I think you made one of the biggest rookie soldering mistake in this video.
I am talking about the Inventables solder sleeves where you put all the heat into the solder and not the copper! With your thin and focused heat gun nozzle the copper wires were still stone cold when the solder melted, thus the solder never wetted the copper. With a typical hand held heat gun you would heat the entire cable and solder at the same time... however depending on the cable you use you might melt the insulation as well I guess...
I think you made one of the biggest rookie soldering mistake in this video.
I am talking about the Inventables solder sleeves where you put all the heat into the solder and not the copper!
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 think this LCD is beyond economical repaire. You can buy a T-con, but of course you can't be sure that that is the problem. I had the same problem, had a number of LCD-s half working, and didn't know what to do with them. Instead of throwing them out, you should take off the LCD panel itself, revealing the backlight, and than wire the psu so that it powers up from suspend, and lights up the backlight. Usually, as I found, the power up for the psu is one line, a logic high/low, so as the backlight. And on most of the panels there is a brightness control. Now this one varies. Sometimes its just a voltage for 0 to 3.3/5 volts, controlling the brightness linearly, and sometimes it needs a pwm signal, to do the adjusting, so by just wiring it high, sould make it blast at its full brightness. It's pretty cool actually, I threw 3 of these above my desk and they seem to be very bright but not unpleasant to look at because the light is so dissipated. I think it would make a good video.
Keep up, nice videos - a regular follower of the blog, and an electrical engineering student.
We used to use these in the military for quick wire repairs... we used a small heat gun but the key is to use a round deflector for heating the crimp evenly on all sides. They are a bit of a challenge, but they do work well. Also as a side note I may be able to get you the service/parts manual for this tv, I'll have a look tomorrow and see if I can get this for you.
Hi Dave. You probably already got this idea. I Googled your brand/model and came up with:
http://www.samsung.com/au/support/model/UA46B7100WFXXY. This looks like Samsung Australia. They do have live chat - no guarantees, of course. You can also get a live phone conversation at shopjimmy.com. They specialize in supplying parts and support for LCD TVs. John
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
memory after effect suggests tcon, something with bias voltage (maybe two chips in upper left corner of tcon are bias drivers?) or broken connection
Well, that LCD screen works alot better than the 23" Apple Cinema Display I have here...
http://youtu.be/_JxAz1g_bZ0Cool thing is, it kinda looks like an animated background where the Earth is nuked....
to me it looks like a bad edge connector on the panel. to verify this you can take the panel out and run your hand across the connector and see if there is any change if it does become good you can wedge something between the metal frame and the panel. its not the proper fix but it can last for a bit.
We also had a Samsung LCD TV fail recently. Ours was different though. The power supply failed.
No bulging caps or any blown parts. It just died. I checked and the standby 5V power supply died and killed the rest. I didn't bother trying to order stuff and try to fix it. Just sent it to the service center where they will probably just swap the boards.
It was interesting though when I also saw the low profile construction. Everything was done to minimize height. I'll try to get pics when we get it back.
At first I was thinking "Oh no, not another television", however the way Samsung built it to reduce the overall depth is quite interesting. I'd assume the low profile design is also why it does not have an IEC 320 C14 inlet connector for a power cord.
I've seen this failure mode before with smaller LCD displays and with those, the failure was of the bond between the flat flex cables and LCD panel itself. At the time I investigated what it would take to repair the connections, but ultimately I just replaced the LCD assemblies.
One minor clarification, those black SMD capacitors on the driver board are not the traditional solid tantalum MnO2 (manganese dioxide) types [aka flaming capacitors of death] we are all so familiar with, but rather a tantalum polymer type.
1 2 3...no more burning tantalum capacitors and charred boards
vented/bulged caps are bad, non vented/non bulged caps are NOT automagically ok.
^^^^^^^^^^
This.
Back in the days of the capacitor plague induced mass-failures of PC motherboards, the low-ESR capacitors on FIC branded motherboards (and OEM motherboards manufactured by FIC for Compaq and others) did not swell and burst their tops. They did however increase in ESR which would cause system instability. FIC was pretty good about swapping out or repairing their own FIC branded boards, while OEMs like Compaq were not so easy to deal with since most systems that seemed to be failing were just outside of their original warranty.
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.
I'm not qualified at all to give advice but I once had an almost identical fault on a 26" NTSC Samsung LCD.
But in my case the problem was intermittent. This was useful because while it was working I could have it open, spray around with microfreeze spray and see what reproduced the fault(cold solder joint(s)?). Sure enough it was the T-Con board and after attempting to reflow to no avail, I was able to find a replacement board on ebay for $15. And that worked perfectly!
Perhaps try heating it or letting it run a long time to see if the fault temporarily goes away.
vented/bulged caps are bad, non vented/non bulged caps are NOT automagically ok.
That's why I've used an ESR meter on them. Seems ok.
vented/bulged caps are bad, non vented/non bulged caps are NOT automagically ok. post hoc ergo propter hoc
And I hate to be that guy but I believe the logical fallacy here is "Denying the Antecedent". Could be wrong though.
But the insight that the SMPS might be above an ESR meter's 100kHz signal is interesting. Do you think the effect of testing capacitors in circuit or the ESR meters' bandwidth would begin to skew measured ESR faster in these sort of circuits?
It is most likely the TCON board. Or more precisely the VCOM/GAMMA buffer. If one of the analogue voltages going into the panel is missing or wrong it will result in wrong coolurs and image sticking. You can clearly see it at the end when them image gets worse over time.
If this was a graphics card than the fault would obviously be memory.
What is that memory doing on the TCON board ? Is it maybe a frame buffer?
What is that memory doing on the TCON board ? Is it maybe a frame buffer?
Older TFTs do not have memory, because the TCON does only what its name suggests: It generates timing signals for the TFT panel.
It would be interesting if anybody knows what kind of processing is done on modern TCON boards because I have seen this several times but could not find any information about it.
My guess is, the framerate is doubled or quadrupled. Most modern TFT panels used in TVs run at 100/120 or 200/240Hz instead of the common 50/60Hz. 1920x1080 @ 60Hz uses 145MHz pixel clock. The connection to the main processor board does not look like it can handle 580MHz pixel clock ( = 14GBit/s!). Therefore it makes sense to increase the framerate directly at the TFT panel.
@Dave
If you want to give it a try repairing the TCON board, look at the ics marked on the picture. It is hard to tell without knowing the part numbers, but my guess is one of those ics is the gamma buffer. They are used to generate the voltages needed by the panel to generate the individual colours. Because LCDs don't like DC, all voltages are symmetrical to the VCOM reference voltage. If one voltage fails, the display sees a DC voltage instead of the AC signal and this produces all kind of weird effects. Those horizontal lines you have seen are typical for a problem with the row inversion.
Because the gamma buffers have to drive the capacitive load of the whole TFT LCD panel, they run hot and fail quite often in large LCD TV.
The +15V you have measured is the analogue supply voltage for the TFT panel.
Try squeezing or pinching the bottom of the screen (where the black plastic is). I bought my dad a Samsung 50" a few years ago, it started having one vertical line, and pinching it made the line disappear. Of course, you have an advanced case of vertical lines.
BTW, I think the problem is mechanical; you're tackling it as if it's electronic.
Hello Dave
When you turned on the TV, there was no visible problem for a few seconds, and then it appeared. This make me think it could indeed be heat-related. It really look like a frame buffer issue (memory or video processor). You had a good idea with the freeze-it, but when you tried, you had a static image displayed. If the processor doesn't update the video memory while you freeze the board, then the "corruption" may just be left intact in the memory. If the data doesn't get to the memory right, it would explain the corruption. Also, noticing one line out of 2 being bad on the display, and also noticing that there is 2 memory chips, one of the 2 chips could be "bad".
I would suggest doing the freeze-it again while displaying a dynamic image, like the screensaver or a live feed.
It could just be a simple thing as the memory chips not being heat-coupled to the back panel cover (missing or not-touching thermal pad for example, after all, proof being the stripped screw, it seem that the cover had been opened before).
Just an hypothesis...
Regards,
Eric
Oh no.... Not another Samsung....
This one's from a few years back but shows a common mode of failure (lasted ~2 years) that I'm sure scrapped a huge number of TV's and led to one big Class Action lawsuit that Samsung settled in the USA:
And then here is a video thanking Dave and EEVBlog for helping me fix my Philips TV (lasted ~1 year before I had to order a set of parts from ShopJimmy.com):
I wish I had a dumpster like Dave! I'd never have to buy any TV's!