Just wondering if anyone knows about any hardware differences (aside from the GAME ROM chips) on a Black Knight. Here's why:

I picked up a Black Knight to aid in writing System 7 test ROMs (https://pincoder.ca) and have found and fixed many issues with it using the (not yet released) test ROMs I wrote. The final problem with the machine is that it appears to possibly have some solenoid and lamp issues when running the Williams Version 4 roms.

The two most obvious errors are:

1) In attract mode: Most of the bulbs in the lamp matrix do not light in attract mode (The most obvious being the sword lights).
2) After starting a game: The lower left 3-blank drop target reset coil kicks in and STAYS ON.

After inspecting the wires and ensuring the proper wires are on the proper solenoids/lamps, and successfully running all of my system 7 test ROMS I can only conclude that the version 4 game ROMs are not compatible with this particular unit. It came with some dead version 3 game ROMs, and I have replaced them with working 2816 EEPROMs with version 4 copied on to them. So I'm now wondering if in fact the version 3 ROMs need to be used.

So my theory is that the version 4 ROMs are used in games with different wiring (ie playfield etc) than a version 3 game. I'm really doubting it, and I would like to eliminate this possibility. I've been looking for the Black Knight version 3 images on the net, but can only find V4. If you know where I could find some please send me a PM.

About the test ROMs:

They are in-game ROMs and test all of the RAM, CMOS, PIAs, CPU, chips on the CPU board and driver board as well as displays, solenoids, lamps, and switches. They even test the ROM sockets. All of these devices pass their tests. The solenoids and lamps can all be activated and deactivated individually using these tests, and all of the driver transistors are behaving as they should.

If anybody wants a pre-release version of these system 7 ROMs just PM me. They are almost complete. I still have to include the commas in the displays test.

Any help is greatly appreciated!

Craig

Yes everything works fine in tests. Including the all of the lights in the sword. And nothing locks on either.

It's almost as if the game and system ROMs think there are different devices attached than there are. Like what I would expect if I put jungle Lord ROMs in or some other game/system ROMs.

I also wrote rom socket tests where I replace all if the other ROMs with a test chip loaded with a specific repeating pattern and check that the pattern is retrievable. That passes too so it can't be the ROM sockets or that the ROMs are somehow inaccessible.

Thanks, found them Unfortunately v3 behaves the same.

I did notice today that the sword lights work upon a cold power up for about five seconds and then they gradually fade away. Still baffles me as even with the machine warmed up the test ROMs still give normal results. I'll put a scope on it and look for rough power...

Thanks!

For the simpler tests, the LEDs on the CPU board indicate pass or fail. For the others, results are shown via displays/lamps/solenoids. They're geared to allow you to test individual components as you like without having to take out the CPU and Driver boards, so you can test the entire game's circuitry and look for specific results.

Interesting. Thanks for the info!

This one's got me stumped. So where I'm at now is:

1) Under the test ROMs all bulbs work just fine all day long.

2) Under the Williams ROMs (v3 and v4) all bulbs work fine for about 5-10 seconds and then all bulbs except column 2 start to fade off over about a two second period and then they don't ever come back unless you turn the game off for a few minutes.

Power appears clean as a whistle to me. Thinking its's a column drive issue, I've replaced IC13, IC14, Q74, and Q75 hoping to fix just column 7. Nothing changes.

I've got a scope on 2J5-9 (Column/Strobe 2, works fine) and 2J5-1 (Column/Strobe 7, troubled column) and I can see the waveform change as the Column 7 bulbs fade out.

Running the scope with the test ROMS of course, doesn't show the fade out.

However, the scope does pick up the frequency difference between the two ROM sets: Williams is refreshing at about 60hz, while the test ROMs are actually slower at about 40hz. I can't think of why a slower refresh would work better on a light bulb. I'd thought that maybe Q74 or Q75 wasn't able to keep up to the switching frequency any longer but replacing them with new didn't change the outcome.

So now, oddly enough, I have to try and speed up the test ROM refresh and see if I can get the bulbs to stop working.

The way I do it, I refresh one column at a time. I wonder how Williams does it. Has anybody gone through the Williams stuff with a debugger? That's another obvious step for me, but I thought I'd ask here to save some time..

Ha! No, but the radiation from Chernobyl probably has something to do with it!

Interesting. Blanking signal is good, and I have to assume the power supply (including lamp capacitor) is also good since I tried a whole new setup from an old pin I had salvaged and there is no change whatsoever with that. The setup meaning: everything from the power cord right up to and including the power supply board (so both bridge rectifiers, lamp capacitor, transformer etc).

I did change IC19 (ROW inverter) and got some improvement: It now takes about 30 seconds (instead of about 10) for the lights to quit. I then changed IC12 (also a ROW inverter) but changing IC12 didn't make a difference.

I've got some SCRs and related transistors coming in tomorrow so I'll see what happens..

As it turns out, ALL columns except 2 have the problem. So it does appear to be a massive column problem. However, IC19 is in the row circuit and changing it did result in a longer time before the 6 columns faded off (~30 seconds instead of ~10).

IC10 is a Motorola chip, and not socketed so I'm reluctant to change it and drop a new one in without knowing for sure - I don't have many spares and I'm not proficient at 40 pin IC removal without destruction involved. I suppose this is where a desoldering station would come in handy. Plus, the pulse and signatures don't appear to change as the bulbs fade off.

I'll try changing a few SCRs and related transistors for a given row and see if there's any change..

So I went and bought a TXINLEI 8586 2-in-1 soldering station and practiced on an old board before using it to remove IC10 (PIA 3). I got the chip out fine without wrecking the board (haven't tested if the chip survived yet), socketed and installed a new 6821 and the bulb fading has disappeared! Funny though, I didn't notice anything unusual when I had the scope on IC10. I've got a few other things to fix on it and then I'll get back to testing the old 6821 and hopefully I'll be able to come up with a test ROM for it. I'll post the results here.

Thank you all for your help

Thanks I'll do that. They are pretty cheap tips anyway, and after only a few hours of soldering the first tip has already eroded into a stub. Too bad my Weller tips don't fit. No big deal though I'll just keep using my Weller for soldering and this for the heat gun. It was only about $20 more to get the 2-in-1 model anyway, but I certainly wouldn't recommend it. The heat gun works nice though