EEVblog
HTML-код
- Опубликовано: 18 авг 2015
- Dave & David2 have another go at repairing the dumpster dive 46" Samsung LCD TV.
This time a further teardown checking out the LCD panel TAB connections, and then reflowing the T-CON Timing Control board.
You know how it ends. If you don't like the lack of a happy ending, don't watch.
Forum: www.eevblog.com/forum/blog/eev...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
EEVblog Amazon Store (Dave gets a cut):
astore.amazon.com/eevblogstore-20
T-Shirts: teespring.com/stores/eevblog
Donations:
www.eevblog.com/donations/
Projects:
www.eevblog.com/projects/
Electronics Info Wiki:
www.eevblog.com/wiki/ 
Наука
![BETA SQUAD vs AMP FOOTBALL MATCH [OFFICIAL STREAM]](http://i.ytimg.com/vi/ChEmvP52p0U/mqdefault.jpg)
I love these attempt to repair videos. Like a good detective story it keeps you wondering. I hope you'll continue your investigations and figure this thing out!!
Hello Dave, I've been watching your channel for about 2 years and you are the reason I'm somewhat interested in electronics. I've got no idea what you say actually means but I feel like I'm learning
RandomVideos It's true watching Dave talk about electronics and explaining things makes it so much easier to understand electronics. When I started watching Dave's videos for the first time I didn't even know what those round pads on the pcb's were. turns out there Fiducials. And well thats a thing I'll never forget now.
RandomVideos Glad to hear, thanks! Similar to when I was reading electronics magazines when I was 8. I didn't know what most of it meant, but then years later it all clicked and WHAM came all the knowledge in a flood.
Yeah, he is a good engineer and explains things well but his repair videos always seem to end up making me cringe! :) Usually accompanied by a few [facepalm]s and shouting back at the screen in anguish... :)
Keep up the good work, though, Dave!
+EEVblog Just like me you are Dave! I had boxes and boxes of pulled apart stuff under my bed when I was younger, I only realize now what I actually had and what I could've made with it :/ Never too late to get back into it :)
Loving these repair/diagnostics videos, as a tight ass dumpster diver these videos only fuel my passion to fix what others would gladly just bin.
Thanks Dave(s)
Love this fault finding problem, yes get the Scope out.
This has been fascinating Dave.
What a result if you narrow the fault, even if you can't repair it, great experience.
The fault does't appear straight away, suggesting a heat fault, maybe it is with the tecon board after all.
If the heat fault is in that board, you may need another ASIC.
I love these repair videos. I'm learning so much form all this.
I just love this kind a consumer electronics diagnostic videos. Keep them coming. :)
The problem is the LCD panel, this is a very common problem for Samsung TVs. If you disconnect one ribbon cable from the Tconn to the panel the problem might go away (only half the picture will show) and you will know the side that is causing the problem. You can also try some freeze spray on the COFs to see how it reacts.
These are very good vBlogs in this channel. Thanks Dave.
Love these videos Dave, Just getting into the electronics field and love to see how these lcd panels function.
Problems like this are so interesting. Keen for Part 3!
I like this kind of videos, its got like an investigation/ mystery theme to it. so exciting. Reminds me of the Yamaha Surround Sound Receiver tear down/ Repair. love it.
I used to heat the boards to the oold CRT TVs on the gas stove. Once I heard cracking, I would drop the boare, component side down, and all of the components would fall out. That is how I would quickly collect boxes of components for fee. Great video.
Frank
Dave, from my experience with my TV, it is indeed the T-con board. You realized that the lines are exactly where the letters of the menu are, as if the menu "caused" the problem. If you put a digital image to it, the lines will probably disappear after some minutes and the image will be totally clear until you bring up the menu again or if you try to watch anything analog and the image will probably get even worse then. I'm telling you this, because that was exactly the symptoms of my set. Replacing the T-con solved completely the problem here.
According to what I read, the chip the causes the problem is a small squared one with the code AS15-G (at least in my set)
From my experience with the same problem on LG 42" LCD TV (exactly the same manifestations), hotbar connections were the problem. That sticky conductive transparent tape (forgot the name) that bonds flatflex conductors from LCD glass to that upper board was not conducting any more. With blowing hot air from hot air gun, and more importantly putting rubber bars that mechanically pressed the connections, TV was again having good picture. That worked for about the year and then started going bad again. Problem with my TV were old CFL lamps that were getting pretty hot (glowing with more red color), and whole tv screen got hot and those contacts to the flatflex were ruined. After that 1 year of extra life, and 6 years form TV, i just bought new one...
The TV manufacturers won't like it if you find a bypass for their "Built in obsolescence" lol..
A very interesting project though. 👍🏼
Green Silver It's always funny when it ends up being 3 dollars worth of capacitors. I think they've caught on that people are fixing these issues easily, so they've stepped it up to corrupting something or just having the chips shit themselves after a while.
It is easy to make a chip die prematurely. In the NMOS process, a little too much boron in the glass passivation layers will slowly cause the boron to slowly diffuse into the chip substrate, killing the chip. This was the problem MOS Technologies (Commodore Business Machines) had with the chips used in the C64. As they heated up (they ran hot) over time that made them diffuse faster and faster.
All the vertical lines, line up exactly with the OSD text, so its looking like it is possibly an issue driving the framebuffer RAM's, I've seen the same style of issue on a much lower tech LED sign display, caused by a decoder with leaky outputs not fully turning off after drawing a pixel, so you ended up with faint trails on the next lines
can you explain for me a little more please ?
Dwwd
Dwwdd
Ten points for keeping it up. I'm sure you guys are closing in on whatever the problem is. Sorry I can't help to diagnose but the character issue seems important. It's odd that stuff is being drawn in placed where the letters never really go. The center bit moves up and down a bit but the 'PC' string pretty much stays in place. Something's not turning off when it should...Good luck guys. Hope you find it.
i love it when we are in like flynn!
hook up a video signal to it and watch it clear up after the menu text dissappear.
I have to agree, hook a Pi up to it, and check if it's just the menus, or full video problems
dreamfox32 Now that could be useful in isolating the problem. If it completely clears, the "character generator" could have a fault, or the connection from it to the next stage in the video processing. I imagine it's not a separate chip but integrated onto one of those custom jobs on the "processor board" as we're calling it. If I had to guess I would say those chips are not going to be available and full board replacement would be required.
dreamfox32 Did you miss the test image in the previous video?
hendik hendrik Wasn't the image test was generated by the processor board, and not by a completely external source?
dreamfox32 How would that repair the fault?
It would be great if you had a service manual showing testing points. A real time saver.
EEVblog Dave, this whole stuff can be because of low gate open voltage, there is a dc-dc converter on tcon board somewhere which drives gates, check voltages
Did this TV ever get fixed?
One Hint for the problem not being linked with a faulty panel might be that all the vertical lines appearing on the display are somehow bonded with the characters displayed! Red collumn -> orange "PC" letters, cluster of lines in the middle -> "Check signal cable." and the one in between -> the ":" in the time (?).
The fade-in and shadowing says "floating input somewhere" to me
mikeselectricstuff It seems doubtful that anything is addressed by column in a way that would break things across the whole column before *and* after the text except for something related directly to the panel.
I'm somewhat familiar with how this failure mode might occur with OLED panels, but I'm not that familiar with LED panels. With OLED panels, you need to reset each element to a certain voltage before writing a new scanline. It would seem here that something about each vertical column isn't getting reset between scanlines and some voltage or charge is building up across the whole column. And it's panel wide.
mikeselectricstuff That's what I was going to check with the scope if I get around to it.
I had a similar issue with a tea damaged LCD panel from an iPhone 4, but it was more in the opposite order. The panel would start up fuzzy and eventually clear up after being on for ~10 minutes. Some characters made bigger lines to be displayed as well. It has 3 chips on the panel(chip-on-glass?); two drivers apparently for each side of the long edge, and the wide chip tcon/short edge driver. The LVDS bus is fine and then phone has no issues with a new screen.
Definitely something floating...
Gustavo Ferlizi Are you British? Surely you gotta be British if you managed to damage an LCD panel with tea.
If you look closely, lines do align with "P C" (the red ones) with check cable, and with time dots. Looks like some sort of filter accumulates values or something, definitely something with the scaler/frame buffer or something like that.
Keep going. It is very interesting watching you trouble shooting. I am at the stage where I want yo to find the fault and be able to fix it. That would be great. It is always a pleasure to watch someone who would try to fix something rather than just throw it away. Good on ya.
this would be great to watch LIVE! :D
Another cool and very interesting video!
I would bet on memory problem with some of the cells. I would:
1. Install custom linux image from: www.samygo.tv/
2. run memcheck on PCU board
3. swap/change memory modules on video board
seems like it could have something to do with a chip heating up, just based on the way it fades in. you could try blowing hot air on different chips while it runs, or look for a particularly hot chip with the thermal camera.
Maybe interesting: hoe about doing a reverse repair! Introduce an error and look at the effect. Two examples 1) Cut a trace (repairable) on a FR4 based board leading to a flex cable, goal discover what a bad solder joint would introduce. 2) introduce a "stuck at" 1 or 0 error on memory databus.
Danny Bokma
Yeah, I like the idea. It could have more educational value than a repair, and maybe more spectacular results.
nativewinner Spectacularly smoky and then the fire system kicks in :)
I have repaired similar TV's and it was an issue with the ribbon cables coming off the lcd not making good contact in the connector. I would use packaging tape to make the cable a bit thicker for the fix.
Surely it's 1920xRGB=5760 columns?
mikeselectricstuff Not the first time Dave forgot pixels have rgb elements.
mikeselectricstuff Err, yeah, Doh.
+Gone Tomorrow I would assume mobile phones have a lot more on glass logic to get passed that.
***** but its not only theirs 3 color for each pixel but you also have two lines correspond to each subpixel and a transistor in the corner.
EEVblog I may sound like an amater , but does the smaller BGA on TCON board looks like it's gone...? Looks like it's burned out... Warped... Shines more than the big one?
keep up the good work Dave and David for my two peneth I would try and use your thermal camera and see what heats up at the same time your lines appear don't think you would have found it with the freezer spray as when you change the menu it takes a while for the old lines to disapear.... I have found a couple of faults that way before...
My bet is framebuffer memory on the main processing board - some rows in memory flip bits, some stay flipped, some decay - this is what happens with videocards. So possibly a memory line doesn't get a refresh signal or something like that... And because it does some postprocessing on the faulty data, it can produce different artifacts depending on when and how the bits flip.
The problem seems to be more evident with antialiased fonts drawn by osd software lib (you have the issue near the curved lines and not on the horz and vert lines that are instead perfectly row/column aligned), this suggest a timing problem with partially lighted leds...
I think you should input some video test signal from a pc software generated image and see if we have changes with full red, full green, full blue, 33% rgb, and so on...
If you look, the lines only appear where the text is, I'm guessing T-con for sure.
Can you use the oscilloscope to check the io for the pixel column that's having the problem. You can look online for what voltage you need for what colour and see if you can see anything unusual then trace it back to the chip on flex and then back to the board etc. Or see if you can a service manual/schematic. Great video.
Been thinking about this one, leaning towards the signal gen for the OSD.
All I can do is wish you both good luck on this restoration attempt.
It really would be great to you beat the gremlins spoiling an otherwise beautiful TV.
A bench with a clear top and a mirror underneath might be handy for doing this job solo. Our Sony has the exact same failure, it happens with on screen text even on the broadcast picture, putting the TV on a radio station clears the issue temporarily.
Really good videos, i was wishing that you succeed, always want to learn more, i felt as if i where there with you, don't like to be beaten. please post more on this one.
Some of these ghosting effects remind me of what some LCD displays do when you turn them off. In particular, when I switch off my old Game Boy Advance, every few rows appear to persist for a bit.
My hypothesis is that for some reason, the pixels are having a hard time turning off. When a row is being driven, the voltage for a particular column could leak through to other rows, and without a strong turn-off force, this leakage gradually causes pixels to turn on.
One possibility is that if a negative voltage rail is used to turn off pixels, then it's possible that rail isn't making it to the TCON board. Thou shalt check voltages on the TCON board. ;)
The lines appearing only where the characters are looks like a RAM addressing issue to me, like a floating / stuck address line. Since it fades, it's probably floating, so you get an average over time of white and black (text and background) on the whole column. That also explains why the lines have colors close to the text color and the striped pattern (the regular spacing of thin black lines in the red columns)
What’s the fix
You didn't put any pressure on the flat flex where it's bonded to the LCD glass, which was the location of this issue on the 2008 Samsung I had. Notice how the lines follow bright areas in the picture at any row. Maybe one of the return lines for the row drivers has lifted, and every pixel in the row slowly floats high if one is high?
Im keen too see if you do ever get this going, it would be very satisfying if you do
In the event of a suspect main ASIC or memory, feed it a video signal (VGA/HDMI) and try to change the resolution. If you have a memory error the patters will move as the picture is placed differently in RAM.
on tcon board level it will always fit exectly the same way in memory.
Yes, as it's behind the scaler board. If it was a memory error i wouldn't appear slowly like that, but be there from the beginning.
I think I would try to generate some test patterns from PC/HDMI like a single white line (h and v), half of image white etc. I would focus on a horizontal bar of white lines (5-30 lines). They should make the ghosts appear. Probing the signals should suggest at which stage the ghosts come in or where h lines disappear. Using smart display patterns you can make the signals look nice and predictable and easy to recognize when they are wrong.
This is an internal panel short, you can narrow down which signal lines the shorts on by putting capton tape over some of the pins on the ribbon end, on one of the 2 ribbons going to the panel buffer boards
Sometimes using this method you can get rid of the lines while retaining a good picture, also one of the ics on the tcon will be red hot while that shorts connected to it, that's why the lines slowly appear as the ic starts to struggle with the overload and the chip heats up
The Horizontal lines appear to be split between the two sides of the panel(like, even rows on the left and odd rows on the right) and the left horizontal lines "driver" (when the panel is the correct side up) is completely gone (hence no display). But since the vertical "fault" lines appear even on the side with "no" horizontal drive that could be some indicator that the fault originates on the main board that generates the signals for the rest of the components (it skips processing half the rows except where there is an error or "overflow" that produces the vertical lines). Hope this was helpful...
its funny to see the no signal and input information move up and down, normally only happens on plasmas to prevent burning, but lcds dont burn pictures so im suprised to see it
Check the top strip with the IC inside the one on top left is usually faulty just apply pressure on the flexible and it should work!!! i usually use a thermal pad for the IC to mold inside of it en ensure a good contact then you close it and make sure there is enough pressure to hold the contact
Reflow all the boards, hack all the things, drink all the beer :)
I'm that way myself; I can become obsessed with making something work, far beyond any practical accounting of time-spent vs. benefit derived
This trouble shooting is great!!!!! Find it very interesting. Hope you get to find a fix for it though.
The lines match up perfectly with what is on the screen, note the PC, : and e. I suspect it will happen on any part of the screen where it aligns with graphics. Could be an issue with the OSD alone. Maybe bad display RAM?
Just to echo another comment or two on here - Mike said floating input, does look that way. Someone else mentioned reflowing processor board and considering OSD text having effect I do agree it's probably the processor board. Something perhaps needs reflow on there.
GadgetUK164 There is also a very plausible theory about the gamma buffer or other analog drive system at fault.
+EEVblog Interesting. When are you going back to it to find out?
I love this guessing theory, I cant wait to see fixed and see what was the problem
+Light Man Yes,
Don't batterise, analyse!
Those blippy pixels right at the bottom (especially at the bottom of the red lines) are quite interesting...
You could try hitting the actual panel with freeze spray, maybe there's a problem with the transistors in the panel at the bottom of the display.
it is shockingly impressive the level of futuristic detail and tech applied into the flat screen.... I tend to think that if such effort was applied into... say fixing the environmental problems... like perhaps deforestation.... then there would not be any problem
I have the same exact problem on a 42" Toshiba from about the same time period. Vertical lines appear in columns containing text, and horizontal "missing" lines are constantly there. It seems to come and go over time; I've been able to use the TV for the past few months without too many problems, but sometimes it just gets in a bad mood or something and the lines cover everything up.
if the image is filling a column, it _could_ possibly be an issue other than bad column drivers or connections. If you see weird pattern in how a column fills, then look at the horizontal drivers activating/leaking onto multiple rows at once.
*and you guys consider it later in the video, good deal!
Looks like a chip on glass warming up problem to me, try your hot air iron round the outer edge of the screen while its turned on.
Check out the way they matched the length of the traces in that cable. The railroad could do this on the inside track on curves.
I've seen slightly cracked LCDs do this kind of thing. At the edges or corners, or when someone's spilled or cleaned things with the wrong chemicals, eats the metallic traces on/in the LCD itself. Can create high resistance connections, so the LCD itself 'charges up'. The common aspect, though was always a toast LCD, I'm afraid.
Good luck, keep up the cool videos. Stu.
You can buy a working T-Con board on E-Bay for about or less than 60.00....From a HDTV with a broken screen they sell all the good parts from it....power board and main boards....inverter...
After seeing what you have tested and poked i would call a dry joint on the processor board... And if i had to point to the exact bit i would say a bad earth on the LVDS connection.
Note to self: I must go outside again one day. Only another 780 episodes of EEVBlog to go...............
This is why TV repair is best left to specialists who have the test equipment. For an electronics engineer wanting to make spare money in their time, it's just not economical enough, but perfectly OK for a blogger on RUclips if it means a large audience.
There have been up to now 10.000 people including Dave and David learning and being entertained from this dumpster-dive TV. Not sure if there's a way to get more bang for the (zero) bucks. ;-)
nice traying to save the LCD TV nice video 👍
hey . have you tried just plugging an input in and see if it works ?? the lines don't se to be affecting the graphics,just the ASCII .
Well, looks like I was wrong about the TAB's being the issue. It looks very similar to a tab issue, but upon your further investigation, it became obvious that was not it. Ah well, those panels are a marvel to see anyway, so wasn't a complete waste of time.
+EEVBlog If you still have the LCD TV Dave, have you considered the clock management?
Clock based timing issues could potentially explain the ghosting as well as the lines which could be following your screen input.
But to see if the lines are just input based as in menues etc just hook the TV to a PC as the secondary screen and make the background white so the TV only shows a "pure" white screen. If a "pure" color can be shown without lines will help you eliminate possible sources for the fault.
I believe it's to do with the LCD itself I've seen this before where the LCD broke and caused lines and dead pixels so a replacement of the LCD would be required
It is entirely possible is not the LCD if so it's the caps which have been fixed but then the tecon board can be the problem but then if it's not that then it's the main board but I've not seen TVs that are built like this and I don't believe after all this time that it's possible for me or most people to be any help but good video I really found it interesting how much could of been wrong with it
Dave! Check Vgh and Vgl voltage. please check the voltage with a oscilloscope.
Another painful troubleshooting session!
Holy cow, you actually called him David for a change - after a dozen times saying he doesn't like being called Dave!
Let's see:
1. The lcd panel cables are good, if the lines would appear instantly after powering the TV on than flex cables would be faulty
2. the panel is not damaged that is a fact, a damaged panel would not have straight lines but bleeding colours
3. power supply is good
4. Image processoring is working because there is video output to the panel
5. The issue is on the video output source
So if you put all this together you will se it is related to the T-con, something on the board is damaged because when it is heating up the lines appear slowely (vertical lines) these other horizontal lines are most likely because of some small voltage disruptions on the T-con. I can not say what exactly is damaged on the T-con but you can try and look deeper in to what could be wrong on the T-con's circuit, if the chip is damaged than it is best to buy a whole new board
Here is one, I don't know what model do you have de.aliexpress.com/store/product/LN46B650-or-LED-UN46B7000-BOARD-2009FA7M4C4LV0-9/710086_614153320.html?storeId=710086
Lower the brightness level and color intensity and when the menu text dissappear it goes to normal viewing
dreamfox32 Well, it _might_ be reasonably tolerable, but even if the image produced is flawless, I think that would be to miss the point: to make an interesting electronics repair (attempt) video. I don't think Dave is hurting for TVs.
If that was mine I would probably just leave it at the appliance recycling place, and buy a new and better TV. They are so cheap now that they're practically disposable. That said, my 40" Sony LCD from 2007 is running like a champ. It's 10cm thick, but I'm not looking at the edge of the damn thing. :)
Did either Dave mention if all the inputs produced the same result? Especially curious about the other HDMI inputs as well as Component and VGA (does it have this one?) inputs. I don't remember seeing any of the others tested on either video so far. Only the PC (HDMI). Just isolating the issue.
Either way, I love these dumpster repair vids. Very educational to hear Dave's thought process on troubleshooting and testing. I hope they can figure it out!
As you have problems on both sides of the split it's probably on the TCON that the problem lies in, as the likelyhood of both sides of the panel TAB bonds going bad at the same time is pretty low. As the problem looks "analog" the scaler board (what you call the main board) is out of the question. A lot of "analog" stuff goes on in the TCON.
Good thing you mentioned that, just like Dave 2
Danny Bokma Exactly.
Danny Bokma Is there anything wrong by getting a second opinion? I have real experience repairing TV's, which Dave or Dave2 does not seem to have.
you had the same explanation, and of course there's nothing wrong in that.
***** Yes, that's the conclusion we came to. That's why I then reflowed the T-CON board.
I was initially suspecting that some of the power rails' voltages were drifting after being turned on, but then when you noticed that the streaks were associated with the text, that made that seem unlikely. (And when you tested the voltages, it had presumably been on for several seconds.) But some of the recent replies to Mike's comment explain how that could actually be the cause. Weird, though, that the text that spans almost the full width of the screen at 22:40 doesn't cause streaks. Maybe that's because it's not as bright as the other text, as Mila Miglia said-the bright icons at 22:25 certainly cause brighter streaks.
I'm really happy that you did a part 2, but in accordance I'd be way more thrilled if you did a part 3 and fix this puppy. I know it will take all the guns but please can you do it. Love your videos rodders520.
I'm not saying I know for a fact that the panel isn't bad, but one of my best friends does tv repair as a side business and he's had this exact problem a couple times and it was the t-con.
try your thermal camera on the back, see what warms up, freeze that, as it related to text, how about a video signal input? please try that with no generated text on
without having watched the first video, or having the time right now, I'm wondering if this could still be a thermally caused issue. If it was, applying heat or cold to various components would be helpful to find the problem.
I am getting the same issue with water damaged cell phone - it draws lines where are some contrast items. In my case i was thinking that the water+impurities are connecting whole lines together - so if anything in the line light us it affects the other pixels.
Also. Please see if the lines react to text or patterns on a video input. My bet is they don't because they are caused by weakly grounded physical memory during the compositing process that overlays onscreen display elements.
The vertical issue was all over the test image as well? Perhaps not specific to ASCII characters rather to intensity?
YUS PART TWO
i think, that there is something wrong with internal power rails - when it drives pixel bright and relaxes transistor, it has lack of power to return transistor back. or there is current leakage in the panel itself from horizontal lines to vertical
Challenge! Since its occuring in the uncompressed video signal I bet/vote memory buffer on tcon board. Or ASIC ofcource, but I think that CMOS process was better controlled than that of the more challenging memory.
The way the fault fades in has to be a vital clue. Also - why not put in your own signal?
The display on my smartwatch did the same thing, with the blue and red vertical lines as well as the horizontal 'scan-lines'. The fault fixed itself over the course of a day or so, and appeared to have come from the wireless charger. Probably not relevant but it could be... :D
As Mike said, it smells as "floating input" and it stinks of "destroyed weak pull downs in the control bga"
If this was on a PC i would say it is a graphic memory overflow that doesn't get cleared every frame and if the TV is entirely digital maybe there is a graphics rendering chip where the problem might be and it appears to be affected by the characters draw function.
You know your video camera is high quality when you can see the specs of dust and lint floating in the air on the video.
Have you tried connecting an input via HDMI yet? Ignoring the built in menus & overlays.
Thermal issue with the graphics processor or RAM. Bad ball grid array joint Maybe. Sometimes the fine leads that connect from the terminals to the die in ICs can come loose either from thermal stress or a fracture on the lead's pad on the die can cause this type of error as well. It does have some hallmark indications of thermal caused lead to die failure
great video, have you considered the processor chips themselves are bad? that's what It looks like to me, especially have you've checked all the regular suspects. The other thing it could be if that LCD is anything like the old iMac LCDs from 2006 is heating actually damaging the LCD itself, but if that were the case, the lines would be present as soon as you turned on the set. This fades in like it happens as something is heating up, so it could still be a defect triggered by heat.