Williams Stargate cocktail - project repair

tmcw

New member
Joined
Nov 27, 2010
Messages
62
Reaction score
1
Location
Dublin, Ireland
Got one of these recently as a project. Not working, but cabinet, top-glass and control panels in lovely condition.

Boards look pretty good too, though the power supply board looks like it got a bit scorched at some point, but the damaged trace has been repaired, and all circuits are good.

CPU board has been upgraded for the 4164 RAM, and has the relevant adapter which checks out ok.

I've reflowed the solder on all the connections, but I'm still getting a 1-3-1 on the 7-seg LED display. Monitor displays a constantly cycling "rug" pattern, though it's not exactly the same as I'd seen running in MAME; it seems to exhibit some vertical "banding" which isn't entirely of the same randomness as the typical rug pattern.

Anyhow, I've checked the voltages coming off the power board, and getting the following:

Pin 1: -4.99v
Pin 3: 11.43v
Pin 4: -14.37v
Pins 5, 6: 10.10v
Pins 12 and 13: 4.82v
Pin 14: 4.57v

From this I gather that only the -5v is ok. Pin 4 is supposed to be -12v; is -14.37v bad here? What about the others? Any suggestions on what I need to do to get this power supply good again?

Also, I was wondering if something on any of the boards could be damaged by NOT using the 4164 adaptor with the new RAM in place?

Thanks.
 
You need to rebuild the power supply, and replace both the burnt headers and connectors.

Agreed. The 12V is not used with the 4164 mod, but it is bordering on failure if it is that high and if it is the older Defender power supply (larger power supply with big heat sinks on board) it can damage the sound card.

ken
 
Thanks for the replies, I thought as much.

I've ordered the parts, but given the weather we've had here, it'll probably be another 10 days at the earliest before I get the parts.

Any tests I can do in the meantime? I'm wondering if there is a way to hook up the CPU board to a PC power supply, so that I can test some parts of the board. I've done it before with a Williams System 6 MPU and driver board on the bench.

Also, any way I can test the sound board for proper sounds, or does the CPU board need to be fully functioning to do that?
 
If the PC Power supply provides +5V, -5V and +12V you are good to go.

On the main power connector, hook:
* +5V to pins 12,13,14,15
* -5V to pins 1
* +12 to pins 3, 5, 6
* Ground to 7,8,9,10

Pins 2 & 10 are no connects

If your power supply has -12V hook it to pin 4 otherwise run -5V to pin 4.

The sound card is the only thing using -12V and it regenerates it anyway and then doesn't use it.

IIRC, the sounds are a little quieter when running with -5V but just crank the volume.

Once you have power to the sound card, if you press the diagnostic switch, you should get one of the sounds repeating over and over again. You need to shut the power off to reset it.

ken
 
Thanks for the info. I'll give that a go today as we are snowed in here, and I need something to do. I have a PSU that gives all those voltages.

Thanks for the info about the sound card, when I press the diagnostic button, I do get a sound repeating over and over again, so maybe a good sign (a kind of thump sound). I was a bit worried about it at first, as the System 6 sound test cycles through all the sounds in diagnostics. So hopefully the board is good.

I'll be back!
 
Ok, I had a go at this earlier, but no change.

A few things:

If my CPU board has the 4164 RAM upgrade, do I need to use the -5V? I couldn't see that the -5V is used anywhere else on the game.

The PSU I used was outputting about +5.03V and +11.85V, but when I had it connected to the CPU board, they were dropping to 4.55V and 11.30V. Is this to be expected, and are they still ok (as in should the board work, all other things being good) if the values are ok?

I took out the CPU board, and was inspecting the reverse when I noticed some traces that were a little scratched, but on continuity, they proved ok. On the underside of chip 6F though, I noticed pin 3 has it's trace broken. After seeing where it used to go, then looking at the blue wire that goes from the top-side of this chip, to pin 3 on 5A, then from pin 4 on 5A to the point near IJ4 (connects to pin 36 there), I surmised that this is "factory". Am I right in thinking this, and if yes, what's the story behind it?

I'm still getting the 1-3-1 error on the 7-seg. Assuming that the voltages above are ok with the PC PSU, what is my next course of action? I have a logic probe that I've used to troubleshoot PIAs and switch matrices on my System 6 board, but just need some prodding in the right direction. Of course, if the voltages are bad, I'll wait until I can rebuild the power board.

Also, when the machine is on, 3 or 4 of the RAM chips get roasting hot, whilst the others remain cool. Could this be something relevant?
 
Last edited:
Sounds like those 2 RAM chips are bad. If your power supply is having that much drop on voltage something is bad and it isn't going to be able to work at 4.5V. Pull those 2 RAM chips and give us a new voltage reading.
 
Check to make sure they are 4164 RAMs. 4116s will get very hot. If they are 4164s then check the RAM chips to make sure they were inserted correctly. There is a small notch or dot on the end that marks the end with pin 1. These should all face away from the battery holder.

If they are correctly inserted, then they may need to be replaced.

ken
 
The RAM chips are 8264A's, which I understand from the spec sheets, are equivalent. They're all inserted the correct way.

There were 3 chips that were getting quite hot, so I took them off the board, but it lead to only a marginal increase in the voltages, to about +4.6V and +11.5V. The ground was giving a reading of about -0.2v to -0.3v - is this significant?

One thing I noticed with the 3 "hot" chips removed from the board was that the "rug" pattern now looks normal, although the "wipe" does a kind of blank to a blue screen with vertical yellow lines then fills the screen with the "rug" wipe, and the wipe is about twice as fast as normal.

So I guess I need to get some new RAM chips at the very least. So I can use 4164's, 41256's, 8264's; any others that I can use?

I presume I can mix n' match 4164's with 8264's? Or can I?
 
Something on in your boardset is pulling down your power supply heavily. As far as I know, a power supply shouldn't drop that much from just running a boardset like that. With only 4.6V the game won't start like it is. I am unsure what to look for with droppage like that but I don't think that it is okay.
 
Ok, tried something else:

I took the 4164 power adaptor off, and injected the +12v, +5V and grounds from the ATX PSU directly onto the pins that require them. The two +5Vs are now between +4.75V and +4.8V, and the +12V is about +11.95V.

Sounds a bit better, do you think? Maybe with a new header and fresh solder, the two +5 voltages might get a little better.
 
Ok, tried something else:

I took the 4164 power adaptor off, and injected the +12v, +5V and grounds from the ATX PSU directly onto the pins that require them. The two +5Vs are now between +4.75V and +4.8V, and the +12V is about +11.95V.

Sounds a bit better, do you think? Maybe with a new header and fresh solder, the two +5 voltages might get a little better.

Don't sweat the 12vdc, it is not your issue.

The 5vdc IS your issue, it is still too low. You may need up to 4.95 to get some of them old ass chips to run properly.
 
Hehe, hoping I won't have to replace too many of those old ass chips at the end of it.

On your reply, I went back and had another look at the +5V being delivered at the CPU board. Since there are two pins requiring the +5V, I decided to get two seperate PC PSUs, checked that they can both deliver the +5V (and they do, about 5.05 each of them), and used one to inject to pin 3 and the other to inject to pin 5.

At pin 5, the voltage was reading about +5V, and at the RAM chips, a smidgeon lower, about +4.96V or so. A good sign, I think.

At pin 3 however, the voltage was dropping to below +4.5V, about +4.45V, IIRC. So perhaps the problem lies in this circuit?

I've only tested the two diodes on that circuit so far, D9 and D10. I've tested diodes on the system 6 pinball before, and know with the diodes on that, that I should be getting about 0.6 in one direction, and no reading in the other direction.

D10 gives me a reading of 0.588, but the D9 (1N5817 DIODE, SILICON) is giving a reading of 0.168. Both give no reading in the other direction, so at least they're not open, but is that value for D9 too low? Is it to be expected on that type of diode?
 
Last edited:
Ok, I had a go at this earlier, but no change.

A few things:

If my CPU board has the 4164 RAM upgrade, do I need to use the -5V? I couldn't see that the -5V is used anywhere else on the game.

You only need the -5V if your switching PS doesn't supply -12V to drive the sound card.

The PSU I used was outputting about +5.03V and +11.85V, but when I had it connected to the CPU board, they were dropping to 4.55V and 11.30V. Is this to be expected, and are they still ok (as in should the board work, all other things being good) if the values are ok?

Switching power supplies commonly drop when getting an actual load, although, that seems a bit much. What is the power rating of the power supply for each voltage?

I took out the CPU board, and was inspecting the reverse when I noticed some traces that were a little scratched, but on continuity, they proved ok. On the underside of chip 6F though, I noticed pin 3 has it's trace broken. After seeing where it used to go, then looking at the blue wire that goes from the top-side of this chip, to pin 3 on 5A, then from pin 4 on 5A to the point near IJ4 (connects to pin 36 there), I surmised that this is "factory". Am I right in thinking this, and if yes, what's the story behind it?

That is standard on a Rev B motherboard. There was a signal that was not used on Stargates that needed to be inverted. Rather than hold up production for 2 months waiting for the next rev of the motherboard, they went with an engineering fix. Here is a link that explains it in more detail (link).

I'm still getting the 1-3-1 error on the 7-seg. Assuming that the voltages above are ok with the PC PSU, what is my next course of action? I have a logic probe that I've used to troubleshoot PIAs and switch matrices on my System 6 board, but just need some prodding in the right direction. Of course, if the voltages are bad, I'll wait until I can rebuild the power board.

Also, when the machine is on, 3 or 4 of the RAM chips get roasting hot, whilst the others remain cool. Could this be something relevant?

The 4164 chips should remain quite cool. If any of the chips are getting hot, they are most likely bad.

At the moment that low +5V is your major problem. You may need to try a beefier power supply (to keep the voltage up) or see if swapping the chips out helps the voltage. Resoldering the power input headers and putting new connectors on the input power plug may also help.

ken
 
Thanks for the replies.

Tested the D9 out of circuit, and it was about the same reading as in circuit. I also have the same type diode in a battery pack for my pinball machine, and it's a similar reading (a bit higher, about 0.215), so that's probably not the issue.

I've reflowed the headers, but it hasn't made much of a difference, slight, maybe.

PSUs are 250W. Like I said, not using the same PSU for pins 3 and 5 seemed to make a difference, just wondering why I was seeing a drop in voltage at pin 3.

3 of the chips were getting hot, so I've taken them out. I'll order some more. I presume it's ok to mix up compatible chips? Eg, 8164's alongside 4164's?

Until I get all my parts together, and can rebuild the power supply and connectors/headers, I think I'll suspend any more testing. Was a similar situation with the pinball machine until I redid the 40-pin interconnector and some of the more suspect connectors, too many variables there. I think I'll replace all the headers too, as they all seem to be of round cross-section. Square cs should give me more reliable connections.
 
Back again...

Got my parts, and finally had some time to replace the capacitors, 2 ICs and the 2n3055 transistor on the power board. Also did the "JustSayNoTo4116" mod on the CPU board, because I felt that the 4164 adapter that was on it was a bit flakey.

My +5V and +12V regulated are both now good all the way to the CPU board and RAM chips, but the +12V unregulated is still low, just under 11v. I'm still getting 1-3-1 and the display on the monitor is just garbage, not a proper rug screen, but I can see it "wiping".

I found Dokerts video on how to test the bridge rectifiers, and all 3 test ok. BR3 was approaching the high side (about 580) but from what I understand from the schematics, it's involved in GI, so not my concern ATM.

So, what's my next action with the +12V unreg? From what I can make out, only the C12 (18,000uF, 20V) and BR1 are involved here after the transformer. The -12V unreg is also not great, at about -14V. Any other tests I can do around here, or is it just replace the C12 at this point?

Also, just curious about this, but when I was looking at pictures of Defender power boards online, I was noticing the R15 is missing on mine, it looks like it burned up at some stage, and in it's place is some heavy gauge insulated wire, on both sides of the board. From the schematics, I can see the note "W1 connected only for cocktail applications", W1 effectively bypassing the R15. Anyone know why this is so, is it related to marquee lights that would be on an upright, but not on a cocktail? Is it possible the power board was once in an upright, moved to the cocktail, then burned out the R15?
 
R15 is removed and W1 (jumper) is installed for a CT. Yes, the board may have come out of an UR. The power coming off of the iso is a little different for a CT compared to an UR.

The 18K caps are notorious for failing and should be changed. You can get the replacement from Ed at GPE for $5.50. (www.greatplainselectronics.com)

Still getting a 1-3-1 error, verify that the socket is good. Hell, it is just easier to replace the damn socket and see what happens.
 
On a cocktail, you also need jumpers w2 & w3 (between the big cap and the fuses) unless you have a 15pin connector from the transformer (white and yellow wires).

That is the main difference between coctail and upright power bricks. Coctails used old Defender transformers and had no wires on pins 14 & 15. Uprights got newer transformers that had the 6.3VAC on pins 14&15.

Unreg -12 is normally -14VDC. The original transformers were spec'ed for 115VAC. Most wall AC now is 122-124VAC which bumps the unreg voltage up. It is only used to run the sound card.

If you've done the 4164 mod to the board correctly, you should not see +12V Regulated on the MPU board. It is no longer used. As is the -5V. The +12v unregulated is only used to trigger the CMOS memory protect on power down, so this has nothing to do with the RAM. It sounds like the CPU bus memory addressing chips are toast.

ken
 
Back
Top Bottom