BVM cap replacement - corrosion bad, advice wanted

I picked up a (fairly) cheap Sony BVM-2015S that was made in 1992. While I'm sure it has a lot of hours on it, it looks like the picture tube has been replaced at least once, which it seems Sony used to do every 30,000 hours of operation way-back-when. The picture quality is pretty good, and I'd like to keep this thing alive.

After turning it on and smelling electrolyte, I opened up the unit and found a series of five mid-size capacitors that had leaked fluid all over what I believe is the main power distribution PCB.



It's the two yellow ones, the two in between, and the similar-sized one on the left. By the time I took this photo, I had already cleaned off much of the fluid from the board.

Eventually, I was able to remove the caps, and it looked like this:



A lot of the green coating is peeling away, but at least the copper traces are still intact.

What are not doing so well are the solder pads. I suspect that the very first cap to go was the top-yellow one, and the solder pads under it appear to be dissolved away.

A lot of the pads that were under the other caps are in poor condition as well. Not to mention, the resistor and the diode at the bottom of the picture don't look so hot, either.

Then there's the black thing in the top left corner. This is actually two components back-to-back, and I think they are transistors. One way or another, they seemed to suffer the worst corrosion, and on the bottom of the board, you can see this. Look at the two groups of three pins:



Even at 500 degrees Celsius, my soldering iron couldn't get this to melt again. Flux, alcohol, vinegar and baking soda don't seem to do much to it. I'm assuming that the pins are either shorted, or they're out of circuit entirely.

-------------------

So, what should I do about all of this?

Tell me if my idea sounds stupid.

This board goes inside of a big metal case, and there is a lot of room to spare. For an amateur like me, an easy-sounding solution is to put all the new caps and replacement components on a small board like this. Then, I could run wires from the leads of these components to the next solder pad on the corresponding trace where the old components were on the main PCB.

Is this dumb? Is there a reason why it wouldn't work electronically?

-----------------------

A few other questions:

1. When I replace the caps, is ripple-current important? The caps have voltage, temperature and uF ratings on them, but not ripple current, and that sounds like an important factor in capacitors. Do you have any other advice about selecting new caps?

2. Bizarrely, the leaky caps still give good uF ratings with a multimeter. Their resistance is a little screwy, but aside from buying an ESR meter, is there anything I can do to diagnose bad caps?

3. If my daughter-card idea isn't dumb, what kind of wire should I use?

4. Does anyone know what those small black things are in the first and second photo? I mean the things with the white dots on top. They just say SONY with some number that doesn't turn up anything on Google. They don't seem corroded.

5. Assuming that those are transistors, what should I be careful about when replacing them?

---------------

Thank you!

OK, I'm going for the daughter-card idea. All of the parts are in.

EDIT: Never mind, I think I got the answer to my next question. Those back-to-back components are fast-recovery-diodes.

On the off chance anyone else benefits from this, I found the full part number in a service manual: ESAC25-04N. They are not transistors.

A spec sheet can be found here:
http://http//www.eicsemi.com/datasheet/ESAC25.PDF

I think that this is an acceptable replacement:
http://www.marutsu.co.jp/pc/i/49673/

I'm going to check later to see whether the original parts, which I clipped off the board, are useable.

By the way, I was right about one thing: the heavy corrosion caused those six points on the main PCB to short each other out.

Can you believe that in spite of five blown capacitors and a shorted-out diode complex, the TV both functioned and had a nice picture anyway?

Hey I know you posted this earlier today, but I am curious to know how your daughterboard idea turned out. Looks like it's a miracle that the unit was able to function properly for long.

I'll definitely keep you posted.

It looks like I can't quite get the daughterboard to fit inside the metal case, so I've decided on a different approach: put it outside of the case (there is definitely space) and rig up a connector so that I can easily remove the card and open the case back up if/when needed.

I replaced all of the other large caps on the board today, and even though they did not appear to have trouble-symptoms while they were attached, it was easy to see minor leakage and bulging on the undersides once I got a good view.

A guy on the shmups.com forums who seems to have had experience working on BVMs in a professional capacity said that it's the capacitors on this power-supply board that are the cause of most people's troubles. Seriously, to anyone who has a BVM, I'd get those caps replaced sooner rather than later. If they've never been replaced before, I guarantee that some of them are already on their way out. I have to say, replacing caps before a major leak and corrosion event is about 100x easier than doing it afterward.

The latest thing I'm wondering is this: Do great big shot-glass-sized caps need replacing, too? Those are definitely not cheap. At the same time I'm doing this, I'm also repairing an MSX, and there's a similar-sized super-cap in there as well. I don't mean to be a cheapskate, but if they don't need replacing, I'd just as soon save the $20 it would take to buy and ship three new ones to me.

Are those giant caps known for going bad as well?

Replacing any caps that you can before they become a problem is ALWAYS a good idea, especially in any device with surface mounted electrolytic capacitors (Pioneer LaserActive pac units, etc.). The tricky part is that caps that are on their way out do not always have an obvious sign such as bulging. With that said, I have not heard of the larger caps failing on these units, but no part lasts forever and these machines are only getting older by the day so a swap is not a bad idea if an identical replacement can be found.

Sigh.

Lord knows I tried to save this thing, but I couldn't. My repaired board went in, the switch went on, and with a sad whine and a plume of smoke, it was over.

In case you're wondering, I had abandoned the daughterboard idea, for the most part. Excluding those two fast recovery diodes, it turned out that the solder pads on the bottom of the board were in recoverable condition after all. I put in all the new caps where the old ones used to be and just used wire connected directly to their legs to get around damaged traces on the top of the board. The fast recovery diodes went on a tiny daughterboard the size of a postage stamp that fit easily inside the case.

Anyway, it obviously didn't work. I don't think I screwed up on any of the caps themselves, but it's possible that I did something dumb with the wires, or that I damaged some other part along the way. Alternatively, it could be that the board was so close to death that being washed and soldered-to pushed it over the edge. I don't know...and that's particularly frustrating. Whatever burned and caused the smoke doesn't seem to have been on this board, either.

Are the potentiometers on these things highly vulnerable to water? I've read that potentiometers in general shouldn't get wet. I tried to keep them dry, but it occurs to me that one of them could have caught a drip.

I've successfully replaced caps on other boards before, and I am quite confident that if I had gotten this BVM before any of the caps had actually leaked, I could have easily gotten it to work. Whether it's because I did something dumb or because the board had one foot in the grave already, it was ultimately the corrosion that killed this thing.

Or so I will choose to tell myself. I really hope I didn't do anything that boneheaded with my repair...

Swap your caps early, folks!

Fixed it!

There were a couple of corroded pads that were shorting! The TV just ran for 30 minutes with no trouble! Thank freaking god!

Just in case someone finds this informative, let me post a little more.

The BVM is basically working. I've logged about 10 hours of gaming time on it already.

However, it exhibits an odd behavior, which it also did before my repair of the power-block PCB. If it draws a white horizontal line followed by a dark(er) horizontal line, the darker line becomes wider. In fact, a pure white line can affect two or three lines beneath it, depending on the circumstances, resulting in what looks like a bulge at the edge of the image.

By checking with patterns that do not have a high contrast, I determined that this is probably not a geometry or convergence problem. It's strictly related to the beam exceeding a certain level of brightness.

My theory is that there is probably a high-voltage capacitor somewhere that is failing, or perhaps many of them. I opened up the set and had a look at the EA board, which is described in the service manual as "Horizontal and Vertical Deflection Output and High Voltage Regulator Circuit."

Sure enough, this board has many large capacitors on it, and guess what: most of them are bulging. Replacements are currently in the mail. I'll post an update when I get them soldered in. Since no electrolyte leaked, it should be an easy swap.

After that, there is one more thing that stands between me and a perfect picture, and that's a slightly tilted image. I'm going to have to try rotating the deflection yoke a little, and I'm not going to lie: I find this somewhat terrifying. Any words of wisdom?

Rats. Even after replacing all of the capacitors on both the EA board and the BK board - the latter of which is the final RGB amp - the problem remains.

Replacing the EA board caps did fix something, though. I was playing Popful Mail on Sega CD, which is naturally 320x240, and it has windowed cutscenes surrounded by black. For some reason, the first few lines of the window used to skew to the right. This didn't happen with full-screen images on the BVM, and the cutscenes are perfectly rectangular on my other TV. Thankfully, with the cap-swap, this issue is gone. Also, the overload lamp lights up less frequently, which is almost certainly a good thing.

Today I bought a special driver to help me loosen the clamp on the yoke. It is possible that some of my distortion-woes come from trying to correct the tilted image with other settings. I'm also going to play with the purity rings to see if I can line up the convergence a tiny bit better.

I'll try to take a picture of the effect I'm seeing. If anyone has any idea what could cause the lines beneath a white line to fatten at the edges, please do let me know.

Finally, if anyone ever wants the manual for a BVM-2015p, which also covers my 2015s, just let me know. I bought it recently, and I'd be glad to share it.

EDIT: I fixed the tilted picture! It was a piece of cake! No time to test the purity rings, but that will come soon enough.