[jeffescortlx]

Samsung HCM5525W.
Got it with a bad channel in the convergence chip for the Red CRT, no control side to side.
I allready went ahead and replaced the STK393-040's. Now I can manually move the RED in all directions.
But, the auto convergence can not finish and work because the Blue CRT coolant is so contaminated that the sensor's for the auto convergence cant see it.
I allready have the Blue CRT tore down and cleaned the crusty layer off. But before that, when I was trying to manually coverge the Red, it just did'nt seem right. Some sqew and twist in the crosses, and I had to max out the movment to get it to line up over the green.
So my question is, does it need to first complete a auto convergence to "ball park" convergence so that it's closer for the manual convergence?

 

[lcaillo]

If you never stored the data or initialized the auto convergence you may get it in the ballpark when you put it back together and run it again. If you stored any changes, you may need to start from scratch. This can be very complicated for a novice, particularly if you have moved the centering magnets.

 

[jeffescortlx]

The convergence menu I was in, was only the consumer menu. I did'nt go into the service menu or any thing.
I did'nt move any of the mechanical yoke adjustments (now that I think of it, I did'nt see the normal looking slider aduster dodads on the yoke) and the original owner said he did'nt touch any thing either. But he did say he took the cover off and looked inside.

I just hope the auto convergence can complete and put things back where it belongs.
As far as I can tell the convergence circuit is working like it should, I can move in all directions, and I ohm'ed out all the low impedance resisters about 5 times to make sure non where open or high.
I'll have the blue CRT cleaned up and back together maybe by tonight with fresh coolent.
You'll hear from me again if it's still a problem!
Thanks for the help.

 

[lcaillo]

Just do your best to line up the blue manually when you put it back together before you try the auto convergence.

 

[jeffescortlx]

whoohoo! It's working!

The auto convergence was able to work right once the nasty coolent was cleaned up. I had to do a tiny bit of mechanical focus on the blue since I had it all apart. The blue is almost perfect, as good as I can get the focus by just looking at the screen from the rear while it run's auto convergence.
The green has a tad bit of a halo effect, so I'm sure that coolant is going bad also.
But the picture is good enough for now. I'm just happy it's working. It's useable now at least, not the greatest picure I've ever seen, but it's getting there.
Thanks for the help!

 

[Krankshaft]

The auto convergence parameters are set from the REAL factory convergence screen (not the line up the tiny + user convergence).

It's when the TV generates a cross hatch and you must move certain intersecting points both left and right and from top to bottom to make nice evenly spaced straight lined rectangular boxes (start with the green gun and a convergence overlay and turn the red and blue off). After green is lined up with the overlay remove it and turn on the red or the blue and line them up to the green. I believe to store new touchfocus info on my Toshiba was press 7 from the convergence screen then hit touch focus.

Just FYI touch focus will get you in the ballpark but it's no substitute for a manual service mode convergence. The halo may be the red and blue not in perfect convergence with the green.

The best way to adjust focus IMO is to use a crosshatch pattern your TV can probably generate one through in service convergence mode turn off all of the guns but blue with your remote and adjust it's focus.

Great job!

 

[lcaillo]

Auto convergence is really just a reset to the last parameters set in the service menu. You adjust it, save it, then initialize the sensors so that it can return the limited number of parameters they measure to the last settings initialized.

Halos that are hazy are normally optical, offsets that are clearly focused are convergence problems.

 

[jeffescortlx]

This TV has 4 optical sensor's, middle of the top, bottom and one in the middle of each side. When it goes through the auto process it display's a line, one at a time over each of the sensors, and will do this 2-3 times for each sensor. So I would'nt think it would have to rely on pre-programmed settings when it can get it's own real time convergence. Or at least that's how I think it works...but I'm no TV tech!

But, I did do a manual convergence, just had to tidy it up a bit.

When I was talking about the halo for the green color, I was looking at just the green gun. This was during the auto convergence when the blue and red guns are off. So it could'nt be from a converenge issue. I took another closer look at the green-only color, the halo effect was'nt as bad as I first thought, and decided it was'nt worth my time to change the coolent at this time.

Now when I play a movie and step back 10 feet or so, the picture looks great! Has good bright crisp colors and a sharp picture.

I could'nt belive how bad the blue coolent was. After I first put in the STK's the blue picture looked as if some one smeared grease all over the lens.

 

[lcaillo]

This TV has 4 optical sensor's, middle of the top, bottom and one in the middle of each side. When it goes through the auto process it display's a line, one at a time over each of the sensors, and will do this 2-3 times for each sensor. So I would'nt think it would have to rely on pre-programmed settings when it can get it's own real time convergence. Or at least that's how I think it works...but I'm no TV tech!

The way it works is like this. You converge the set, save the data, then initialize the system. This makes the current location of the red, green & blue at each sensor the reference point. When you run the system it moves the image for each color to match the reference data stored when initialized. It does NOT do convergence. It simply matches the location that it detects to the previously stored data. In the case of the Samsung, with only four sensors, IIRC, it only does static convergence. Some others do some interpolation of some other parameters other than position, but they are still limited to what was previously stored. None of the consumer systems for "auto convergence" actuall measure convergence of the three images. They measure relative to a previously stored reference for each.

 

[jeffescortlx]

Why is it when the blue gun was so contaminated that the sensor's could'nt see it and it could'nt finish auto convergene, it was'nt able to align the gun's. Red and Blue were way off.
But once the sensor's could see the blue, it could then finish auto convergece and was able to align all the colors and "fix it's self".

If it just goes off of previously stored data, why does it even need any sensor's at all?

I'm not trying to argue, just learn how this stuff works.

 

[lcaillo]

When the sensors do not get enough light, or when they get so much ambient light that it cannot distinguish the test pattern, it does not know how much correction to apply.

The convergence data that is stored is used to create the correction waveforms and d.c. offset applied to the convergence amplifiers. As the amps and related parts age, and as the optical parts age (primarily the coolant in these sets) changes in what you get at the screen occur with the same input to the amplifiers. The purpose of the convergence correction is to get it back to where it was when the set was initialized. The measure the location and apply the bias needed to match the original data. If there was not some reference, they could choose one of the colors and try to match the others to it, but the systems are not precise enough for this to be a reliable method. Some attempts have been made to make professional grade products do this, with very mixed results, mostly poorer than manual adjustment.

 

[jeffescortlx]

It seem's as if the TV has the parts to make it's own convergence correction, it has the eye's to see it. I know it might not beale to make a perfect alignment with only 4 referance points on the outside edge.
I think by calling it "auto convergence" or "auto focus" is misleading. They should just call it "back to previous default settings" lol.

 

[lcaillo]

The problem is the precision of the system. Just look at the size of the patterns that they use. They are also limited to the outside edges, so even if the convergence was adjusted on the fly, it would not be very comprehensive. A system to do full screen convergence automatically would be very expensive and complex, and would still have to be initialized to some reference geometry.

I agree and have said for years that it is misleading, and really rather useless, considering the cost that it created.