely white (all FFs) BMP image full screen, would I be
able to detect an LED that\'s not lighting up? Would there
be a dark area on the screen, or is the failure of one LED
not enough to cause a detectable darker area on the screen?
When I try this, I don\'t see anything unusual. The display
is a little darker in the corners, but I think that\'s normal.
 
On Monday, July 25, 2022 at 1:41:31 PM UTC-4, Peabody wrote:
ohg...@gmail.com says...
There is another possibility however. Most Samsung LEDs
are built with a zener inside each die. Ordinarily, the
zener does nothing but if the LED fails open, the entire
voltage of the array will appear across the zener
causing it to conduct and (hopefully) short hard
effectively jumping out the bad LED. If it does, the
problem will fix itself until another wayward LED crops
up or if an interconnect gets resistive.
If I connect my computer to the TV via HDMI, and display a
competely white (all FFs) BMP image full screen, would I be
able to detect an LED that\'s not lighting up?

Probably, but Samsung does a good job with their lens design on their back lit models. Think about it - there is a fair amount of dead space between LEDs as they\'re laid out and generally you don\'t see the individual LED illumination even on a plain white display. There will be *some* unevenness if an LED goes out, but it\'s generally not dramatic, but your all white bit map pattern might show it. Try different brightness settings when you\'re looking at it. I\'ve also seen some older Samsungs with half the LEDs shorted and the customers didn\'t complain at all about the picture until one opened and shut the array down.
 
On 2022/07/25 10:50 a.m., ohg...@gmail.com wrote:
On Monday, July 25, 2022 at 1:41:31 PM UTC-4, Peabody wrote:
ohg...@gmail.com says...
There is another possibility however. Most Samsung LEDs
are built with a zener inside each die. Ordinarily, the
zener does nothing but if the LED fails open, the entire
voltage of the array will appear across the zener
causing it to conduct and (hopefully) short hard
effectively jumping out the bad LED. If it does, the
problem will fix itself until another wayward LED crops
up or if an interconnect gets resistive.
If I connect my computer to the TV via HDMI, and display a
competely white (all FFs) BMP image full screen, would I be
able to detect an LED that\'s not lighting up?

Probably, but Samsung does a good job with their lens design on their back lit models. Think about it - there is a fair amount of dead space between LEDs as they\'re laid out and generally you don\'t see the individual LED illumination even on a plain white display. There will be *some* unevenness if an LED goes out, but it\'s generally not dramatic, but your all white bit map pattern might show it. Try different brightness settings when you\'re looking at it. I\'ve also seen some older Samsungs with half the LEDs shorted and the customers didn\'t complain at all about the picture until one opened and shut the array down.

One of my employees brought in a Samsung 55\" UN55JU7100 that has a
shorted primary side of the switching supply. Hard short at the output
of the first bridge rectifier. Noticed a few small value caps that are
slightly bulging - is there anything else to check/replace on this
series other than the shorted MOSFets and suspect caps?

I\'m going to suggest he add a small fan to the backside of the TV to get
longer life...

Thanks!

John :-#)#

--
(Please post followups or tech inquiries to the USENET newsgroup)
John\'s Jukes Ltd.
#7 - 3979 Marine Way, Burnaby, BC, Canada V5J 5E3
(604)872-5757 (Pinballs, Jukes, Video Games)
www.flippers.com
\"Old pinballers never die, they just flip out.\"
 
On Monday, July 25, 2022 at 2:07:23 PM UTC-4, John Robertson wrote:

One of my employees brought in a Samsung 55\" UN55JU7100 that has a
shorted primary side of the switching supply. Hard short at the output
of the first bridge rectifier. Noticed a few small value caps that are
slightly bulging - is there anything else to check/replace on this
series other than the shorted MOSFets and suspect caps?

I\'m going to suggest he add a small fan to the backside of the TV to get
longer life...

No, you\'ll be good to go. Just change the shorted MosFets and the caps, and it should fly. If you\'re very unlucky, the gate driver IC might have failed but usually the fuse will blow quick enough. If the customer changed or jumped the fuse, the gate driver IC will most likely fail on AC plug in.

You don\'t need a fan, but they way to extend the life of these is to lower the back light setting. Any TV I repair automatically gets a small mod to reduce the back light current regardless of where the customer sets the adjustment in the picture menu. Most LED TVs use a fixed LED voltage on the high side of the array and a MosFet on the low side which controls the current. I raise the value of the MosFet Source resistor (senses current) to fool the feedback.
 

Welcome to EDABoard.com

Sponsor

Back
Top