Toshiba convergence (1 Viewer)

[Scot Kight]

Has anyone come up with a way to reset the convergence ic's to stock yet?
If not, can I use CMUT set to some number other than 47 to bypass the convergence controlls and reset the convergence back to the unadjusted values? Then with cmut set to 48 (I believe) which should completely shut off the convergence, try to align the tv up? once thats done, just do the nitpicky stuff with the convergence adjustments?
What I am guessing is that when I set it to 48, it should be a minor difference than when set to 47 (48 = all color convergence ic disabled, other numbers have other combos of colors disabled, 47 = ics enabled). In my case the convergence is inches off from stock, and the 3 colors are inches away from their converged counterparts... is this proper?
Thanks.
Let me know...
Scot

 

[Chris White]

You can reset the convergence to the factory defaults. You will find instructions here: Reset Convergence
------------------
The Whites' Home Theater Website
Pictures, DIY Projects, Glossaries, FAQ, Links, Toshiba Tips & Tweaks

 

[Scot Kight]

Not looking for user level convergence, I am looking for service level convergence.
I should have been a bit more specific in there.

 

[Sean_S]

Scot,
I've been looking for a way to do this for about a year now. There should be a way w/o having to reset the QA01 or QA02 eeproms. It would be very helpful for centering the CRT's mechanically so that minimal stress would be put on the DC board. I suspect that convergence would also tend to drift less if we could accomplish this type of centering.
However, it may be the case that when you put up the white cross on black background internal test pattern that the DC circuits are being bypassed. If this is true, if you center the CRT's while viewing the white cross on black background the CRT's would be centered on a valid reference point, and you could then use the convergence controls to move the lines into congruity w/o worrying about overstressing the DC board. Obviously, you don't want to try to center the CRT's when the image's location on the tube face is influenced the the DC settings. The only reason I mention this is b/c it is the method mentioned in the service manual. Anyway...
I have centered my CRT's 3 times using the white cross, and my red and blue guns still have convergence anomalies in the upper lefts and rights. It's really frustrating. And playing w/ the CPC magnets....tough to get both the 3 CRT's centered and the astigmatism correct.... If anyone could provide thorough step-by-steps on this type of operation it would be great. I've read extensively on Keohi's site and Chris White's, but I think I'm mostly lacking in experience. Pix of the CPC magnets would help. Maybe 3 more attempts and I'll GET whatever I'm missing, b/c I've been told several times that mechanically centering my CRT's should correct my problem.
I'm assuming, Scot, that you have convergence anomalies too.
Sean

 

[StephenMSmith]

Hey Scot. There is still no way to reset the convergence IC memories, but you can use CMUT to find the true center for each converge memory. I have done this and found some benefit, even though I had tweaked my geometry to the absolute max prior to the CMUT realization (thanks to Louis Carliner for that one...).
These are the steps:

1. enter designer menu w/the full convergence grid displayed
2. verify static convergence params (STC's) are zero'd out, ie. set to 80
3. set CMUT=48. This bypasses the horizontal winding coils on the CRT's (not the convergence IC's altogether!), showing you the real CRT centerpoints, before the convergence IC affects are factored in
4. set STRH=7B and VLD=44. Adjust STRP to value that gets green grid closest to screen center (ideally 95, but +/- 2 or so should be OK)
5. use the centering magnets to set the grid centerpoints of all 3 CRT grids to exact center of screen
6. reset CMUT back to 47
7. recenter and readjust all the grids in the service convergence (using an overlay ideally)[/list=a]
   It could be quite a bit of work, depending on how closely your CRT's centerpoints are when CMUT=47, but if your a geometry nut like me, it's work it.
   Steve

 

[Sean_S]

Although it wasn't directed to me, thanks for the post Stephen! That's great info.
I'm now gonna try to tackle my upper corner geometry errors. Is there a param like CMUT which will diable the vertical convergence coils? Or is this a stupid question? The reason I ask is that I feel that my CRT's are mostly off-center vertically.
Also, any advice as far as the astig magnets vs the centering magnets? They seeem to interact, or am I adjusting the wrong ones?
Thanks in advance for any response b/c usually I'm not able to get back to my pc for a few days.
Sean

 

[Scot Kight]

Stephen Smith! You are the man!
I would estimate the work to be about 5-6 hours of effort. So if anyone else even thinks about doing this, watch out!
I havent done it yet, so that time effort maybe LOW.
Remember you will have to do the convergence in ALL of the modes, and setup ALL of the geometry again. Not quick or easy. But it should be the best way to go.
Scot

 

[StephenMSmith]

Sean,
You can fix everything there is using just the service convergence. if you align the service convergence grid lines to the service manual specs (this is where a screen overlay really, really helps), then you also
correct any and all geometry problems. You should use the geometry service menu parameters to get the size, position, linearity, etc. as close as you can, then finish by adjusting and/or moving the grid lines directly. If you don't have an overlay, you can use a program like CorelDraw or something to print precisely-sized boxes on a sheet of paper that you can hold against the screen.
I have a write-up on the astig/alignment magnet adjustment procedure at Errol's site that you can read. You would do this adjustment b/t steps 6 and 7 above.
Altogether this is at least a 5 or 6 hour job as Scot says, and maybe even longer depending on how many TW modes you actually use (I don't bother w/TW3 or 480p 4x3).
Steve

 

[Robert Packman]

What is the site link...thanks

 

[Sean_S]

Thanks, Steve!
I would like your opinion on my situation -- if you would be so kind.

I realigned the centering on my CRT's again last night. I know they have to be centered since I used the CMUT param to make sure the green was centered. Using the overlays I adjusted the VLD, STRP, etc. params and the geometry params (they were almost dead-on as far as making the vertical lines straight and evenly spaced) using green only. Then converged all CRTS.
However, I still have anomalies in the corners: in convergence mode the upper corners of the top 3 horizontal lines have curvature that I can't correct with the geometry params or by moving the convergence points. The vertical lines are straight as arrows. Should I whack out the geometry settings and bend the points back using convergence controls to see what happens, or am I stuck with this? My geometry settings are currently at or near defaults (para, cnr, trap, etc.). The curvature is most evident in the red and blue CRT's. That is why I thought my mechanical alignment was off -- I thought that the image wasn't centered on the red and blue CRT's and was pushing the limits of the DC at the corners....
Since the distortion is more intense in the red and blue, it makes the convergence look off in the corners even though the convergence points line up precisely. I wish I had pix. The curvature kinda looks like a ~ in the corners on red (it would look like and upsidedown ~ for blue). And the green is straight, so if there is text in that area it looks all out of focus.
AARRGGHH....this is frustrating.
Sean

 

[StephenMSmith]

Sean,
Unfortunately, those uncorrectable corner anomalies are normal for the red and blue CRT's and are most prevalent on the furthest side of the screen, ie. red in the right corners and blue in the left. I have the same thing even after spending 100 hours on my set's geometry. On my set, I would describe it as a subtle "bump" near the edge of the far corners of the red and blue CRT's. But I can really only spot it's picture affect by staring at a scrolling stock ticker.
By completing the static centering/alignment as you've done, I think you can conclude that you've got it as good as it can get. My geometry params are all at the EEPROM default values as well (I did all adjustment using the service convergence only). You can try playing w/the STRH param with the full mode service convergence displayed. This will fix either red or blue but not both! 7B appears to be the best compromise b/t the two.
The only thing left would be increasing the overscan, but for me at least, I'll take the subtle corner geometry anomalies in return for < 3% overscan on the sides (Full) because it is *very* hard to spot any ill effects on regular program material (for my set at least). Can you really *see* these anomalies outside of the service convergence or the scrolling CNBC stock ticker? If so, then maybe you can improve some more.
What are your overscan %'s?
Steve
------------------

 

[Sean_S]

Steve,
I set all modes except 1080i to the overlays, so my overscan should be around 4-5% for all modes other than 1080i. As for 1080i I have not measured overscan. I would say it's in the 2-3% range as the edge of the raster is just beyond the screen borders.
The convergence problems are most evident at 1080i since I have decreased the overscan as much as possible. The anomaly is almost non-existent in all other scanning rates and stretch modes. Increasing the overscan helps by hiding the anomalies, but I won't trade screen real estate for losing the anomalies. It's not TERRIBLE, but it's frustrating. You see, I almost exclusively watch in 1080i through my Dish6000 -- I don't mind the upconverted SD channels b/c it's worth it to be able to flip through channels from HD to SD seamlessly.
Anyway.... I have played with the STRH equivalent for 1080i which is STH1 and have it set to the best setting already (but I don't have the value right now since I'm at work). You're right though, this is the only param I could find that had a positive effect -- I chose a setting which minimizes the effect for red and blue both as much as possible; although, if you adjust it enough in either direction you can correct red or blue (but not both) almost completely.
The times when it is most noticeable is during OAR presentations -- the black bars have little bits of red, green, and blue at the upper corners. During normal viewing it is not noticeable, as you describe yours. It requires a high contrast area w/ straight lines to notice.
Well, I suppose I'll live with it. But I may try adjusting the STH1 param to correct the problem in red. Maybe since blue is the least noticeable color it won't look as bad. Any other suggestions will be appreciated.
Thanks,
Sean

 

[Robert Packman]

A strange phenomenon . I have a Toshiba 65X81 hooked up to dish for HDTV and DIRECTV for all else..HDTV is hooked on colorstream 2..SHOWTIME when leterboxed seems to look somewhat curved(like the old scope or cinerama) not to that degree but it makes me think of that..ALL OTHER leterboxing be it DVD's Laser, regular HBO and shotime comes in without this..I also noticed this when the old DISH demo used to have clips from showtime (ronin)..Can anyone explain..thanks

 

[StephenMSmith]

Hey Sean. Well you've definitely got all pieces that you can work with figured out.
It's just a matter of how much time you want to put it to it. I did some experimenting w/moving the raster area around to see if I could avoid the problem areas, but eventually gave up. It's a lot of work!
But every now and then, I still get the urge to fine-tune the all the CRT magents and redo my geometry for absolutely no reason. I suspect you probably do the same .
Steve
------------------

 

[RudolphT]

I do overscan in 4 values from seeing the edge. I too will go for the corner anomoliesin change for the content.
Did another service convergence (necessary after a WID or HIT change).
Cannot find STH1 in the service menu scrolling through the values. Any help?
Tim

 

[StephenMSmith]

STH1 is located in the geometry section of the designer menu (around the middle). However, if you have one of the latest H80 series, then the designer menu is not accessible...
You can download this spreadsheet for a complete parameter list and reference: Link Removed
Steve
------------------

 

[Jeff D.]

Interesting discussion guys. I thought I would chime in as I have noticed a bizarre effect on my 56H80.
It is usually noticeable when watching credits roll, but can also be seen on the odd horizontal or diagonal pan. There is a geometry error on my set, about 1/3 to half way up from the bottom of the screen. Hard to describe, but there is definitely something wrong, as the text during the credits sort of "shifts".
In any event, I noticed this on a 50HX70 at work as well. So I did some investigating. I measured the height of each of the rows in Service Mode convergence. It turns out that, right in the area I noticed this problem, there are two rows that are slightly shorter than the others. Verifying with a credit roll, I could now see what was happening. The text "shrinks" in height right over that point.
Now, the LIN and VSC adjustments seem to only affect the grid at the top and bottom extremes - they don't affect this area of the screen.
My question is - is there any way to correct this? I tried maniuplating things with the convergence mode, but I can't really achieve what I want to. Adjustments here only serve to sort of move the whole row down, as opposed to moving one line to make the row taller.
I hope I am clear in my description and I hope someone might be able to help with this.
____
Jeff
------------------
"You're a walking stick! It's funny! Now GO!"

 

[StephenMSmith]

If I understood correctly, then you should be able to fix that using the service convergence itself. Just move the problem grid line down a few clicks for all 3 CRT's.
Steve
------------------

 

[Sean_S]

Jeff,
Ditto what Steve said. I can think of no other way to correct that. I had a similar problem about 5" above the screen center. Looked like the credits were 'rolling over a bump'. I was able to greatly reduce it by moving convergence points only.
Sean

 

[Jeff D.]

Just thought I'd pass on an update...
I finished a full geometry and convergence adjustment using the Toshiba PSOL5601 templates for the 56H80. The results are amazing. Firstly, my little geometry error is 100% corrected - credit rolls are absolutely rock solid.
Geometry on all picture modes is much improved. For whatever reason, the geometry was off quite a bit from the factory. This was noticeable in all modes, but most obvious in TheaterWide 1. Corrected with the template, TheaterWide 1 is now by far the best mode for viewing 4:3 material. The main centre image area remains undistorted, with only the sides being stretched. This is how it should have been in the first place.
____
Jeff
------------------
"You're a walking stick! It's funny! Now GO!"