Easy SNES/SFC YUV component video!

Quote:

---

Originally Posted by Drakon

Fine then continue breaking fragile chips. You're the last person who should be giving advice. You seem to have never built anything at all. You have zero experience and don't understand a simple concept like using glue to hold fragile parts together.

---

Or, you could just be careful. There's nothing wrong with using a little glue here and there, but you don't need to cover the thing in glue like you said.

And I do know how to solder and have done it many times. The difference is I don't own a soldering iron personally. All the soldering I've done was in a school or lab environment. I know things can be fragile and break easily. The point I'm trying to make is, a little glue or other kind of adhesive can go a long way. There's never a need to go overboard like your previous mods have shown and smother everything in hot glue. A few small dabs here and there would get the same result.

From experience in building models when I was a kid I can tell you that when it comes to glue, the better mentality to have is less is more. A small amount of glue will hold something down just as well as covering it in glue will. The difference is the former approach looks much cleaner in the end and is easier to work around.

These chips are being built into a project box and from what I've read they break if you breath on them too hard. Who cares if they're encased in glue atleast it makes it so they won't break. You don't have to look at the chip it's inside of a case. My idea is to use a dip adapter and have the big dip solder spots exposed so you can still solder all the pins freely. If you have a better suggestion to stop these chips from breaking please say it.

:sweat:

Wow... total fail. It turns out my problems were caused by me resoldering Composite Sync to the wrong pin of the BA6592F. That's the 3rd time I make this same mistake.

Quote:

---

Originally Posted by Drakon

These chips are being built into a project box and from what I've read they break if you breath on them too hard. Who cares if they're encased in glue atleast it makes it so they won't break. You don't have to look at the chip it's inside of a case. My idea is to use a dip adapter and have the big dip solder spots exposed so you can still solder all the pins freely. If you have a better suggestion to stop these chips from breaking please say it.

---

Having all that glue around it can make it hard to work with as well. What if you need to get to it and redo some of your soldering? Oh well you can't because the chip is now under a pile of glue, and trying to break that glue off could very well damage the chip you are trying to protect with all that glue.

It's not all about looks, it's also about use and being practical. If you are concerned about the chip moving why not just put a few small dabs of glue to hold it down? That accomplishes the same goal but now it's not completely covered in glue and you can still get to the pins if you need to redo any soldering as well as do more soldering to other pins if necessary.

As for a suggestion, since this is clearly an issue of operator error, I'd say a good suggestion would be to be more careful and pay closer attention next time.

Quote:

---

Originally Posted by Ace

:sweat:

Wow... total fail. It turns out my problems were caused by me resoldering Composite Sync to the wrong pin of the BA6592F. That's the 3rd time I make this same mistake.

---

So did that damage the chip or is it still usable?

SMT ICs don't move when all the legs are soldered onto a surface socket (not sure what it's called). I've done such a thing before and no glue is required once the chip is fully soldered in place. And when I say all the legs, I do mean all of them. Lifting legs is extremely foolish and puts the chip at serious risk. I've found SMT chips tend to be sensitive to heat, so less is better. Going by what Ace has said before, the BA6592F is especially sensitive to leg lifting, so just solder the entire chip into a sturdy surface, test for cold joints (it does happen with SMT chips) and then you should be good to go.

Trekkies is right, you don't have to smother everything in glue. Good solid joints don't often need to be reinforced so heavily in glue. If you solder a wire and it has a very good bond, the wire itself will break, not the joint... For practical applications, yes a bit of hot glue can be useful, like for panel mounting jacks/switches without screws, but the whole thing doesn't have to be covered in glue. That just complicates repairs even further, if it ever has to be done. Glue won't fix a cold joint, it will only mask the problem.

I think wiring Composite Sync to the wrong pin has killed the chip. The desyncing issues are back, resoldering Composite Sync doesn't help at all and the chip is getting unusually hot. This is what happened to the first BA6592F I used as it too had Composite Sync wired up to the wrong pin. God, these ROHM chips are so fragile, it's not funny.

EDIT: Oh that's real nice. My Luminance transistor has taken a shit. It doesn't work at all.

Chip too hot = shorted, possibly internally. You need a new BA6592F, sorry man. :( I may at some point try my hand at these chips, but I'm still waiting for my CXA2075s to play with.

Not sure if this is useful info, but a quick way to extract a broken SMT chip is to very slowly and carefully, clip each leg and desolder/remove it, one by one. Trying to clip all the legs at once might cause problems, since most SMT chips have very little gap between each leg. Also, in my experience, SMT solder pads are very prone to lifting if you accidentally pull on them. Just be very careful with the removal and prepping and things should be good (I mean prepping by clearing pads of extra solder and making them smooth).

Huh... well this is interesting. If I just directly wire Luminance to the TV and bypass the transistor I have amplifying the signal, the picture is a lot blurrier than normal. One thing's for sure, I need a new transistor to amplify Luminance as the one I have appears to be completely shot.

I haven't been keeping up with this thread much, but is Ace the only one actively researching this chip? Is someone else here doing tests too?

Quote:

---

Originally Posted by Guntz

I haven't been keeping up with this thread much, but is Ace the only one actively researching this chip? Is someone else here doing tests too?

---

I've been doing tests here and there, but not in the capacity I'd like, or anywhere near the capacity that Ace is. I've been laid off since mid June, so a lot of my time has gone to job hunting and what not. Now that I actually have a job that starts next week, things should get back to normal for me and my time management.

What I can say is that the results on my two Genesis units, Master System, and Neo Geo all look pretty damn good. My HDTV actually accepts a 240p signal, so I've been using that to test color balance, since it's been professionally calibrated. I have to say, the colors look pretty spot on. I'll have a go at it again later on to see if I notice any changes.

I've also noted that the BA6592 chips are temperamental, so whenever constructing a circuit for them, I have it protoboarded first, so I don't waste all that time soldering the circuit together, only to have it not work. Also, I second Guntz in soldering ALL the pins to a SOP to DIP adapter - if you do it right, there's no need for glue or anything like that to hold it in place. Also I agree that lifting pins on the damn thing is a bad idea.

I'm done with the BA6592F. I'VE HAD ENOUGH OF THIS GOD-FORSAKEN CHIP!!! Every single one of them ended up like this:

-Complete failure of Luminance
-Complete failure of either B-Y or R-Y

I can't deal with this chip anymore! I'm gonna try something else: the BA7230LS. I'm gonna buy some and rig up a converter using this chip. At least it's not a surface-mounted piece of shit!

You say it took you a month or so for chip failure? How often were you using it?

Quote:

---

Originally Posted by Lastcallhall

I've been doing tests here and there, but not in the capacity I'd like, or anywhere near the capacity that Ace is. I've been laid off since mid June, so a lot of my time has gone to job hunting and what not. Now that I actually have a job that starts next week, things should get back to normal for me and my time management.

What I can say is that the results on my two Genesis units, Master System, and Neo Geo all look pretty damn good. My HDTV actually accepts a 240p signal, so I've been using that to test color balance, since it's been professionally calibrated. I have to say, the colors look pretty spot on. I'll have a go at it again later on to see if I notice any changes.

---

Have you posted before what your current application circuit is composed of? Perhaps it'd be of help to Ace if he doesn't know the whole picture (no pun intended). Maybe a photo of your best working encoder circuit would assist Ace.

Quote:

---

Also I agree that lifting pins on the damn thing is a bad idea.

---

Technically speaking, lifting pins on chips is not an intended usage of the chip, both SOP and DIP. It can be done on DIPs because the pins and chip are much larger and far more durable, but doing it on SMT/SOP chips is a very, very bad idea. They can snap very easily, even with an adequate amount of heat. Even if you do manage to bend the pin, the amount of heat needed to do so, can damage the chip.

Quote:

---

Originally Posted by Ace

I'm done with the BA6592F. I'VE HAD ENOUGH OF THIS GOD-FORSAKEN CHIP!!! Every single one of them ended up like this:

-Complete failure of Luminance
-Complete failure of either B-Y or R-Y

I can't deal with this chip anymore! I'm gonna try something else: the BA7230LS. I'm gonna buy some and rig up a converter using this chip. At least it's not a surface-mounted piece of shit!

---

At the very least, just take a break. Take some time off from the chip and do something else. Come back to the BA6592F with a fresh mind and a more calm soldering procedure (though I don't know exactly how you work). Maybe you should work with some CXA2075 chips in the mean time, they don't seem to be as prone do failure (but can still break for little reason).

S-Video is not an option. I'm not going back to either that or Composite after seeing the kind of quality Component gives compared to those signals. No way.

In the meantime, I've decided to recap a VA3 Genesis Model 1 I was supposed to trade with a friend for his VA2 Genesis Model 1. He gave me the VA2 last week and I gave him my VA6 Genesis Model 1 until I finish recapping this VA3 Genesis Model 1.

Quote:

---

Originally Posted by Lastcallhall

You say it took you a month or so for chip failure? How often were you using it?

---

The box was being used nearly every day, especially recently with my two 240p Component test videos I posted on YouTube. If I wouldn't have gone and attempted to tweak the RGB inputs as per what Chessage mentioned (which I was gonna try first with fixed resistors due to not having any 500ohm POTs), I'm pretty sure the stupid box would still be working.

Another thing, MAKE DAMN SURE YOU'RE SOLDERING COMPOSITE SYNC TO THE CORRECT PIN!!! Since I've soldered Composite Sync to the Composite out, the BA6592F will no longer output Luminance, and now, it's no longer outputting red. Son of a bitch!

Quote:

---

Originally Posted by Ace

S-Video is not an option. I'm not going back to either that or Composite after seeing the kind of quality Component gives compared to those signals. No way.

---

I dunno about that... S-Video shares pretty much the same clarity as component, the only difference is color definition. Further more, S-Video is still very inexpensive to implement, and a decent percentage of TVs support S-Video (HDTVs, not so much). For a long time until now, component was prohibitively expensive, either in parts or labor. Only now have we realized the existence of this one chip.

On top of all that, the CXA2075 is a perfect replacement for the KA2195D which can be found in most Genesis Model 2s. A BA6592F is quite a bit more complicated to implement, since it can't be installed in the old socket pad without significant changes to the surrounding circuit (not worth the trouble and trace cutting). Changing out the 2195D for a 2075 is pretty simple once you know the differences. That means it takes less work to give more Model 2s a new lease on life. The only other thing needed would be a CCAM, then you have a very respectable Genesis. :)

For purists though it seems, only the VA4 of the Model 2s is any good. :?