All right, we're back on this: Samsung Dumpster-dive 46 inch LCD TV And it might look a bit funny because here's the frame. It looks like it's a front, but you can see the circuitry because we've actually taken off the front bezel here with all the associated I'll flip it around with all the associated front panel our stuff on. it's got the the lead down here. It's got the speaker's over here and here at the bottom speakers.

this speakers on the back. Of course. they're like the subwoofers on the thing for the low-frequency performance, but it's got four speakers here and it's got the sensor board in here as well, so that's all capacitive touch sensing under there, so that just came out easily with that four screws. Unfortunately, one of the mounts for that actually broke off.

So yeah. so I'm not sure if that's indicative of the issue we're having, but what? I Believe the issue is our link in the previous video. If you haven't seen this thing, we couldn't really fault the power supply. All the measured values on the T-con the timing control board here all looked fine.

Voltages over on the main processor board looked fine, so it all points towards the panel itself now. I've actually? well what happens here? Okay, you got your main processor board. signal comes across this flat flex. You have played around with these connectors, nothing doing there at all.

And then you've got the timing Control Board which actually outputs all the timing and video signals for all of the LCD information. Now, because this is a full HD panel is a 1920 by 1080. so it's got 1920 columns. so if you actually count the number of conductors in there between these two cables going to the panel, there's of course not going to be 1920 columns on there for the damn thing, right, let alone the horizontal stuff.

So um, yeah, so these actually go to little boards you might be able to Here we go. I'll see if I can get the camera over, you might be able to see that little board in there and see that board. There's a board which were runs all the way along here goes all the way right out to either side. So there's two boards that go in different directions like this and they would have all the column drivers on them that actually I go down.

So it's most likely that they're the way this thing is constructed is that this connects up into here and then you've got this other they we'll call it like the driver board or something that is tucked away inside here and then there'll be multiple flat flexes again with like a hot bar solder attachment probably direct to the flat flex bonded flat flex bonded directly onto the PCB itself. using a hot bar technique it's usually called. that's most likely I Don't know I Haven't opened up this one a particular one before, but there's probably a few of them along there and then these and they'll be circuitry to actually demultiplex that stuff into the panel. Or the panel actually could have the circuitry on the flat flex itself.

That would be like a chip on flex technology or something like that. So I think there's most likely to be an issue right up here? Yes, I have actually, you know, feel, fiddled around with these and reset them and everything while the TV is going and couldn't see any difference whatsoever. So it's not that so. I think it's further on inside the panel.

So what we're going to try and do now is somehow get access to that so we can actually see and play around with those potentially. And there's the whole front panel. Actually, it looks rather jazzy on its own. I Really like it.

It's not like a shame to put the glass around. The thing actually looks kind of like industrial. David - what are you reckon? way better, way better. He thinks it's way better than having their wanky glass around the outside.

Yep, yeah, absolutely all right. So this has screws in here and if we take those out, maybe this top this whole top bezel will lift off and then we'll should be able to see the flat flex connections. Let's give it a go. Yeah, there we go.

we got it. Oh, we're in like Flynn and it's exactly as I said. There's a few more than what I thought, but I'll show you that in a sec. Right here we go exactly as I suspected.

There's actually um, 16 of these flat flexes. Yep, they are hot by our attachment just as I suspect. I'll show you a close-up in a second and they've got the chip on board decode as well. So a thousand or so if these are all column drivers, then it's full.

HD is 1920 columns, 1920 pixels across like this. So divide that by 16. That's a hundred and twenty. So there must be a driver chip in there embedded on the Flex which drives 120 columns like that.

So here it is. Here's the flat Flex coming over from the T-con board. you can see that that's hotbar soldered directly down to there. The reason that they hunt well, it's actually like yeah, reflow / hot.

By the reason that they call it hot bar is because during at the PCB assembly house, they like literally have a bar, a metal bar that comes across and it's hot and it you know they have a G which comes in and then just presses you know down onto there like that. they sort of you know have little jigs to align it all up and then the hot bar just comes down and buh and then just sold as everything in one operation. So they've done that and therefore the that's for the T-con connection and then these all split out to these 16 different flat flexors here. Once again, hot hot bar attachment.

Don't trust that any further than they can throw it. This one looks solid. We can see the solder fill, it's in there kind of. I might get the macro lens out in a second, but these ones, these ones are definitely hot by our attached and these don't look as solid as these ones down here, but and much higher density of course so you know there could be something wrong in here.

or I can show you. Well, yeah, because it's on the top. I can show you the underside of the chip up here up there. that's a chip actually driver chip embedded onto the flat flex.

So there we go. That's a closer up shot of that and you can really see the difference in the pin pitches on those two. So the one the bottom ones, the one coming from the T-con bored and I've got high confidence in that. As I said, like I've wiggled that around and everything.

But yeah, as far as the column issues we're seeing, there could be a tab connection issue along here or it could be the chip on top and the bonding and all that sort of jazz. You have to excuse the video here. I've got to look through this mylar flat flex or whatever it is and you can actually see the connections. You can actually see some test pads in there that they've got.

So this is the underside and that chip is actually a physical. Like a large chip like that. it's actually physical. The dye is physically long and thin like that, so that that's just the way it's organized.

Logistical: e To get all the traces out the other side. so these so you don't see, you've got a relatively small number of traces coming in here. okay, like serial input or whatever it is and then here. Then you'd have like the hundred and twenty traces coming out the other side, which you can't see at this angle.

So that's why they've actually manufactured the dye like that. Now, the interesting thing about all this is you're probably wondering: where are the horizontal drivers. Well, the home. I Think this is the vertical driver chip on here, and the horizontal might actually be come in, perhaps directly from the T-con You can.

actually, maybe if you look closely, you can actually see some traces then bypassing the chip and going out into here. and then we've got another connection there. You can see that on there. they're all little traces running through there.

I Show this up in my new detail in previous videos. but yeah, there's got to be some traces snaking around the side of this whole thing. For the horizontal, the horizontal outlines as well. Because you can't just have column drivers, you've got to have intersecting horizontal ones as well.

You can. actually, this is the corner of the glass. LCD You see some traces snaking their way around here here. You can probably see like the individual pixels in there.

Now if we, maybe if we yeah you should be able to see that There we go. All the individual pixels hard to tell on my LCD screen here. yet there we go. You can see these traces actually sneaking around the side like this and going all the way down to the horizontal down here.

So I'm not sure if there's any chip on glass drivers under here for the horizontal. so I'm not actually should sure the exact mechanism they're using there to get all of the 1080 traces that they need to. They might be doing half on one side half on the other. maybe? I'm not sure.

The mechanism can't really get the entire glass panel out to take a look. Now if we actually zoom in on the side here, you can actually look, look at that. You can almost see you see something in there, something very strange and staggered through. It's almost like there's wires jumping over.

It almost looks like they're jumping over like that and making connections. So yeah, I'd love to see more detail, but it's actually really difficult to manipulate this thing under the microscope. and you know I don't at this angle your mighty actually even be--i'll it might come out better but that is that is fascinating. Look at that so staggered arrangement of three and somehow they're getting all of the horizontal drive out of that.

All right. Here we go. We've actually powered the thing up without its front panel bezel so it doesn't have what the front panel power capacitive touch switch doesn't have the front. you know all the touch switch Ystad obviously didn't need that to I power up.

Now you can see the clear vertical stripes here. so now it's trying to poke it. so let's give it a bull. Let's uh, touch our now flat flicks up there.

put some pressure on that hot bar. Oh nothing. Nothing. Okay, let's go these ones over here.

Something with it. Well I'm starting to think it may not be there. I'm starting to think it may not be these either. I Would have expected.

yeah, if you push these down, maybe ya know it's gonna be hearing both wards. Yeah, there's a separate boards. Yep, they are two separate boards and those are two separate ribbons. Yep, so you'd need it.

Looks like two super favors. Okay, we might be wrong. It may not be these connections, so they look, they're They're pretty solid. They're pretty so I'd expect to see something something changing there but not a sausage.

No, that is rock solid. So is it. Is it. The T-con Is there something wrong with the T-con board because it doesn't look like there's anything wrong with these flat flexes as you saw in the previous video where actually I got out of a thermal spray the freezer spray and we and we froze those big ASIC chips on there and it didn't do anything so there was any sort of you know dry joint BGA issue then that it doesn't always show up with the freezer spray but you know odds are that it should have shown something and it didn't So I don't know.

There you go. So much for the flat flex 3 Ah bummer. Love the relays clicking. Yeah I Don't think I trust a product that doesn't even realize, just instills confidence in you when you power something I realize go click, click click I my input mains.

so that's yes. See, it's not there. nothing and let us start it. Then there's some couple little red dots there.

Oh and then then it start there it is. it's coming in. Is there a video like filter? Is this something? I'm like yeah I know because this doesn't pass through any analog pass. it's it's entirely.

It's entirely digital and this and it remains regardless of where thee of what you know, we do the test signal that's still there as well. So look yeah, that's so that that is very, very strange. This way, worse does get worse with time. So that's not indicative of a flat flex is you? So there you go? Have to choose a different path.

Choose your own adventure, right? We've got a theory because those horizontal lines went. they'll common all the way through. and we've got two separate boards like this. It's effectively split it, so it seems to be happening further back in the chain somewhere.

So what we've done is, we've disconnected this side of the paddle and we'll see what happens. Apply power Physically disconnected this side. So here we go. Yeah, I can see it There we go.

so we're still getting the lines. We're getting half those lines. Yep, there we go. But all the horizontal lines are still there and the vertical lines are all still there.

So it's yep. We'll just swap that over. And of course I think we'll see exactly the same thing. but this half will work so that in that will indicate yeah and see it coming.

That look, you see how it fades in, know since worse. that greenness. Yeah, that's you know. It does kind of seem worse.

And also we noticed before that a couple of these little pixels up the top Here for this red bar. They were like start, like would like vanish off and on. So it's worse. So it's weird.

it's worse. welcome and we're not getting the menu signal right. Where's that? This should be? Half a menu there? We're not getting a menu all right. So I got it back.

Yes, their feet and yeah, Dave we're socks. He's complete weirdo. I David Sorry. doesn't like being called Dave Not here comes Ericom's wait for it folks.

Wait for it. Fading In fading in. that is not indicative of a panel fault. That is not indicative at all.

I Don't think and look at that. the bars are back. Yep, yep, it's different. It was.

yeah, it was more solid before and we got no menu. Why weren't we get in a menu? I'm plugging this site so I'm not sure the exact do you know mechanism between horizontal drive. Maybe there was no yeah I think what was happening. Maybe that side is controlling more like the the horizontal or more of the horizontal or something like that.

I don't know I Haven't thought about the architecture of the driving mechanism behind it yet, but I I've got high confidence in the panel I I'm not suspecting any of these flat flexes at all. all the chip on flex drivers, all the hotbar attachments I'm not suspecting those because we're fiddled and diddled with them and nothing's so. a city near thinking I Think we both agree it's got to be the T-con board because if it's the processor board, um, yeah. I Like you wouldn't be getting the menus - probably in Syria there probably be like eight I Reckon there's like, you know, a bunch like a doll 16 or something differential pairs over that sexy ribbon cable that will show you the back that connects the processor board of the T-con board.

And if it's a serial problem, you'd expect a really widespread right currently carnage, right? They'd be carnage, You wouldn't get the menu with, you know, doing all its funkiness. Although if you remember in the first video which David to is not seen the we were getting the ghosting. a weird ghosting on the menus. so that's gone and that's gone I only saw that once so but I did capture it on film celluloid.

You know we shoot this in like 70 millimeter and we're using the 8-track sound system, you know? So yeah, this camera takes up Martha Love it's cozy. Yep. so if you want the original source footage for the videos, just send me a self-addressed stamped envelope and I'll send you the original 8-track tapes tapes and do that. A good ghosting is back since.

I what did you do I don't know - David was tapping something I was sad David with typically I remember I was like tapping the ground or something I was tapping something. Oh yeah, sure, any changes. no no it was. it was like plastic tapping like that original one.

like that first tap he did I wasn't I wasn't taking it seriously Woods day was zoned out tapping the duck TV he found whatever the hell it was. Oh, we do know that I don't you can see me I'm foo. Um, we do know that the yet the main ASIC on the processor board is driving a whole bunch of differential pairs as you'd expect over that main cable that black cable we saw inside over to the T-con board. it's being driven directly.

That sounds like the tapping that you were doing. Yeah, okay, that sounds the same. Okay, well, that's the main AC on the processor board. although it was shortly before that sprains some of the memory, some of the basic memory that's surrounding it, so it only takes one bit together.

Easy to get every section. Yeah, every second row liner or some row missing. Yeah, that's right. Yeah.

so I don't know. Here's our money. Now on the processor board. you were definitely tapping the processor board.

Row I was yeah. I Haven't been tapping the T-con Well, we've got to choose something either the T-con board or the processor board. And of course we're going to choose wrong. Murphy will ensure that happens.

So we're going to. um, but usually like T-con boards are more popular, our failure modes in fact, by far the most popular. But we do suspect something on the process of all because we were tapping around here and we made a difference and we're freezing around here. made a difference.

So we don't like the look of the so we think so. Either the ASIC or it's the memory and little pain in the arse BGA Memory chips in there are. they're actually easy to reflow. We don't have to take the bought out or anything like that because they're a low thermal mass.

Everything else. you know, we don't need a preheater underneath or whatever ideal. just pre heat on the top a little bit. Then we'll go in for the kill doing those.

And yes, so that's really quick and simple Just to rule out those memory chips there. so we'll just give it a go. You know I don't like our chances, but ah, are you feeling lucky Punk? No. Go for it.

Plug it in, set a camera between your legs. Are you just pleased to see me? All right? we have. What we've done is we've reflow 'dry heated the the D rams. The surrounding the is our yep, our friend the red.

Yep, not. it's exactly the same. Alright, so weary flowed those memory chips and nothing. Wow, thanks for playing.

And the other thing we're perplexed about is why it only seems to be yet. Well, vertical issues here only seem to be on things which line up with ASCII characters. You see, we've got our characters here and the vertical lines like our each I'm aligned with each ASCII character. The PC over here is like orange, but we get in like a red or it was what is it? yellow.

But we're getting the red stripes lining up with the ASCII characters and this one's interesting. Lines up with our one colon there, but we don't have I know there's a fight. one stream. Lisa Yeah, there is a faint one trail in the outer pixels of those two characters, so maybe it's like it's almost as if like the density of the character is causing a you know, a vertical stripe issue.

It's just bizarre. See what? We don't get anything on these vertical lines here, which is solid bar. Oh, is it yeah, yeah. big solid bars.

so it's like graphical. It only happens on text insertion side of things. Almost like it, you know there. Plus, we've still got the issue with the horizontal lines right, but that may be may or may not be related.

There could be two separate faults, who knows, but one of them seems to be definitely related to ASCII characters. It's weird. you can see we've got a different menu up now. We've put the Bezel back on so that we can access the menu and see it follows the it follows the text and menus and things like that.

So it's really, it's really rather bizarre and it stays there for a bit and then we can eventually get it to do other stuff. It's just yeah, it's something sticking to all the menus. Something like that. Anyway, all right.

I'm starting to strongly suspect this T-con board now I Think it might be a red herring with the tapping on the processor board making it go completely away. although we did see it go completely away. So anyway, I'm the T-con board. It could be a bad BGA joint under either the memory all these are two Asics Here it's you know unlikely to be these are flat packs and things like that.

These BGA's are classic culprits for that. I've already actually refloat the two memory chips here and it didn't do anything. So the next step is to get Medieval on its ass and whack it in the reflow oven. I'll put it through a standard temperature profile I can't remember what a temperature profile I've got set up for this beater layout one as you seen in a in a previous video.

but yeah, it's like a pretty stock standard art temperature profile so it should work the business for this. I'll just whack that in there and it'll take you know, five minutes or something. hopefully. I heat everything up, reflow all the joints, and if there's any problem in the joints at all, hopefully it would fix it fingers crossed.

but I don't know I don't think the chances are too high I'm not sure what the actual failure modes are with this particular T-con board in particular anyway, but yeah, I meant T-con board classic culprit in this scenario for the LCD But yeah, I've like you know. I've tried the flat flex ribbon I've been I've gone to town on this flat flex and I can't find any issues barring going in there and actually measuring the receipt, getting differential probe, and measuring the receiver points and everything to make sure everything's okay. Easier just to reflow the thing so we'll give that a ball. and yes, this is actually a double-sided load so you know ideally you want to glue on the second side.

Components can't actually see any glue on these puppies. but anyway, we'll give it a go I'm most likely there is, so they might have tiny little dabs of glue on them. hopefully. Anyway, the surface tension of the solder should keep out the components there on the bottom.

there only very low mass components. so yeah, fingers crossed yet again. And if you haven't seen my little beater layout reflow where controller before it is quite nice so you can buy it from beetle layout and it's got a learn mode as well. We can actually you know set up the thermal profile for your particular oven which is excellent and I just worked the yaw.

So I've already like our pre program this for a solder up profile for this oven. so at the moment it's in an art preheat mode that will go into soak. Then this is where the reflow happens. It gives like a little at peak to the temperature profile, then dwells for a while and it's all finished so well it'll take a while.

There we go. It just finished its reflow process and a little bit more left. As you can see it's up nearing up around 230 degrees C in there give or take. and the only real problem with these toaster ovens is that they don't cool down very quick.

So the way you fix that, we're finished now and we can just whoo open the door. All right Here we go. every flowed that T-con board. So if there were any bad solder joints on there, hopefully you know it might have done something.

Reef load them as the name suggests. So let's give it a bell. I Don't think we'll get that lucky, but you never, No, never, no, no, no No. I can still see some horizontal stuff happening.

No. I think we'll see our red stripes come back? Yep, there. Yep, they're going to start. No, there you go, re flowing.

The T-con board didn't do a thing well. good news is didn't damage it. Hmm. I Just wanted to point this out.

Look at the beautiful length matching on these LVDS differential pair traces here. Check this out. you'll see because of the shape of the thing. okay shape of the thing.

The the pair on the inside here is going to be a little bit shorter than the pair on the outside pair on the inside. Here we'll need a lot more wiggles in it to actually get a longer length to match the outer one. So if we actually zoom right in here, you can see that these outer might see they have little Wiggles in there to match the length and the outer ones actually stop there and there and it gets progressively shorter. They stop at different lengths until so this pair in here is the exact same length as this pair out here so that the timing between all the data is exactly the same.

Very nice. Well there you go. I think I've had about enough again for today I haven't spent a whole day on it, spent like half an hour on the damn thing. But anyway.

um yeah. I don't know if anyone else has got any good ideas about this and let us know. we might have to start cracking out the scope or something like that. Geez, that's a bit rough, but yeah, refloat.

That board could reflow the processor board as well I don't want to do that right now. So anyway, I'll just leave that video here I Don't know. that was just like half an hour of us just bumming around with his Monda at TV sorry. Anyway, if you liked it, give it a big thumbs up and all that sort of stuff.

Catch you next time you.