[michaelt]

Within a week the left & right edges occasionally would tear (jitter) then settle down again after being on for a while. Then it went blank, the red power LED started flashing and the power button did nothing. After removing power for a few minutes the set remained off. The final time it lasted for 1.99 DVD movies (missed the final sequence), went into the LED flashy mode forever.

My google search for a clue finally got me here where I read the entire thread about convergence modules, failures and repair (very informative). Notably missing was Toshiba. After finding this forum, reading all the posts, and the rules I decided to join (register). I am looking for the root cause of this problem and maybe I can spread some of my knowledge in other areas later.

So, where am I now?

• the set is broken down. (This set was designed to be built, not repaired and go to the dump.)
• no obvious heat or physical damage, just a little fine dust (it's been cleaned every year).
• no coolant leakage.
• both of the main boards (power & signal) have had a close look-over with a bench microscope. Except for one heat sink tab, all the solder joints look "ok" (I can't say good because most are hazy, quite few have way too much solder and a lot of flux remains all over); all caps look good.
• the twelve "hot" resistors in the convergence driver paths all look new - no discolorations or cracks.
• hard to look directly at the 392-110's because the hyper module blocks a direct view but they look fine.
• all 10 fuses on the power board ohm out to zero.

Now, with (almost) everything out, my question is: what to suggest I do first?
From what I read so far, I could:

1. remove the convergence modules and power up. A quick check to see if they are the problem. (easiest)
2. replace all the hyper module electrolytic caps. Time delay to remove board, size, order, wait for delivery and replace. (medium)
3. start measuring & scoping out the boards. (hardest)

Your thoughts please.

 

[michaelt]

I have made a little progress.

I purchased the Service Manual from one of the recommended sites: completeservicemanuals.com.
I am very satisfied as these "pfd" files were not just pdf scanned images of the originals, they are the pdf originals.

Both the Power & Signal boards are now out.

All fuses are good. I double checked the Convergence +-16v & +-32v fuses.

I am going to replace the Convergence chips just because the Signal board is out and by removing them I can test if they are the cause of my failure (thanks for the tip).

The Hyper board caps I would like to replace, again because it is out and I do not want to repeat the removal any more times than I have to. One problem: I can not determine the physical size (footprint). Being an SMD device it is critical to get it right. Can you tell me the package type or give me a part number that I can look up the dimensions from (DigiKey or Allied would be great). The Toshiba part number just gives me a lot of sites for quotes or buy but no specs. I don't want to remove that board just to measure a cap.

A side note; this model does not have a removable Light Box as indicated in the Service manual. Any live testing must be done in the set. With very little cable slack, it is going to be difficult to say the least.

Is there an easy way to disable the HV to avoid another shocking experience? I didn't see anything obvious in the schematics. (One 40KV hit in a lifetime is enough.)

 

[michaelt]

Progress is slow but I do not want to miss anything and have to rip this apart again.

I have given up on trying to "bench" this puppy. It was not designed for it. Not worth the effort.

With the SMD caps as a major issue, I decided to get an ESR meter. It came in kit form and needed a few modifications to upgrade it. Works well. Worth the money.

Measured all the hyper board caps and found most to be within limits but aging. A little digging shows these caps are at their design lifetime: ~8 years. One failure (open), three grossly over limits (>*10), twenty over limits. The lifetime depends greatly on hours and even more on temperature. Since this board is enclosed in a metal box with minimal air flow inside a larger box with minimal air flow, ambient temperature must be raised by 10-20°C for lifetime calculations. It came out to be about 8 years and that is just how old it is. I ordered all of them from DigiKey.

While doing the cap test, I ordered the Hitachi convergence kit from Encompass early in Nov. They were on back order and expected in on 11/29. Now the delivery is "in 6 days" and holding. I also found these at Sears Parts Direct, in stock, delivery in one week, so they got the order. We'll see how they deliver.

I used the ESR meter to check all the other board caps and found none over limits. A bit surprised but maybe these run cooler since they are not in a "box". I have one more board (Digital Converter) in a box, blocking the convergence chips, so I'll remove it to check caps and make the convergence chip replacement easier.

More later.

 

[michaelt]

Finally back working on this since the Christmas interruptions. I set a limit of $200 to spend on repairing this tv. So far I've spent ~210. Most of it was ~$80 for the ESR Meter which I get to keep afterwards.

Most was ordered online:

• ESR meter kit from Anatek ($87) + meter lead sockets from RadioShack ($10 local store) allowed me to use the test leads from my Mastech DMM (much better than the banana-alligator leads in the kit).
• convergence IC's: Sears PartsDirect was a no-show. They listed part in stock & 10 day delivery but they have no idea since these come from "unknown outside suppliers". (Never had this problem with them before with parts for tools & tractors.)
  I went back to Encompass, the rep checked all the possible kits - the last one showed just one in stock - I bought it ($51), I got it. It had the STK394-160 IC's. I rate them as A+.
• SMD caps: got all of the Hyper board caps from DigiKey ($32). Be sure to check the physical sizes! I had to chose from three different series to get the "best". Price difference is minimal.
• thermal grease: ArcticSilver 1.2g (I had no way to tell if the .35g size would be enough) from NewEgg ($18 +$3 ship) - got it in 5 days even with it going to the wrong town! The .35g was $10 from RadioShack.
• misc: got a .031 tip ($3.50) for my 35W Weller iron, much better to get into the few tight places; Chem-Wik .050 desoldering braid ($14). (DigiKey)

Capacitor removal was done with two soldering irons, one for each terminal - heat & lift (flick?) off. The only problems were when on side was attached to a plane without any thermal reliefs. Using wick to clean the pads prevents the new caps from wandering. Flux new cap, heat on corner of the pad, solder to the other corner, it melts and flows over the lead, nothing moves. Solder the pin with less copper trace first - faster and it holds the cap in place especially when it is a plane.

I re-measured all 49 caps after removal to see how good the in-circuit test was. Of the total 43 bad caps, in-circuit the meter correctly failed 29 bad caps but it passed 14 bad ones & failed 3 good ones. The 14 bad ones passed are attributed to the low impedance of the circuit itself. The 3 false failures I don't have an explanation for other than a measurement error. Seven of the caps were open (measured infinity). How this set worked amazes me.

The convergence IC's were replaced by removing the Digi-Conv board to gain free access to the heat sink screws. Desoldering was done with the braid wick only. Nothing abnormal, no signs of heat damage.

A soldering gun was used with a sucker to release the board shield tabs, wick cleans up the reminder.

Reassembly went well except for the connectors. Most located easily. There are two open connectors: Digi-Conv module PH005, 6 pins and Signal board P113, 5 pin. It would have been good to take a picture or two before before tearing it apart.

Now the problem: initially it did not power up properly. Still had the blinking red light. After a few times plugging & unplugging into the wall, I heard the speakers pop and a relay(?) click. From the rear, there was a white flash near the middle (green) crt a few times. I left it plugged in for a minute or so and then it powered up with full video (un-converged). A quick auto focus concluded with an "Auto Focus Error" message. ????
After 30 minutes the picture would shake occasionally. Going to use a large fan to test if it is a heat problem next.

Was this "normal"?
Don't know but now it is fan, convergence & pray time. More later.

Thanx for the help. Any more suggestions welcome.

If anyone wants more detail - just ask.

 

[michaelt]

Well now it is the shotgun approach.

I am going to defer the full convergence until I root cause the current problem. It is only out about 1/4" at the far corners.

I brought out all seven power supplies so I can monitor them easily from the front to see if one of them is fluctuating when the picture shakes.

The initial power up brings me back to the delayed flashing red power light. Got to see if the "flash" is happening again around the green gun when power button is switched on. (It won't power off - got to pull the plug.)

I'll be back.

 

[michaelt]

The flash is now gone since I got the set to power up. All connections checked & re-checked OK.

When the picture shakes/jitters all the voltages except the +5 rise ~5% and jitter. Normally I would expect to see a drop not a rise.
Since the +5v is on its own transformer-rectifier-regulator, I now suspect something in the main DC power circuit driving T862 (Q801, et al).

Suggest any usual suspects?

 

[michaelt]

Based on other posts I went for the re-solder of the xray protection and primary dc voltage drive circuitry. A few connections were questionable but nothing clearly bad or poor.

Tried again last night. Power on was normal, all voltages nominal and stable. Noted the brightness was was higher than we normally have it. All was whitewashed, dropped the brightness level from 50 to 23. Watched a DVD for about 45 min then the +7v drooped, spiked up and went to zero. TV now dead. Looks like something failed, took out a major component and/or a fuse.

Too late to troubleshoot - went to bed.

Today is a new day to continue looking.