EEVblog #447 - Samsung Plasma Followup.

[firewalker]

Maybe you should ask norcal715.

https://www.youtube.com/feed/UCJTrnevHaK73BbOxfteY5Gw

Alexander.

[tom66]

The coloured noise may be maldischarge. You will need to touch the high voltage section -- adjust the Yrr and Yfr on the Y-main board, until the picture improves. I have fixed a few Samsung plasmas like this with maldischarge, fairly common.

[Spawn]

Dave, I got one those VGA to composite converters different brand and it is doing same flickering from day one, I am not sure why but it is pretty useless.

[Thor-Arne]

I have a newer model Samsung 42" Plasma, it has issues with channel separation when using the analog inputs. In my case this shown as ghost images, perhaps it's something like that happening.

[nitro2k01]

When you had the the converter box into the composite input, I noticed you had an OSD menu pulled up, and FLICK was selected in red. Maybe you accidentally put the menu pointer on an item which generates random noise bursts to test how a monitor reacts to an unstable composite signal?

[codeboy2k]

The coloured noise may be maldischarge. You will need to touch the high voltage section -- adjust the Yrr and Yfr on the Y-main board, until the picture improves. I have fixed a few Samsung plasmas like this with maldischarge, fairly common.

IF that was the case, then I would expect the HDMI input to be just as bad, but it's clean.  I expect the problem is more likely still inside the Philips SAA7 ADC chip.

1) in the beginning, all the composite analog inputs that feed into this chip would not work. Now it's been seen that some of them do work after Dave put on his lab coat (at least the ones that Dave could test). 

2) the path from the HDMI input is probably direct to the Samsung ASIC's as Dave also believes, and this bypasses the phillips MUX/ADC chip, which then goes into the ASIC's

when you consider 1 and 2 together, it looks more and more like the culprit is the Philips MUX/ADC chip, or possibly something around it, causing ADC issues. If Dave is able to test all analog inputs, and sees that colorization on all analog inputs, I would suspect the ADCs have gone wacky in the chip. Since RED and BLUE makes PURPLE, I am led to believe that both the R and B ADC's have gone bad, possibly causes by some overheating event in its past, as other forum members have made note about.

[tom66]

The display could be switching between 50 and 60Hz, which can affect maldischarge. However, I would be considering the main board more myself. 

[Bored@Work]

Not wearing shoes? I'll start to worry if Dave stops wearing pants in the videos 

[Neganur]

seems like one resistor is missing, although I don't think it's responsible for the thing not working properly.

[SeanB]

It was never there, it is a pad for an extra decoupling capacitor. I would put a guess the Phillips BGA is going intermittent internally when hot, they are rather prone to failing, at least the older satellite decoders here that used a similar chip and a small ADC on the IF converter that ran almost white hot did so. Not helped by them being about the right shape to be pizza boxes, and having about the same amount of cooling.

[Stonent]

I had wondered about tapping on the Philips BGA chip with the handle of a screwdriver to see if the screen cuts out or flickers.
I'm wondering if laying flat and standing up may have pressed a ball into place. I'm still thinking cold solder joint, or crappy RoHS solder thermal stress break.

I once got a small HP laser printer working after baking the motherboard for 10 minutes in the oven.

[SeanB]

It did get both cryocooling and percussive maintenance by Dave, so not likely to be a dry joint or thermal. Might be a bad memory chip, they were not cryofrozen or fondled, and the ADC does deliver to them, while HDMI goes to a different memory so possibly only one is failing intermittently.

[NiHaoMike]

Don't waste any more time on it since the HDMI works. As for the stand, couldn't you build one out of wood?

It's also interesting that there's a big empty space to the right, as if there was another model that had some hard drives or something.

[tom66]

The empty space is probably for a PIP tuner.

LG did produce some older LCD and plasma TVs with a built in PVR. Genius LG however made it so when the HDD died the entire TV stopped working -- which occurred surprisingly often as they generally used inferior drives.

[John_L]

I would be checking/eliminating.

Power supply associated with analog input channels, check voltages and stability with scope.
ALL individual analog input  video and audio channels. These are obviously going to individual ADC inputs.
Check analog signal path with a scope from input jack to ADC input on the chip. You have multiple composite inputs, compare them.

If all of above is eliminated then you can reasonably assume that it is analog processing chip with ADC input channels fried.

Dave, a bit of constructive criticism. You obviously don't have a lot of time, but try to organize yourself better. In number of your fault finding videos you are jumping all over the place with too many assumptions without using systematic approach and measurement tools. Also, if you want these type of presentations to be credible then you should make all reasonable effort to bring them to proper conclusion.

Keep up the good work.

[tri-be]

How many cans of spray did you use? :-)

You were giving the 24.576 MHz crystal next to the chip a pretty good spraying. It might be slightly involved in correctly interpreting the colour signal from the composite video.

Composite video is pretty lame and you're spoiled by digital video. Perhaps you're picking up interference of some kind from your lab/office. Those L+R+Video cables can be pretty grotty, and you have the cover off the back of the set.

The chip is probably supposed to do various picture improvements under some circumstances. Maybe it is something to do with that. Hopefully you can turn it off with a remote control.

The first DVD segment, before you started with the freezer spray in the video, there seems to be to be a barely visible horizontal band that slowly drifts down the TV screen, made a bit more obvious by your edits to the video. This could be just an effect of the different scan rates of the DVD output and your video camera, but could also be interference from another video signal, or other RF interference. It seemed like it got bluer in the top area while you activated the sprayage, but that could have been the camera, or some artifact of the video compression process.

The second DVD segment, where you get the colour splotches, the various cables around the TV had moved a lot. The VGA to composite box has gone (the composite cable was pretty tangled up with this previously), so perhaps you're no longer getting video interference from the VGA cable or converter output. So perhaps you have a cleaner video signal now, and the chip processing is doing some crappy "improvements" to the slightly different picture.

Composite video is lame anyway, and you'll probably never want to use it. Don't worry about it.

A pity about the component video though.

[creakndale]

The intermittent input signals are reminiscent of cracked solder connections either around the input connectors or around the tuner shell.  I've seen circuit designs relying on the connector shells and the outside of the RCA jacks to jumper "common" (ground reference) from jack to jack.  The outside of the RCA jacks are also part of the connector's strain relief.

You'll also see the tuner shell is probably being used as a circuit common path.  I believe they do this to save trace space by just using the metal part of the connector(s) and tuner shell as part of a daisy-chained common circuit path.

The failure modes are:
1) Mechanical stress due to the operator plugging and unplugging A/V connections into the the TV.
2) Mechanical stress due to cables that put weight or stress on the connectors.
3) Thermal stress from the TV on-off cycles.

Any of these can cause cracked solder joints resulting in an intermittent loss, or high resistance in the common (ground) circuit.  The fix is to remove the board, carefully inspect all the thru hole solder connections paying special attention to the input jacks and tuner section.  These cracks can be very difficult to see and require the use of a magnifier and good lighting.  If a cracked joint can't be located, I just resolder all of those connections.

As Dave plugs and unplugs connections and physically moves from the front to the back of the TV, the input jacks are being moved every so slightly.  With a cracked solder joint, that's all it takes to get this type of intermittent to occur.

creakndale

[Fezder]

first time i see ''cold'' spray in action, so that's where it's used, interesting....well i've heard of its usage, but newer saw .

[tec5c]

first time i see ''cold'' spray in action, so that's where it's used, interesting....well i've heard of its usage, but newer saw .

That's a video I made about a scope of mine that had a faulty hybrid IC. Freezer spray was used to confirm the fault the IC is not made any more and the only ones you can find are quite expensive. So I wanted to make sure it was actually that IC before I bought an IC worth over $100.


Dave, the discolouration, as tom66 said, would most likely be maldischarge. Maldischarge is misfired pixels, which is a fault with the panel itself. It can sometimes be 'tweaked' out by adjusting the panel voltages. Ie Vs, Va etc. There are usually test points and trimpots located around the SMPS board, monitor these with your DMM and adjust them slowly to see if you can introduce and more importantly, reduce the discharge. I can't remember exactly which panel voltage is usually the main influence but I know it's either Va or Vs.

Jason

[AndyC_772]

Maldischarge is misfired pixels, which is a fault with the panel itself. It can sometimes be 'tweaked' out by adjusting the panel voltages.

That wouldn't explain why it works on HDMI just fine but only shows discolouration on the composite input.

I'd guess that condensation from the freezer spray has affected the crystal oscillator circuit, causing the frequency to shift slightly. This would affect its ability to correctly decode the colour information from the composite signal.

Let it warm up and dry out properly and it should be fine.

[Sionyn]

the unpopulated space is for smart card cam for watch pay tv services like the nds system

[Fezder]

tec5c, nice video, but could you explain little bit, what was that fault that ic caused? was it that fading of the signal or what? i could'nt realize it .

[tom66]

Maldischarge is misfired pixels, which is a fault with the panel itself. It can sometimes be 'tweaked' out by adjusting the panel voltages.

That wouldn't explain why it works on HDMI just fine but only shows discolouration on the composite input.

I'd guess that condensation from the freezer spray has affected the crystal oscillator circuit, causing the frequency to shift slightly. This would affect its ability to correctly decode the colour information from the composite signal.

Let it warm up and dry out properly and it should be fine.

Although I do not think it is the problem, it is wrong to say switching the input will not affect maldischarge. HDMI and computer may be 60Hz whereas PAL video that the Aussies use is 50Hz. The switch in frame rate affects the priming wave. I am leaning more towards the main board myself though.

Maldischarge is a combination of panel ageing (MgO deposition) and poor control board firmware design. Good firmware adjusts for MgO deposition as the panel ages meaning maldischarge "never" happens. At least, that is true on Panasonic plasmas... Samsung and LG didn't get the memo apparently.

[tec5c]

tec5c, nice video, but could you explain little bit, what was that fault that ic caused? was it that fading of the signal or what? i could'nt realize it .

The quality of the video isn't the best I'm sorry.
As the IC heats up, the display gets dragged off to the left of the screen. Even with the horizontal position knob on the scope turned all the way to the right (clockwise), you can barely see the edge of the display. The IC that is being sprayed with freeze, and which is/was faulty, is the horizontal output IC. The IC developed an internal leakage path pulling one of the driver outputs high to its respective voltage rail. (Not captured on film - while monitoring the voltage at pin 2 of the IC [location - U800] you can observe the voltage start to increase as the IC heats up, finishing up at +5v when the display has been pulled off hard to the left of the display.)
Spraying instant freeze spray onto the suspected IC once the fault has occurred brings the voltage back down to 0v and the display returns back to it's normal position. This is what's demonstrated in the video.
I hope that makes things a little clearer for you 

[Fezder]

Ah, thanks tec5c, that helped to figure .