MaximRecoil

"Crude" as in, not a perfect, thin, even application of tin like you'd find on a "tin can" (which is tin-plated steel), or tin-plated QD terminals (tin-plated copper). Even DIY tinning jobs on copper cookware don't usually have mounds of excess tin all over the surface. It seems like an unnecessary step anyway, since the only reason to tin-plate electrical conductors is for corrosion-resistance, but the traces will be covered with solder mask anyway, which seals out oxygen and moisture, preventing corrosion. In any event, I have a higher opinion of the quality of Atari's boards now, though their pads are still relatively delicate.

Thanks for the picture (and see the edit to my post above). Yes, those are definitely solid mounds of tin, which means they used a crude method to tin the traces before applying solder mask, a practice which evidently went out of vogue for the most part at some point in the '80s. What I would like to know is how they avoided creating short-circuit bridges between the smaller / closer-together traces while crudely tinning them.

Can you post a picture of the board you sanded to bare copper (closeup of the part you sanded)? By the way, Nintendo arcade PCB traces are bare copper for sure: Edit: If they used a crude method to tin the copper traces prior to applying solder mask, then it is certainly possible that the bubbles are in fact solid mounds of tin. I'm not familiar with that step because we didn't do that in the PCB factory I worked in; our traces were bare copper like the Nintendo Punch-Out board above. However, the theory of solder seeping beneath already-applied solder mask to form those mounds is 100% impossible.

The only time of year that I associate with the Atari 2600 and other vintage consoles is Christmas, because that's when many people got their consoles and/or games when I was a kid (and I'm sure that's still the case today, though I couldn't care less about modern consoles). I remember when the neighbor kids, Ross and Jeff (brothers), got an Atari for Christmas, 1980. That was a huge deal, because they were very expensive at the time ($145 on sale at JC Penny that year). I got to play it a few times; Combat is what I remember most, and I thought it was amazing.

If you sanded down to copper on a bubble, then it is, of course, the trace itself that is bubbled up, thus lifted from the board beneath that bubble. If your traces on whatever board you are doing this on are silver in color then they are plated, most likely with tin. Again, there is no way for solder or anything else to get between a trace and already-applied solder mask. Solder mask is like paint; if you try to force something beneath it, it will just chip off. The solder mask is, of course, already applied to the bare PCB before anything gets soldered to it.

When I said: "The NES' composite output, as good as it is, obviously isn't going to look good on an HDTV, but on an SD CRT TV, it is beautiful." I was referring to a real NES. As for the clone system, I would expect its composite video to look better on an SD CRT than on an HDTV, but I wouldn't expect it to look beautiful like a real NES, because not even consoles from major companies like Sony and Sega have composite as good as the NES.

All of those boards with bubbling are cases of the trace lifting/separating from the board itself. It isn't possible for a trace to both be bubbled up and still attached to the board, unless the board managed to bubble up with it. As for it just being solder mask, like I said, it would have a completely different appearance if it were just bubbled up solder mask. You can look at the color and see that it is bubbled up copper. Also, press on the bubbles with your finger. If the bubbles were made of solder mask (which is very thin and has very little potential to form structures), they would easily break if you pushed on them. You'll find that those bubbles are quite structurally sound, because they are made of copper sheet. That absolutely does not happen. For starters, solder mounds coated with solder mask would be a different color than copper traces coated with the same solder mask, so that disproves your theory right there. Secondly, do you realize what kind of a mess you'd have if solder could get under the solder mask and onto the traces? You'd have short circuits everywhere, because all of the small, tightly spaced ones would be bridged with solder. These are wave soldered; which means they pass over a "sea" of molten solder. In order for this to work, you have to have solder mask covering the traces, and it has to be 100% effective. Your idea that big mounds of solder can make their way throughout the trace network by seeping beneath solder mask is 100% false. Not only does it not work that way, but solder mask isn't like a cloth sheet which can simply conform to the shape of something slid under it. If you force something under it, it will simply chip away. It is like paint. Think of paint on a car. You can dent the body to a certain point and the paint will remain adhered to it, but you can't slide something between the paint and the steel and have the paint conform to said something. It doesn't compress easily because the bubbles are dome structures formed in copper sheet, with the dome structures being supported all around their bases. Yeah, it's not a case of air bubbles in the solder mask. You can simply look at it and see that the bubbles are the same solder-mask-painted copper color as the rest of the traces. Impossible (see above). Also, please explain how solder coated with solder mask could have the exact same color as copper traces coated with solder mask. There is no interruption/change in color at the transitions from flat traces to bubbled traces. Solder is silver, copper is not. The traces most likely got bubbled due to heat distortion in the wave, along with not having a strong enough adhesive attaching them to the board.

No. There is no solder on the traces. That's what solder mask is for; to keep solder off the traces.

As I said in my OP: "Poor quality PCBs don't really affect reliability in and of themselves ..." No, the traces are bubbled up, i.e., separated from the fiberglass board. Look at the picture I linked to. If you think solder mask can take on such an appearance, then I don't know what to tell you. Also, solder mask is applied to the entire board (except for the pads of course); if it could assume such bubbling formations, the bubbling would not be perfectly restricted to the trace areas. On top of that, you wouldn't see the same color on the bubbled areas as on the non-bubbled areas if it was somehow the solder mask which was bubbling up. The color of the trace areas is a specific result of the color of copper as viewed through a thin, translucent coating of solder mask adhered to it. If the solder mask could somehow bubble up while the traces remain flat, the color would be different. And it isn't particularly common. None of my non-Atari console and arcade PCBs from the '80s and '90s have even any hints of bubbled up traces. I have seen it on non-Atari arcade PCBs, usually from the '70s, but none that I own (I own non-Atari arcade PCBs from Technos, Nintendo, SNK, Data East, and Capcom).

The traces are actually bubbled up. Solder is only on the pads, so it doesn't have anything to do with it. Also, the separation is between the traces and the fiberglass board, not between the traces and the solder mask (green insulation coating). All of my Atari main PCBs (6 2600s, 1 7800, 1 Missile Command arcade PCB) have bubbled-up traces, some worse than others. You can see plenty of bubbling traces in this picture of the solder side of a 2600 six-switch motherboard (not mine).

You should sign up for the KLOV forums and ask about it there in the Technical - Videogames - General Repair And Help subforum. They specialize in arcade machines over there.

See this thread. The picture of his Supercharger looks the same, including the resistor connected to nothing on one end.

I just installed a set of 6 new switches from Best Electronics, in my heavy sixer, and it was a breeze. One shot on each leg's solder joint with a Soldapullt DS017 and the switches just fell out (and the pads withstood the heavy suction of the industrial strength DS017 just fine; it is the same model we usually used in the PCB factory I used to work at). In the case of the sixers, the switchboard is a better quality PCB than the motherboard, so I expected it would go smoothly anyway. As for the 4-switch models like yours, the switches are on the low quality PCB used for the motherboard, so I don't know if it would go as well. Lighter suction such as from a typical desoldering gun would be a good precaution, though I don't know how well a typical desoldering gun would handle those relatively large solder joints. I haven't used a desoldering gun much. We had Metcal desoldering guns hooked to shop air at the PCB factory, but I rarely used them, because we made high quality PCBs (they went into elaborate industrial fire alarm systems, which were categorized as "life saving equipment", so they had to be high quality) and the components we usually had to desolder had good breathing room between their legs and the pads, so the Soldapullt worked perfectly fine, and was more convenient. By the way, I put a coat of car wax on all of the brushed aluminum knobs of the new switches. Hopefully this will prevent them from oxidizing/corroding like the originals did.

The original front-loader NES has perhaps the best composite video output of any console ever made; not even consoles from big names such as Sega and Sony can match it, so it isn't surprising that a cheap clone console from some random factory in China can't match it either. The NES' composite output, as good as it is, obviously isn't going to look good on an HDTV, but on an SD CRT TV, it is beautiful. Which is why the "RGB mod" he is talking about exists.

As a replacement for the original (Atari part no. C010816) on a 2600 motherboard?

Those haven't been in stock for quite a few months. They are looking for a new supplier. As for the weight: Pair of original Atari paddles, with Y-cord = 9 oz. Pair of AtGames paddles, with Y-cord = 7.6 oz. The difference in shape is subtle; they definitely didn't use the original molds.

I got my paddles from AtGames today, and now I'm even more annoyed that I couldn't order two pair of original ones from Best Electronics. These are lighter than the originals, thus they feel cheap (I've always thought the originals felt hollow/cheap too, but these are even worse in that regard). Also, they didn't get the shape/contours right, so these feel flat/squarish in your hand compared to the originals. The worst part about them is, they used switches that click for the side buttons (the originals had silent switches). On the plus side, the potentiometers are smooth with a dampened feeling, which is nice, and they measure roughly 1 megohm like they are supposed to, so they work very well.

My TV has component (YPbPr) inputs, but it isn't capable of progressive scan (at least not of the ~480p variety, which is ~31 kHz). It handles progressive scan of the ~240p variety just fine (~15 kHz), as do nearly all CRT TVs ever made. Most, if not all, consoles prior to the Sega Dreamcast output a ~240p signal, which is why you can see "scanlines" on them. The Dreamcast defaulted to 480i, which is also ~15 kHz. In any event, I bought this TV new 10 years ago, and I've connected practically every major console made between the Atari 2600 and the PS2 to it, and I've never had an issue with any game before. Granted, I haven't played a lot of different Atari 2600 games on it; I usually stick with a handful of favorites when I'm in the mood to play Atari.

What about the CD4050BE? Mouser says that's CMOS as well. Also, what would the 4050 chip be driving anyway? Doesn't it head directly to the RF output? If there are no logic circuits between the 4050 and the video output, wouldn't it be okay to use CMOS instead of TTL?

Thanks for looking into that. No, I don't have a Harmony cartridge. And that sucks about Warlords; I just bought that game (it hasn't arrived yet), and it will probably do the same thing on my main TV that Asteroids is doing. I wish Atari 2600 cartridges could be opened up without ruining the label; then I could possibly have a new, fixed ROM burned for games like this. I suppose a Harmony cartridge would make more sense, but I like dedicated cartridges and their labels, that simply boot to the game you want to play immediately. Would it be hard to fix the Asteroids ROM? I'd like to have a fixed version in case I get a Harmony cartridge.

You posted at about the same time that I made my second post, so you didn't see it. In any event, it isn't the cartridge, it's my TV. I'm aware of Space Rocks, and it would be a nice game to have, but I'm also nostalgic about the original Asteroids cartridge. It was 1 of the 5 games I bought the same day I bought my first Atari when I was 10 years old in 1985.

I just tried Asteroids on two different CRT TVs and the picture was perfectly stable on both of them. So apparently, my main TV (also a CRT) doesn't like Asteroids, which is bizarre, given that I don't have any issues on that TV with any other cartridge for Atari or any other game for any other console I own. I wonder if there are any non-standard coding tricks used in Asteroids. Doesn't the Atari 2600 rely on the game code to generate an NTSC-compatible video signal? If so, it would be possible for some 2600 games to be more NTSC compatible than others.

The whole picture shifts slightly (maybe 1/16") up and down every few seconds in Asteroids. That isn't normal is it? It doesn't do it in the Stella emulator. I tried my Asteroids cartridge in a few different Atari 2600s (heavy sixer, light sixer, and a 4-switch woodgrain), as well as in my Atari 7800, and it does the same thing in all of them. None of my other cartridges do this.

Does anyone know if this Texas Instruments CD4050BE is a correct replacement? It is the only through-hole CD4050 that Mouser and Digi-Key stocks. Digi-Key also lists two others: another one from TI (CD4050BEE4) and one from Fairchild Semiconductor (CD4050BCN), but they are both non-stock items. Edit: There is also this one (CD74HC4050E), which is the one included in the BAtari Atari 2600 AV Mod Kit. That one's probably a safe bet. And by the way, that faulty buffer did affect colors other than red in some cases, i.e., it affected the Missile Command score numbers regardless of what color they were. As I mentioned in my OP, in most cases the numbers were fatter than normal, especially on the blue background level, where they were so fat that they were all touching each other. Then it went in the opposite direction with the red numbers on the yellow level, where they were all skinnier than normal. Swapping that buffer with a good one brought both the too-fat and the too-skinny numbers back to normal.

I tested a heavy sixer (motherboard Rev. B) last night without its RF shielding quite a few times, and I couldn't see any difference in the picture compared to usual. I was surprised, considering the 4-switch woodgrain (Rev. 12) I tested without its RF shielding had a lot more RF interference in the picture than usual.