Just to play devils advocate, I go with a completely different socket theory when installing. I've given up on the machined pin sockets whenever possible, I've found that good double-wiped stamped socket tends to be the better part. It gives much more surface contact area between the socket and the chip, and also there's a bit of flex to the socket that greatly reduces the risk of damaging a chip during insertion. The machined pin sockets are so rigid, that you have to have the IC aligned perfectly or risk bending legs.

Also, I've gotten away from putting a socket under every chip. Given how much repair work is always related to sockets and connectors, I just view them as a potential failure point in the future. RAM/ROM and 40 pin chips are all I use them for now, and almost never for logic chips. Once you eliminate all the heat coming off those lamp resistors, and get the batteries off the CPU board, it removes the bigger causes of logic chip failures in the first place. Particularly the switch matrix section, which is almost always a heat-kill when something fails.

-Hans

IIRC it's either 18 awg or 16 awg that's already on there.

-Hans

It's full of a lot of "Um's", but I just uploaded a video today of using the diagnostic testing built-into a System 4-6 machine.

I'll be doing one for System 7 in the near future, there's a few differences in there. No access to a System 3 unfortunately, as that's dramatically different in how it does settings, but I'll figure something out to do a video on that.

-Hans

Most of the sound board issues I've come across tend to be caused by the ribbon cable failing. It's kinked so badly that the internal conductors fail. But it is fairly easily to replace without soldering, and cheap too.

http://www.siegecraft.us/the_siege_blog/blog/2012/06/speech-board-ribbon-cables.html

Since you are dealing with audio, I do like to replace the caps on my speech boards, if nothing else it does clean up the sound quality a bit. Also make sure to use all the mounting screw holes when you install them, and make sure the screws are tight. A weak chassis ground can cause static and distorted sounds.

-Hans

Yeah, sometimes you just don't have a choice if the solder side connector breaks. But if you take your time and go easy, it separates without a problem most of the time.

-Hans

I use the zero ohm jumpers so often, I buy them in spools of 1,000 from Mouser. Same with 1N4004's.

Over time, the resistance value of those things dropped a couple times, but I do question if it's a generational thing or just an evolution of the design. After all, the CPU doesn't even read those resistors.... the PIA chip does and they go through additional logic chips first. And the CPU speed never changed either, 6800, 6802, 6808.... all run at 1mhz. Started with 1k for System 3-4 and into early System 6, mid-way through system 6 they dropped to 330 ohms, and at System 7 dropped to zero ohms. The signal direct signal from the switches never even reaches the CPU board.

-Hans

Very possible. I remember reading somewhere that there were changes to how System 7 reads switch hits, due to limitations in the earlier designs when it came to multiple closures in a row (or was it column) at the same time. Unfortunately I've never found that reference again, so I can't provide any details.

-Hans

The relay only came into play for System 7, and Blackout in System 6. Otherwise, there's no relay to worry about in the vast majority of system 3-6 games.

I usually de-solder at 800F. Some folks may suggest a bit lower at times, but I find you need it for the 40pin, particularly the end pins. I'm also a big proponent of using RA flux for desoldering. You can get away without it, but the flux does help.

-Hans

Usually it's marked something like "Rosin Flux", or "RA type", or "Rosin Activated". Just make sure to clean it off with some isopropyl alcohol when you're done.

Using lower heat for a longer time period is exactly the opposite of what you want. The longer heat is applied, the further into the component that heat is able to soak. A hotter solder temp allows you to get the joint finished faster, and the heat doesn't penetrate as far into the part. Too hot is bad too, you need to be hot enough to get the job done fast, but cool enough that you aren't going to scorch the board. I solder at 700, and desolder at 800, both for lead free and lead. Watch for the 'slump' when it goes up the barrel and then take the heat off. 3-4 seconds tops per joint unless it's on a heavy power or ground plane.

Pace Worldwide has some fantastic videos on youtube for solder instruction. The cheezy 8mm era videos may LOOK dated, but the techniques and instruction are still perfectly applicable today. The guy from the mid 1990's at his desk is equally recommended, a bit less droll too with more good video of actual solder application. I recommend watching both series.

I'm also a big proponent of using the correct flux. If you don't mind cleanup, RA is the best performer, but there's a lot of good no-cleans out there too. Water soluble works very well, but if you don't clean the board enough it's also the most corrosive over time.

-Hans

TIP102's are the hot ticket these days. 122's do still work if that's what you have on hand, closer to factory value.

-Hans

His GI is fused on an aux fuse board in that case, with a separate GI relay mounted to the bottom of the back box. Early production machine, and same GI control circuit as Blackout.

Don't think I've ever seen a game with any symptom that the big blue is failing, nor have I ever seen any confirmed reports from anyone else that one has ever failed.

If you end up making the boards in bulk, nylon standoff's are pretty cheap if you buy in large numbers from overseas suppliers.

If space is a concern, there's also plenty of fuse packages that are much smaller than the old glass ones Williams used. Food for thought.

Those fuse boards look nice and clean, I'd probably be up for a few too. If you make them, will you also be doing the wiring harnesses as well?

If you don't, I'd be glad to offer something up for sale, but if you plan on doing harnesses I'll sit on the sidelines.

-Hans

Biggest time suck is going to be soldering those fuse clips

Digging through my stuff, and yeah, the IRFZ34N is what I used in my Black-Knight about 8 years ago. Still running strong and haven't had a single failure with it, and no issues of note.

Hey, good news! A few people have asked about my old pages for the restoration work I did on my Firepower and my Black Knight. I finally found the files for these, so they're back up on the website!!!! At least, most of it is. Still looking for a couple pages, so there will be a couple dead links in there.

http://www.siegecraft.us/blackknight.htm
http://www.siegecraft.us/firepower.htm

Did you redo the caps on the sound board as well? Don’t forget the sound board has its own independent power circuitry.

Richard;
It’s not the driver itself, it’s actually the big pill resistors that burn up. The TIP122/102/120 darlington originally used takes a fair amount of wattage to drive it. So the resistors overheat badly.

Going to the mosfet will be more efficient and it eliminates the wattage going through the resistors. Makes a big difference. I don’t have the data with me, I’m on the way to MGC, but it was a very meaaureable reduction in total machine wattage and backbox temperature.

I did the mosfet swap almost 10 years ago on my black knight and haven’t had a single failure on the driver board since then. And all I swapped was the mosfet, no resistor changes.

When I was doing board repair work, nearly every board that came in had damage due to those resistors. In extreme cases I’ve even seen them desolder themselves, or the heat cause damage to the switch matrix components directly above them.

-Hans

Ahhh, gotcha. The resistors are like the GI on system 7. Even on a good game, they’re running hot and look fried. Only once have I seen a board where the paint hadnt flaked off them yet, that driver looked pristine, likely never even installed in a game.

All Williams games exhibit some ghosting in the lamp matrix with LED's. Just the nature of the beast. LED's just switch on and off so fast that you can actually see it, as opposed to the slower speed of incandescent bulbs where you can't see it happening.

Happens with an unmodified lamp matrix too.

Looks like the previous transistor burned up, and was replaced with a new piece. Notice that particular transmission has different markings than the rest.