Stargate -5v problem SOLVED!

shuyge

Member
Joined
Jan 19, 2017
Messages
83
Reaction score
12
Location
Colorado Springs, Colorado
I recently picked up a non-working Stargate. This is my first Williams game so I've had to come up to speed on the basic hardware.

IANAEE. I am a S/W Eng. by trade which means that I can understand the basics of an arcade schematic and I understand address and data buss's, address decode logic, etc.
In other words, I know enough to be dangerous..

The CPU board flashes a 1-3-1 but that doesn't matter since I don't have good voltages coming from the power supply. The 4116 RAM can't work without all 3 voltages.
I have two leds lit. #1 and #2. #3 won't light.

I should mention that this is the later power supply board with the power transistors located on another board with a large heat-sink. I'm using the schematics from Robotron to troubleshoot the problem since they seem to match my board.

So the first thing that I did was to purchase a cap kit and installed it. Same problem.

So I'm starting to get a bit aggravated with the power supply. So I start looking harder at the schematics and the -5V supply looks like a standard power supply circuit that I've seen many times.

There's the normal A/C input, a bridge rectifier, filter cap, voltage regulator. Normal stuff. The only part that is odd is that the outputs of the bridge rectifier outputs are hooked up backwards to provide negative voltage. But since this is a -5v circuit, it makes sense.

So I grabbed my multi-meter and started probing the board. I'm actually getting -1.2v out of the circuit. I checked the legs of the 7905 to see what I can find..

From pin 1 to 2, I have -10.8v
From pin 1 to 3, I have -5v.

WTH?

So now I really start looking at the board and I pull it out of the machine again and I start doing some continuity checking and actually looking at the traces on the board. What I see confuses me. The ground that is labelled as 0v that comes from the bridge rectifier isn't connected to the other grounds in the circuit. If you look at the schematic, there are two grounds that are shown in the circuit, the output of the bridge rectifier, and the one that the #3 led is connected to. Those two grounds are not connected in the circuit. The ground from the led is connected to the other grounds on the board and also the ground that goes to the CPU board. The ground used with the 7905 isn't connected to the rest fo the board at all...

So I put the board back in the game and I locate and measure all of the test points on the board:

TP1 - Supposed to be 5.1V, measured 5.2V
TP2 - Ground
TP3 - Supposed to be 12V reg, measured 12.46V
TP4 - Supposed to be -4.9V measured -1.2V
TP5 - Supposed to be 12.8V unreg, measured 14.21V
TP6 - Supposed to be -14.8V unreg, measured -14.62

So then I also measured from the 7905 pin 1 (ground) to TP2 and I find my missing -3.8V.

So it appears that my grounds are not the same and the -5V circuit ground is floating at -3.8V.

Have any of you guys seen this problem before? Is it the bridge rectifier? Something wrong with the input A/C voltage?

I do have a scope, but I'm a noob with it and I don't know where I should hook the ground clip to start probing this circuit...

TIA, Stan
 
Last edited:
Williams PS

Check all the resistors on the PCB and write down the values. My guess, based on similar experiences with these power supplies, is that some of the resistors are out of range. Probably R5 and R17 and I think R23.

I just went through this with 3 WMS power supplies and all three had 2-4 resistors that needed replacement after a full cap kit, transistors replaced and voltages regulators.

You may want to perform the adjustment pot mode to:
http://cambridgearcade.dyndns.org/?q=node/76

Good luck!
 
Many thanks for reading the wall of text for my problem. I'll check the resisters when I get home tonight. Do you remember if I can check them in circuit or if any need to be removed to check them?

Stan
 
Don't have one in front of me however based on the schematics, check for a bad connection on pin 7 on the output. I believe this is the pin that's clipped off as the key pin. The schematic for the older board shows the ground for the -5 going to pin 7. Perhaps a broken solder joint there is preventing the ground from being connected to the common ground plane?

Cracked solder joints on Williams boards are very common. If you haven't replaced the headers, which is recommended, you can pry the plastic up carefully and resolder the pin on both sides to ensure good connection.

Good Luck.

Kirk S.

P.S. The Williams power supplies used some very standard parts and are based upon circuits described in those parts datasheets. The -5 supply is a basic 7905 supply with no resistors and three capacitors. Not much to go wrong there.
 
OK,

So I measured all of the resistors in-circuit and as you mentioned, 3 of them didn't measure out correctly. They were R5, R17, and R26. However, when I removed them from the board and measured them, they all measured correctly so I put them back in place.

I also removed the bridge rectifier BR3 and tested it out of circuit and all of the diodes tested fine.

I'm still at a loss as to why there is -3.8v of differential between the -5v supply and the +5v and +12v grounds.

Stan
 
Last edited:
PS

If R5, R17, and R26 didn't measure correctly I would replace them. I had the same issue and resistors are cheap to replace.

Otherwise, if you have a few dollars to spend and don't mind doing some more work replace the transistors and BR1-3 if it's not working after the resistors are replaced.

Check for cold solder joint as well.
 
Looked at a later model power supply board and it appears that C14 is common to both of the supplies as a connection. It looks like the -5 supply connects to the bottom and the +12 supply connects to the top of the board. It's very common to lose the connection between the top and bottom of the boards when performing a cap kit due to damage to the through hole plating.

Kirk S.
 
good point

That's a very good point to make. :)

It's so hard to desolder these PS PCBs sometimes to re-cap them or work on the power regulators that when you replace the parts you should solder top and bottom of the PCB.
 
Kirk,

Do you have a connection from the -5V ground (pin 1 of 7905) to the rest of the grounds on the board? I do not have a connection there.


That's the kind of info that I was looking for. I'll have to take a look at my board when I get home tonight.

After looking at where C14 is located, I don't think that this is my problem. C14 is the +12v filter cap. Also, C14 is pretty tight to the board as I recall. I can flow some additional solder from the backside though in hopes of making said connection.

In the meantime, I got frustrated and went ahead and installed my 4164 ram and adapter and promptly lost the +5 on the board.

So I got frustrated and walked away for the night. I'll take a look at where my +5 problem is tonight as well.

Thanks,
Stan
 
Last edited:
to Kirk's point, the easiest way to sink a Williams linear is losing the through hole plating in the pads. I don't think I've ever done it on a later Stargate/Robotron/Joust power supply, the Defender ones you tend to lose some pads even with good tools. :p

I've only ever fixed the problem by opening the traces top and bottom (scrap or sand the mask), desolder whatever part you need to reconnect the two halves and run the wire through, lay wire flat on one side, solder, then lay it flat on the other, solder, then reinstall the part you moved to run the wire through.

depending how fat the trace is will determine how many strands I use. I think you can fit up to 3 strands, even with the fatter new header pins.

alternatively if it happens on a header pin, you can open the trace on the component side and tin it and suck the solder back off and then when you install the new header you can dump the solder through the hole to reconnect to the solder side. of course always verify all of this with your multimeter to make sure everything's connected again.

the -5 circuit on these is the simplest one, so about the only way you could break it is with solder/component sides not connecting. I haven't worked on one in a long time and can't even remember what it looks like right now.
 
I checked the board and the connection at C14 had a common ground to pin 1. There is a very thick trace on the top of the board which connects on the - of the capacitor. The opposite side has a thin/thick trace which connects to the outside of the board.

I've noticed that the Williams power supply boards don't have vias, or through holes. They are like the old Magnavox projection TVs. Fine if you have a solder bath, otherwise you need to have a hefty soldering iron to be sure things don't go pffffft...

Sorry, too lazy to take a picture. the + on the -5 supply must be common to ground on the others. Ground is ground... That's where the electrons flow from. The earth has plenty of them... :)
 
Kirk is the winner on the -5v problem

The problem was exactly as I suspected in my first post in this thread.
I did in fact have a floating ground on my -5v circuit. I just didn't know how or why.

C14 plays a much bigger part in the board design than I realized.
I had to tin the top side pads and raise the capacitor as much as I could (I trimmed the legs a bit short trying to seat the capacitor low to the board) and reach the tip of my iron in and solder the top side to the negative lead of C14.

Actually, now that I've re-read mecha's post, it appears that I have put in a poor fix for the problem. I'll have to revisit it later.

Now the question is should we solve the broken +5v that I now have in this same thread or should I start a new one?
Also, are there other locations where there is through-hole-plating that I need to be aware of on this linear power supply?

Thanks again guys!

Stan
 
Last edited:
Back
Top Bottom