EEVblog #780 - Samsung LCD TV Dumpster Dive Teardown

[German_EE]

Some of the techniques used to achieve a low profile were interesting, especially the daughter boards used for the relays. Some of those manufactured with space for a SMD back EMF diode would be a useful space filler on a China manufactured PCB

[Frost]

those solder sleeves have been around for a long time.

Yes, but what's their advantage?

I normally use crimp butt connectors of the CrimpSeal II series
from DSG-Canusa, this are ordinary crimp connectors
with a shrink tube and an inner adhesive layer which
melts during the shrink process.
They work very well, are 100% watertight after shrinking
and very easy to use.

[Dino KL0S]

I was curious about Dave's problem having the solder flow in those solder sleeves since they looked like an interesting thing to have around. I e-mailed customer service at Inventables and here's what they replied...hmmm, wonder if anyone has actually had good luck with them?

Subject: Solder Sleeves
AUG 18, 2015  |  01:42PM CDT
Sam replied:
Hi,

We do not actually have the melting point for this item, but it should be the same as plain solders 183 °C. We sell a heat gun that works really well with them:

https://www.inventables.com/technologies/heat-gun


Samantha Jo Alaimo
Customer Success Specialist
Inventables

AUG 17, 2015  |  06:55PM CDT
Original message
Dino wrote:

What temperature is required to activate the solder in the sleeves? I was very interested in these but in a recent EEVBlog episode Dave Jones had difficulty heating the joint to a point where the solder would melt and flow. Your product video example shows a large heat source and very quick melt, albeit with a smaller wire gauge and sleeve size. Thanks in advance!

[edy]

Sure you could blast the whole thing with a huge wide nozzle heat gun, but as you said, wire insulation would likely melt.
My nozzle was small, but I thought should have been big enough to do the job, I was trying to heat the wires as well.
But yeah, bigger nozzle next time.

I wonder if it would help to dip the 2 wire ends into some rosin flux and get them all loaded up before sticking them into the Shrink-Wrap-Solder tube? Then you would see your flux getting all bubbly and melting away on the copper. It might look like a mess, but perhaps would improve solder flow?

[max_torque]

those solder sleeves have been around for a long time.

Yes, but what's their advantage?

I normally use crimp butt connectors of the CrimpSeal II series
from DSG-Canusa, this are ordinary crimp connectors
with a shrink tube and an inner adhesive layer which
melts during the shrink process.
They work very well, are 100% watertight after shrinking
and very easy to use.

I have no idea why you would want to solder a splice, when you could crimp it?

Just get the normal AMP/Tyco  adhesive sealing crimps you mention...

(solder is a really bad way to join wires, especially unsupported wires subject to ANY vibration or bending, as solder wicks up wire, leaving a rigid section, at which point any vibrations are concentrated and cause a failure.  Not for nothing does the space,aero and automotive industry crimp EVERYTHING)

[artag]

those solder sleeves have been around for a long time.

invented by AMP ( Tyco )

http://www.digikey.com/catalog/en/search?filters=18111

Invented by Raychem, I think. These might be clones. They're cheaper than the Raychem ones.

http://www.farnell.com/datasheets/1680945.pdf

Tyco, AMP and Raychem are all owned by TE Connectivity now, though.

[coromonadalix]

Lcd had a lot of quality problems before, the lcd controllers matrix inside the lcd panel had bad contacts, they were glued/tacked on pcbs and had poor contacts.

I founded a very beautiful 46" lcd tv who had half the vertical display working, very good at cold  but problems when hot.

Google search  told me a lot of lcd failures had that problem, i cant recall the lcd panel maker.

[zze110]

Here's how to get the TV in to test pattern mode

TV in standby mode.  Press Mute 1 8 2 Then power On... should put you in service mode.
Select "svc/service/control" , "test mode or test pattern", "FRC pattern"
Pattern in generated from T-CON board directly.

The tv may have a hidden menu also.   Try 0 2 1 4 exit if needed

In the documentation I found... "vertical or horizontal lines: defective panel likely but also t-con, lvds, or mainboard.   Use test patterns in factory service mode to determine error."

When done, press power button to save settings.

[PA0PBZ]

TV in standby mode.  Press Mute 1 8 2 Then power On... should put you in service mode.

You need the remote for that...

[edy]

A cheap universal remote should do the trick. Handy to have around and reprogram to use for any dumpster dives missing the original remote. Hopefully these newer TV models and old ones are all supported. Most should have a long list of models/makes they are compatible with and should allow direct programming to the TV you want.

[TheAmmoniacal]

Did you catch this video?



In brief: If you enter the service menu, and click through to set the sub contrast down, it pretty much removes the lines: ADC/WB -> White Balance -> Sub Contrast: 70

It seems like you need the remote to enter the service menu, maybe you can find a way to enter that menu with the capacitive buttons? (don't know the model so I can't check the user manual)

[rollatorwieltje]

If you have a phone or tablet with IR port there are several apps that can mimic actual remotes.
For Samsung there's also Samygo Remote, it uses wifi or ethernet.

[bktemp]

After seeing the second video, it is clearly a faulty gamma buffer or some other analogue voltage driving the panel:
You only see the vertical lines on bright colours, because only then the voltage from the faulty gamma buffer channel is used.
The stripes get worse over time because a DC offset build up in the panel and saturates the pixels.

[davidmc]

Just an idea from watching the video, and I cant see pixels clearly enough to check this in the video, but is it possible the issue is actually in the horizontal line drive, with the vertical bars then resulting from the problem rows 'averaging' the signal that they see with their selects floating or similar.

If it were the left side horizontal drive (right when upside down), this would be consistent also with the behaviour where connecting the right side of the screen produced half a menu while connecting just the left didn't. Pne way it still has the good row signals, bad ones discconected, while the other way the only connected rows are the bad ones.

[Kerden]

Just an idea from watching the video, and I cant see pixels clearly enough to check this in the video, but is it possible the issue is actually in the horizontal line drive, with the vertical bars then resulting from the problem rows 'averaging' the signal that they see with their selects floating or similar.

If it were the left side horizontal drive (right when upside down), this would be consistent also with the behaviour where connecting the right side of the screen produced half a menu while connecting just the left didn't. Pne way it still has the good row signals, bad ones discconected, while the other way the only connected rows are the bad ones.

I'll second that!
It should be fairly easy to test: If this is true, the vertical lines will only show in the rows which do not display the correct video signal. so the "striped-rows" pattern should be inverted.
As davidmc said, this would mean that the rows are floating and will build up charge, until the voltage difference is enough to switch on the pixels. This would also explain why the vertical bars match the colour of the displayed signal.

[Rasz]

Did you catch this video?

In brief: If you enter the service menu, and click through to set the sub contrast down, it pretty much removes the lines: ADC/WB -> White Balance -> Sub Contrast: 70

nice find man! this and same effect even with only one flex connected confirms dc bias on the tcon board as the culprit

[TeaNTronics]

I was curious about Dave's problem having the solder flow in those solder sleeves since they looked like an interesting thing to have around. I e-mailed customer service at Inventables and here's what they replied...hmmm, wonder if anyone has actually had good luck with them?

Subject: Solder Sleeves
AUG 18, 2015  |  01:42PM CDT
Sam replied:
Hi,

We do not actually have the melting point for this item, but it should be the same as plain solders 183 °C. We sell a heat gun that works really well with them:

https://www.inventables.com/technologies/heat-gun

no wonder it didn't worked well for dave,
in their website, they say it got 2 temperature settings:
750 °F (398.89 °C) and 1000°F (537.78 °C)
it's not a low melting point solder after all
p.s. it's looks like paint removing heat gun that you can get in hardware store

[Protonus]

Dave - check those Sam Young filter caps on the power supply board using your ESR meter!  I don't see that you did that in either video?  I know you have no reason to think they're bad based on ripple and visual diagnosis.  BUT - I have a very similar vintage Samsung LCD TV (2009, 55", high end) I'm working on, that also has multiple Sam Young filter caps on a very similar board.  One of them was visibly bulged, but even the ones that weren't in the same value (25V 1000uF) had questionable/high ESR values. 

It has the exact same symptoms yours does.

I'm waiting for the replacement caps to arrive, but given the nature of the failure (it worsened with time according to the owner, used to go away on it's own, and it's temperature related) etc, and the fact that one had visibly failed - I'm confident that this is what's wrong.

[Rasz]

Dave - check those Sam Young filter caps on the power supply board using your ESR meter!  I don't see that you did that in either video?  I know you have no reason to think they're bad based on ripple and visual diagnosis.  BUT - I have a very similar vintage Samsung LCD TV (2009, 55", high end) I'm working on, that also has multiple Sam Young filter caps on a very similar board.  One of them was visibly bulged, but even the ones that weren't in the same value (25V 1000uF) had questionable/high ESR values. 

It has the exact same symptoms yours does.

I'm waiting for the replacement caps to arrive, but given the nature of the failure (it worsened with time according to the owner, used to go away on it's own, and it's temperature related) etc, and the fact that one had visibly failed - I'm confident that this is what's wrong.

its not, in this particular case its not the electrolytic caps for a change

[Godzil]

Dave: could you do some close-up picture on the bar on the screen?

I wonder if some of the "magic" filter that TV manufacturer tends to put are not what started to failed.

The other possibility is as the video processor is likely to handle different alpha blended framebuffer, the one used to display text (which could be different from the one used for the boxe) have some problem.

It's unlikely to be a software problem, but who knows, that's true it's strange that it only happen on the text shapes.

[Stephan_T]

Hi Rasz,

vented/bulged caps are bad, non vented/non bulged caps are NOT automagically ok. post hoc ergo propter hoc
smps can work above 100KHz, better to check using oscilloscope or measure cap esr directly, not everyone has a $400 meter with enough bandwidth

At that moment of the video, I was almost about to scream at the screen, but then Dave silently changed the multimeter. I think, there should have been a comment about the reason. The spec of the Brymen 257 (wich I also use) lists for AC: 50 Hz ~ 400Hz. Showing the ripple on a scope would have been more educating.

[Rasz]

The other possibility is as the video processor is likely to handle different alpha blended framebuffer, the one used to display text (which could be different from the one used for the boxe) have some problem.

its not only text, it manifests in columns with high brightness(gamma) elements like icons, or whole test image from previous video

btw I would LOOOVE a video tutorial about driving modern LCD, what is exactly in the Tcon, how it all works, VCom, gamma, shifters and drivers, all that magical high speed digital/analog mess. Preferably done with help from someone who did it for a living.

[Godzil]

Rasz: I didn't notice that but ok, make more sense in that case.

That's still weird, I remember having strange behavior on some LCD panel with the timings was't correct, and having bar like that was one of the problem I had, but it was on really "old" technology like FSTN graphical LCD screen

[rad.wiel]

Hi Dave!  I just found some interesting info on one of Polish forums. http://www.elektroda.pl/rtvforum/topic2285209.html One guy has similar problem and discovered, that there is problem with flat flex connectors and traces made on matrix glass. He used ohmmeter to check if there are connections between test pads on pcb and test pads on flat flex cables. As I can understand and translate he had no connection between test pad marked as "voff" and device called "side driver". He used thin copper enamel wire to make missing connections and this fixed the problem.

Btw. All matrix made in Slovakia for Samsung in 2008 and 2009 has similar technical flaw (model of matrix: Ultra Clear Panel
S-PVA 4ms).

Sorry for my english. 

Photos (driver boards/flat flex can be different for different models) http://obrazki.elektroda.pl/8159757800_1380278939.jpg http://www.elektroda.pl/rtvforum/files-rtvforum/bez_tytu_u_1716.jpg

[averox]

Dave, I don't know if you've seen this comment on your video but this guy seems to know what's going on and there's alot of people who agree with him, here's the comment.

Dave, these TV's have problems with the TAB(Tape Automated Bond) coming loose, where the flex cable connects to the actual panel. The fix is to use some foam or something to help reinforce the TAB's connection to the panel.

Nope, tried that.

You should try giving it another shot as some of the tabs can be simple or downright finicky to get the pressure right. I've had to get really tricky with some tabs before with shoving things under and then back on top to get them to work right. Don't burn your finger though because those pieces of silicon on the tabs do get quite hot. I've seen this a fair amount in TVs I've tried to fix but I wasn't able to fix them all despite trying my best.

Good luck!