Game Plan machines don’t get a lot of love. There’s a lot of history behind the company along with some intrigue. If you’re reading this you probably share some of the GP love and know it is increasingly difficult to keep these games running.

With the word that JimF @ Echo Lake may not be making replacement MPUs anymore for Game Plan machines, I thought it would be a good idea to document how to fix a GP MPU as it is the part that probably fails the most often - usually due to a leaking battery.

If you have a working GP machine and it has the original battery, stop reading, go find your machine and cut the battery out. Now. Right this second. Don’t wait. It’s a ticking time bomb and dealing with the after math of the corrosion from a leaking battery is not something you want to have to do. The battery is not necessary for operation as all it does is keep audits and high scores. Settings, including free play, are made with the 32 DIP switches on the MPU board. You'll get garbage on the displays for HSTD with no battery but that's about it.

In this thread I will be repairing a MPU-2 board. These boards initially came into use with Sharpshooter and were used pretty much until the company shut down. They are backwards compatible with the MPU-1 boards which were used in the early GP cocktail machines. Much of what I’ll cover here is applicable to both but you should check schematics and be sure of the differences in case you’re working with a MPU-1 board.

Some of the techniques I'll show can be used with other boards as well. But this thread will stay focused on the Game Plan MPUs.

July 2014 edit: I and others have reached out to JimF. While he feels his product has matured and demand has slowed, if he gets enough requests to do another run he will. He currently prices the boards at $275 shipped. More details on this thread: https://pinside.com/pinball/forum/topic/game-plan-pinball-owners-and-future-owners-must-read

_{Post edited by viperrwk: Added info on Echo Lake repro boards}

The MPU-1 Board

Here’s a MPU-1 board.
mpu1boardwcaps.jpg

This one already has the battery removed and will get a memory capacitor. The main difference between this and the MPU-2 is that it only supports two ROMs (not three,) it doesn’t have the hookups for the BDU-1/BDU-2 displays used in upright machines and usually did not come with the 6810 memory installed. There were two Game Plan cocktail machines which used the MPU-2 – Vegas and Lady Sharpshooter. All the rest used the MPU-1 board. The MPU-1 board could be used in Vegas with the correct ROMs and if there were a 6810 installed but Lady Sharpshooter requires the additional ROM space of the MPU-2. So the MPU-1 board for all practical purposes should be considered an “early cocktail MPU” only.

The MPU-2 Board

There were three different versions of the MPU-2 board which I’ll call "early", "middle" and "late". All are backward compatible to the MPU-1 and can be easily installed and plugged into any Game Plan cocktail pinball machine (with the correct ROMs of course.) The only major difference between them is in the reset circuit.

"Early" boards are marked “REV. 1” in the upper left hand corner.

mpu2rev1.JPG

These boards appeared in the early upright machines like Sharpshooter - http://ipdb.org/showpic.pl?id=2113&picno=13427&zoom=1 - but I imagine you won’t see too many boards like this.

The "middle" boards are essentially REV 1 boards which were modified at the factory by adding a daughter card to help with 12v detection in the reset circuit (call it “REV 1A”.)

mpu2fullboardrev1awcaps.jpg

A common failure point is breakage of the leads from the daughter card to the main board. I’ve seen these boards in both Sharpshooter and Vegas machines (the earliest MPU-2-based machines.)

"Late" boards, which are marked “REV. 2” in the upper left hand corner, incorporated the daughterboard components into the main board eliminating the need for the daughterboard entirely.

mpu2rev2.JPG

I’ve seen this board in machines as early as OCI! It is unclear to me how many REV 1 vs REV 1A vs REV 2 boards were made of the more than 14,000 out there. If your board is an early type, don’t sweat it. As long as everything is working properly and you have all your voltages present you shouldn’t have any problems.

mpu2fullboardrev2wcaps.jpg

MPU-2 Rev 1 schematics can be found in this .ZIP: http://mirror2.ipdb.org/files/2113/Game_Plan_1979_Sharpshooter_Schematics.zip

MPU-2 Rev 2 schematics can be found in this ZIP: http://mirror2.ipdb.org/files/1407/GamePlan_Lady_Sharpshooter_Repair_Manual.pdf

The Reset Components, and a Note For What's to Come
Interesting to note is that the components on the daughterboard of the REV 1A board are identical to the ones used on the REV 2 main board – 2.2k ohm (1), 56k ohm (1), 10k ohm (2) resistors and 2N3904 (2) transistors.

mpu2rev1reset wcaps.jpg

mpu2rev2resetwcaps.jpg

This guide assumes you have good power going to the board (+5v. +12v, +24v) and that you don’t have badly corroded components. It’s beyond the scope of this posting to document detailed power and corrosion issues. If your board has lots of corrosion, cutting off the components, neutralizing the corrosion and replacing them, and insuring all the circuit traces are making connections is critical. There are plenty of guides and posts out there that show you how to do this. I will cover some of this along the way but every board is different and you never know how the corrosion will go. And don’t assume that if you get corrosion off the legs of a chip that the inside is ok. It’s easy for corrosion to travel into the carrier and destroy a trace. All of this is reason enough why you have to cut off the battery.

Game Plan Flash Codes

GP was like Bally in that it has a diagnostic LED that flashes as the machine boots. If you get to six flashes then the board has fully booted and the MPU is functioning properly. If you get less than 6 flashes it indicates a problem with the board. Oftentimes you won’t get any flashes. This could be a power problem. It could also be a problem with a damaged reset circuit due to a leaking battery. It could also be a problem with one of the socketed components on the board. (More later on what to do if you don’t get any flashes.) Below is adapted from the GP troubleshooting flowchart.

If the reset, CPU, PIA and voltages are all good, you should get at least one flash. If it stops after one flash, swap out the Z80 CTC (U10.)

If you get two flashes, either one or both U6 and U7 RAM are bad. These are difficult and expensive to come by. At this writing Arcade Chips still has 6551s but they are $8 each.

If you get three flashes, the 6810 at U8 is bad. It is possible that you may have a MPU-1 board without a 6810 – that’s ok. You’ll still get a flash as if you have a good 6810.

If you get four flashes and it stops, that says that part of the PIA is bad. This is an unusual situation because the LED will not flash at all with a bad PIA. If you haven’t replaced the PIA at this point now is the time. But if you replaced the PIA in the first place with a known good one, then the 74LS00 at U4 is likely bad.

Finally if you get to five flashes and it stops, the board has gone through all its internal tests and is trying to boot from the ROMs – unsuccessfully. At this point you know the ROM(s) are bad and need to be replaced.

When you get the sixth flash you know the ROMs are good, the board is running the game code and the board is fully booted (though you still may have a problem with the switch matrix, solenoid lines or displays – something I’ll discuss later.)

If you replace any of the above socketed components with new ones and still have the same problem, you have to suspect a bad socket and/or corrosion which may be preventing the board from communicating with that part. Or possibly one of the other soldered in logic chips on the board. The only solution is to find, remove, and replace the offending component.

I get no flashes – what do I do?

If you don’t have any MPU flashes and the power is good and there’s little to no corrosion, I take the shotgun approach of replacing all the socketed components on the board: Z80 CPU (U11), Z80 CTC (U10), 8255 PIA (U17), ROMs (U12, U13 and most times U26, depending upon the game), 6810 (U8) and 6551s (U6 & U7.) Replacing these components is the last step in the GP troubleshooting guide for no MPU flashes but chances are by replacing all these components you’ll hit the bad one. All together with the ROMs they will cost you about $45 + shipping.

major componentswcaps.jpg

My thinking in replacing them all is simple. Unless you do board repair all the time and have all the components on hand, you’ll likely have to order components to fix the board. And it makes no sense to order a PIA, replace it, discover it wasn’t the problem, order a CTC, replace it, discover it wasn’t the problem, etc. Also, if you've replaced them all and it still doesn't work, you've eliminated eight components as being the reason why the board isn't working and can focus your repair efforts elsewhere. So order them all and replace them all then change each one out one at a time until you find the bad one that keeps the board from booting. And $45 is a small price to pay to get a board that is becoming unobtanium, working.

If you have no flashes and want to take it a step at a time, you should at least replace the PIA. If that’s bad, you will not get any flashes out of the MPU. (Yes, though unlikely, you could also have a bad diagnostic LED.) You can have other bad components including a bad PIA, but the bad PIA will mask the problem. And in general the PIA fails more often than the CPU. So if no flashes, always swap out the PIA first. If still no flashes after that, then swap out the Z80. If you still have no flashes after that, chances are something has failed in the reset circuit, likely corrosion related. Regardless, the goal here is to get to a booting MPU.

Game Plan ROMs - Modding the Board for 2716s

The MPU-2 board came from the factory configured for 2316-type PROMs. With any sort of corrosion or tarnishing of the pins, these ROMs could very likely be damaged and unusable. Below are a pair of original ROMs from an Old Coney Island - notice the tarnishing on the pins.

original roms.JPG

The easiest thing to do is get a new set of standard 2716 ROMs. The good thing about the MPU-2 board is that it is designed to accommodate 2316s, TMS2716s, standard 2716s, 2532s or 2332s with a few simple board modifications. (MPU-1 boards can also be cut and strapped for 2716s but I will not be documenting that here at this time - perhaps in a later post.) Standard 2716s are best since they are easily available, and most EPROM programmers can handle them.

To set the board to use 2716s simply requires cutting two traces and installing two jumpers. This process breaks the ground connection to pin 21 of the ROMs bringing 5v instead and drives the logic level correctly for reading 2716s. Trying to interpret the schematics for what to do here is a challenge, to say the least but it is easy to do if you can see what needs to be done. Three steps are done on the component side of the board (steps 1, 2 & 4 below) and one step is done on the solder side (step 3.) Here’s the photo guide to how to do it.

Step 1 - First thing you need to do is cut a trace on the component side of the board. There is a via which is between pins 35 and 34 of the Z80 that then connects to pin 21 of the ROMs. You’ll know this is connected because you can test for continuity between test point 7 (ground) and pin 21 on any of the three ROM sockets. Cut the trace carefully next to the via with a Xacto knife, insuring you don’t damage the traces on either side of the via. If you don’t want to cut the trace you can remove the solder from the hole and drill it out with a tiny drill – you just need to break the connection from one side of the board to the other. You’ll know you’ve successfully cut the trace by testing for continuity between TP7 and pin 21 on any of the three ROMs – there shouldn’t be any.

rom step 1wcaps.jpg

Step 2 – Next, on the component side, install a jumper between the two pads below U26. This brings 5v to pin 21 which is required to read the 2716s. Using a cut off lead from a resistor is the easiest way to make this jump. Remove solder from the holes, install the cut off lead, solder it in place and trim. You’ll know this is working because you should now have continuity between TP1 and pin 21 on any of the ROMs.

rom step 2wcaps.jpg

Step 3 – Now, on the solder side of the board cut the trace that goes to pin 8 on U24. You’ll see a mark on the board that indicates what to cut. When you’re done, there should be no continuity between pin 8 on U24 and pin 18 of U13.

rom step 3wcaps.jpg

Step 4 – Now back on the component side of the board install another jumper between the two pads above U24 as shown – this is the bypass for U24 to provide the correct logic level at pin 18 of U13. You’ll know this is working if you have continuity between pin 8 at U25 and pin 18 of U13.

rom step 4wcaps.jpg

That’s it. The board is now ready to take standard 2716s which you can simply pop into the sockets and they will be used by the machine.

A Note About ROM Versions

Game Plan recycled many of its designs. For example, Sharpshooter, Old Coney Island and Sharp Shooter II are essentially the same design with OCI being the mirror image of the other two. Many of the early cocktail models were the same. Real, Rio, Black Velvet, Camel Lights, Chuck-A-Luck and Foxy Lady were all model 110, Family Fun and Star Trip were both model 120. What this means is that some ROMs are used in more than one game. Sharpshooter and OCI use the EXACT same ROM set. SSII uses the EXACT same ROMs as the other two at U12 and U13. So a working board from a Sharpshooter can be dropped directly into an OCI and vice-versa. However, some have reported that the SS ROMs can also be used in a Sharp Shooter II. While my testing has shown that this is possible, the fact is that SharpShooter II used a different ROM at U26. Sharp Shooter II had a different sound system and the ROM at U26 was different to utilize the newer sound board. While it is possible to use Sharpshooter ROMs in a Sharpshooter II or vice-versa, you will not get the correct sounds to play if you do.

The best place for ROM versions is the Game Plan Pinball web site - http://gameplanpinball.com/roms.shtml. If you need to buy ROMs, John Wart Jr. should be able to burn you what you need: http://thatpinballplace.com/eproms/index.htm

The Board to Be Repaired

This is the board to be repaired. It came out of a Sharpshooter. You can see this is a mid-MPU-2 board with the 12v reset daughterboard on it. This board has had previous work done to it. It is currently not working and not providing any flashes when power is applied to it. Additionally, there are numerous jumper wires on the board, the 12v reset daughterboard is broken off, and someone already replaced the factory ROMs with standard 2716 EPROMs.

top repairwcaps.jpg

bottom repairwcaps.jpg

board backwcaps.jpg

First Step, or Last Step, Verify the ROMs

My first step is to pull the ROMs and check the images. Since these are standard 2716s they can go in most any programmer and be checked. Since this is easy to do I do it first. I downloaded the ROM images from the GP website listed above to check these against.

U12 – bad checksum, doesn't verify
U13 – good
U26 –bad checksum, doesn't verify

So two of the three ROMs are bad. I’ll erase and reprogram these if possible as well as use a new set of 2716s just in case. We now know right away that this board would have never fully booted.

If you have original factory masked ROMs on your board, and/or you can't read them to verify them, and the board doesn't boot, mod the board as described above and install 2716s. It's easier than trying to figure out if the masked ROMs are any good.

J8 and U9 In the Path of Corrosion

Below where the battery sits is the J8 connector and U9, a 74C32 as well as some decoupling capacitors.

J8U9 wcaps.jpg

Someone replaced the caps earlier but the corrosion has taken over U9. The J8 connector is here for one reason – to provide -5v for tri-voltage EPROMs. Since I will be using standard 2716s and it is unlikely this board will ever use tri-voltage EPROMs again, (never mind the fact that GP almost never used them either) this connector will not be replaced.
So I'll remove J8 which has the added benefit of allowing easier access to U9, which gets cut out with a pair of flush cutters and removed. U9 was so badly corroded that it came apart as I cut it out. The 74C32 is obsolete so the best replacement here is a 74HC32 and socket. I’ll discuss the rationale for the replacement later.

j8removed wcaps.jpg

Always Check Previous Work!

I did a closer inspection of the board – the ROM socket at U12 which has been replaced is not sitting flush on the board. In addition there is the aforementioned jumper wire installed on the solder side behind this socket.

u12 socket wcaps.jpg

With U12 being bad, this work has to be suspect. It is impossible to say who or why the work that was done was done but it will now have to be undone. The socket will need to be replaced and there may also be through-hole damage. You always have to carefully inspect any previous work done to a non-working board.

And further examination shows that only step 2 above to make this board work with 2716s has been done. And in fact the white jumper wire on U12 brings the 5v and ground planes together directly through the ROMs! Whoever did this did not understand exactly how to get the 2716s set up on the board.

More Bad Previous Work

Further inspection of the DIP switch banks shows why the long jumper wire on the back of the board. Specifically the trace for DIP switches 8, 16, 24 and 32 has been damaged. Additionally, diodes are missing on switches 8 (of course), 10 and 26.

dips wcaps.jpg

I’m guessing that the diode at switch 8 was removed to prevent the machine from running in free play mode and when it was removed, the damage happened. Obviously that diode will need to be replaced. Sw10 is for adjusting the number of credits for coin chute two. If all are off (result with the diode removed) it will be the same as coin chute one. Sw26 is part of the set of switches that controls the maximum number of credits the game will take. With the diode out (off) max credits is 5. The logic behind what happened here is baffling.

Green Decoupling Capacitors

Game Plan MPUs came from the factory with these green disc decoupling capacitors all over the board. Almost all of these caps are .1uf 25v ceramic. If you want to know what a decoupling capacitor is and what it does, you can read the Wikipedia article: http://en.wikipedia.org/wiki/Decoupling_capacitor

It’s not uncommon for these capacitors to have problems, not the least of which is reacting to battery gases and even sticking to the board. (see photos)

DC2 wcaps.jpg

DC3 wcaps.jpg

In general it is best to replace them all, even if you have an issue with just one. They can be replaced with smaller, more efficient axial ceramics - http://www.greatplainselectronics.com/search.asp?pg=1&stext=CCA-0.1uF-50V&sprice=&stype=&scat=. At $.08 apiece it’s well worth it. You’ll need 13 in all for a MPU-2.

There is one capacitor located above U20/U21 that is not a .1uf – it instead is a .01uf and is for the solenoid/sound select chips on the board (74LS00, 74LS379.) According to John Robertson of flippers.com, this is a critical timing cap for the solenoid select circuit - http://flippers.com/images/GamePlanMPU-bad-cap.jpg. You can also replace this with an axial ceramic - http://www.greatplainselectronics.com/search.asp?pg=1&stext=CCA-0.01uF-100V-AVX&sprice=&stype=&scat= .

Someone previously replaced *some* but not all the decoupling caps on this board. In the process of doing so, they also damaged some of the mounting holes for these caps. These will need to be repaired and all new caps installed.

The Major Components Tested

Next, I removed the major socketed components (CPU, CTC, PIA, 6810, two 6551s.) Sourcing them is a pain. No one supplier will sell them all to you. The 6551s are the most expensive part. All together they cost about $30 + shipping. You hope none are bad but if some of them are bad like the PIA or CPU, you may not get any flashes on the MPU. I already know the ROMs are bad and will reflash them if possible. I’ll also make a complete second set of ROMs anyway just in case. If this board wasn’t in such bad shape, I would simply swap out all the socketed components and try to boot it. Since I want to know exactly what I need to replace on this board and already have a known working MPU-2, I’ll simply install all the socketed components in that board and then isolate.

In addition to testing all the socketed components, removing them also offers you the advantage of inspecting the pins on the sockets - another place where corrosion can hide. Unfortunately, one of the places I found corrosion was in the PIA socket. This, along with the U12 ROM socket will have to go.

piasocketwcaps.jpg

The good news is that the legs are clean and shiny on all the chips – save for the 6810. That’s where the corrosion was by the 74C32. And while the 6810 socket itself looks ok, the via next to it doesn’t, so I’ll remove that socket as well.

Here’s the 6810 and the corrosion on the pins. The problem is that this corrosion weakens the pins which makes insertion more difficult as well as break the small part of the pin from the large. I will try cleaning these pins front and back but this may need replacing, especially if the corrosion has creeped inside the carrier. Some would say just to replace it – new ones are plentiful and pretty cheap, around $2.50.

6810wcaps.jpg

I swapped all the chips over to my working board, hoping I would get lucky. I wasn’t – no flashes. First thing I changed was the PIA. Interestingly, after doing so I got one flash. That tells me the PIA was bad but the CPU is good. And it means the Z80 CTC (clock-timer-counter) is bad. Swapped that out with a new one and bingo - get six flashes. Fortunately the 6551s and the 6810 were both ok but it will cost $8 for a new 8255 PIA and Z80CTC. While I was at it I checked the reflashed ROMs as well as the new ones I burned. All tested good and we now have a complete working major chipset. So the chips that will go in this game.

Original components that will be reinstalled:
Z80
6810
6551x2

New components that will be installed
Z80CTC
8255
2716 EPROMs x3

And the old reflashed EPROMs will go back as well as backups.

Post a video or it didn't happen!

viperrwk

Yes, the connectors between the head and cab are problematic on GP machines. The pins can easily corrode and cause you problems in operation.

Basic booting of the board (6 flashes) should not be affected by the connectors themselves unless you have a power problem. However, solenoid, lamp or switch problems which you might think are a fault of the CPU could be the connectors between the cab and head.

Should be .093. Yes they are a total pain to take apart. If you're going to repin you could do the entire housing to make your life easier.

See above. You can cut the wires at the connector, strip, repin and put in a new housing. Both sides usually need to be done anyway if you have to do it. Otherwise you can try the molex removal tool (your Amazon link did not work for me.)

Yes, that makes sense though I'd have to go do some digging to tell you exactly which ones to get to replace what's there. When I finish posting the rest of this I'll have a look at what it would take.

viperrwk

Yeah - I noticed the pinwiki GP section was "thin." I thought about posting this there first but decided to do it here as breaking it down into posts is feeding into how to orgainze it in the wiki. When I finish here I will definiitely put the relevant content there.

viperrwk

If it's true that Jim will no longer be making boards, it might be time for you to dust off those schematics John!

viperrwk

Yee ha! Nicely done! And the sling is working as well - excellent! Since it's snowing you should go play some more!

Thanks for posting!

viperrwk

Removing Components, Spotting Problems

Removed all the old green disc capacitors and the replacements someone else had put on the board. As stated earlier, these can be problematic so I'll replace them all with axial ceramic. They are cleaner looking and less prone to breaking. I also took a first quick pass at cleaning up the solder pads for U9.

board mid stripwcap.jpg

Unfortunately, I lost some of the solder pads in the process of doing this. One way you can inspect your work and tell whether you've done damage or not is by backlighting the board. You can see below that I lost a pad at U9 pin 4 due to the corrosion that was there. And the pads for the capacitor below J8 which were lost due to the previous work. You can also see the damaged trace above and to the right of U8 with the backlight.

pads back lightwcap.jpg

Below is the mount for the cap above connector J2. Someone had replaced this previously and blobbed the solder on the top of the board where the connector for this cap is located. Whoever did this repair did not use the right amount of heat and damaged the through hole on this plane. Removing the cap made it worse. All of these will be repaired with eyelets and will be as good as new when done (more on eyelet repair later.)

pads back light 2wcaps.jpg

It bears repeating at this point that whenever you do troubleshooting on any board, you have to suspect previous work. It is increasingly unlikely that you will get any game where the boards haven’t been touched at least once. And when you do any board work, think about the person who might have to troubleshoot it down the line. Documenting the work you do and why is not a bad thing and I believe will add value to a game.

Removing the Connector Pins

Next, I removed all the remaining header pins from the board as well as the DIP socket at U8 for the 6810. All the pins used on the MPU are .100 size.

board mid strip 2 wcaps.jpg

The reason is that after all these years and use, the plating on the pins is gone making for a poor electrical connection. Replacing them insures no problems with the connectors. Truthfully the plugs should be repined as well but since I’m only working on the board this is all I can do.

pins no plating.jpg

Removing the connectors revealed that there was some corrosion on the J7 connector that could not be seen with the connector in place.

j7 pads wcaps.jpg

I removed the socket for the 6810 because of its proximity to the corroded U9. I was surprised to find this:

u8 pads wcaps.jpg

The corrosion on the via between pins 19 & 20 made me remove this socket in the first place. But I was surprised to see the corrosion on the pads for pins 5 & 6 across from it. As I indicated earlier the legs on the 6810 didn’t look great but it did pass the test and booted on a good board. I’ll clean this up, put in a new socket, clean the legs on the 6810 best I can and go from there. It may require a new 6810 down the line but that will be easy to change out. And yes, the two resistors in the lower left of this picture need to go.

All the corrosion will get cleaned once I pull the PIA socket and the ROM socket at U12. I’ll be following John Robertson’s procedure here: http://www.flippers.com/battery.html

Accept it, you may screw up! I did!

And now for my screw up. When I was pulling the power connector pins at J1 I got a little ahead of myself and did this:

j1 screw upwcaps.jpg

I pulled the plated through hole and pad out on pin 11! Ack! I was rushing because I was getting yelled at to come to dinner and only had 5 pins left to pull. Here’s what the damage looks like:

j1 pinwcaps.jpg

You can clearly see the plated hole stuck to the pin. Further evidence of the damage can be seen with the backlighting trick.

j1 back lightwcaps.jpg

This will also get a new eyelet for repair as it is a critical connection. Pins 11 and 12 on this board are for the 5v line! So as you are doing a repair, you try not to do damage but you must be prepared to deal with it if you do. If you are not prepared or not capable of dealing with any damage you might cause, best to send your board out to a professional.

The Electrolytic Caps

I removed the PIA socket at U17, the 2.2K switch matrix resistors, the resistors by the 6810 previously mentioned and lifted the 25uf cap by the CPU. The resistors and cap came up ok. Unfortunately I lost another pad at pin 3 on the PIA.

board mid strip 3wcaps.jpg

I lifted the cap because I wanted to test it with my new Peak ESR70+. This device tests capacitors for their equivalent series resistance – important for old electrolytics where they may have dried out and are no longer doing their job. The Peak can test in circuit or out and give you both ESR and capacitance readings. I pulled the 25uf cap because when I tested it in circuit, it could not give me a capacitance reading. I wanted to see if both electrolytics on the board were still good.

25ufcapwcaps.jpg

Above is a 25uf 16v cap by the PIA. As you can see it tests at nearly 36uf – almost 50% more than what is called for. However, older caps were commonly rated for +50% tolerance so with its low ESR, this cap is ok. However, modern caps are usually 20% tolerance so replacing this will be more in line with the spec called for on the board.

10ufcapwcaps.jpg

Above is the 10uf 16v cap. Its capacitance tests close to the required value for the circuit but its ESR is 15 ohms. This is consistent with the table in the Atlas manual for a “typical” capacitor of this value. However the manual also states that it is common for the ESR readings for good capacitors to be much lower than the “typical” values. If I were troubleshooting this board for not booting I would say this cap was not a problem. But since I am replacing so many components and I have to remount the daughterboard anyway, there’s no point in not replacing this as well.

I've tested these caps on about a half-dozen GP MPUs and they usually test fine. Generally, the 10uf cap's ESR is a little high for what you would like to see here. You can choose to replace it or leave it. Unless you have an ESR meter or see bulging of the caps, leaving them is probably ok.

After I remove the U12 socket, the reset daughterboard and the caps, it will be time to mitigate the corrosion.

The slings are 21-50003B coils. $7.50 from Steve Young.

http://pbresource.com/coilgam.html

You could try going with the old style Williams coils/mechs 19-400/30-750 @ 28v. The original GP coils were 23-375/30-800 @ 24v so you get a little more umph out of the Williams coils though GP runs a slightly lower voltage.

I'd have to look at the dimensions to see if the GP coils would fit in the Williams mechs. I've usually just cleaned them up and went with what I have. One day I'll have to look into xref flipper parts for GP as I've not seen it anywhere up to now and other than the coils, don't know anyone who sells it.

viperrwk

I'm not a pro. This is not my day job. I do this for fun and I happen to like and own a few GP pins.

I've helped a few people because I like to. If you want to PM me what you're looking for we can discuss offline.

viperrwk

Thanks vid - too lazy to go find that page.

The telling line about GP:

"No Williams coil equivalent as original coil's core diameter is vary large."

So you'd have to sub the entire flipper assembly with this:

http://www.pinballlife.com/index.php?p=product&id=711

viperrwk

Removing the Remaining Suspect Components

I removed all the remaining components that need to be replaced like the ROM socket at U12. Unfortunately I lifted another pad – due to the previous work on the socket it lifted as soon as it heated up. I also removed the jumper wire across the U12 socket and will make sure the board is cut and strapped correctly for 2716s.

U12 ROM wcaps.jpg

I removed the 12v reset daughter board and the 10uf electrolytic cap underneath it. Will put in a new electrolytic and reattach the daughterboard more securely.

no daugtherboard w caps.jpg

I also removed the jumper wire for the DIP switches on the back of the board. Now it is time to clean out any remaining solder and corroded metal.

DIP jumper wire gone w caps.jpg

Clean Clean Clean

Now comes the process of cleaning. First, I make sure that all through holes are clear of solder. This allows me to inspect each hole and make sure that if any are damaged I can identify them and repair them before reassembly.
First I cleaned up the through holes for the diodes on the DIP switches.

dip diode w caps.jpg

You can clearly see the damage to the trace for the diodes on switches 8, 16, 24 & 32 yet the diodes that are still in place look like they were factory installed. Perhaps this board was damaged and repaired at the factory before it went out. I will finish removing the diode at switch 24. The through hole for the bottom lead of the diode at sw8 will also need to be repaired with an eyelet. After insuring all holes were clean of solder and identifying which ones need eyelets, I started on the corrosion.

There are many ways to clean corrosion. Some people sand the board down, some bead blast it, some try to chemically strip it. I try to be as focused as possible and try to surgically remove as much corroded metal as possible with a xacto knife and then neutralize what’s left. My thinking is to neutralize what should remain and what I want to solder to. I don’t want to get into a debate on what’s the best way to remove and neutralize corrosion. All I care about is to make it go away. And since I'm not in a hurry I can take my time. This is about 45 min worth of carefully cleaning the pads at U8 & U9. Progress.

pad cleaning wcaps.jpg

Remember what it looked like?

j8removed.jpg

I also did some cleaning around the mounting pads for the battery. I’m most concerned for the small holes in the photo since this is where I’m going to attach the wires for a memory cap.

battery pads cleaning w caos.jpg

Interestingly there was a jumper on the back of this board made from solder for the positive terminal. Turns out there is still through hole connectivity and the trace is good on both the positive and negative side of the battery pads so it will be easy to run the memory cap to keep the high scores.

Now I can start to neutralize the battery damaged areas of the board with a vinegar solution using the procedure I linked to previously from John Robertson, then clean and dry. Take a final accounting of all the parts I need, make an order with GPE and while I wait, start installing the eyelets for all the damaged through holes – about 11 in total by my count – four of which I lifted, two of those were completely my mistake.

Further Damage From the Bad ROM Strapping

Before doing any neutralizing, I decided to look at the cutting and strapping for 2716 EPROMs. As previously noted, to set up a GP MPU for 2716s requires cutting two traces and adding two jumpers. One cut and jumper removes the ground from pin 21 on the ROMs and applies 5v. 2716s require 5v on pin 21 for read operations. The other cut and jumper bypasses U24 to keep the signal to U13 low on pin 18.

Whoever worked on this board previously did half the work cutting the ground trace for pin 21 on the EPROMs and jumpering it for +5v but didn’t cut the trace to pin 8 on U24 and jumper pin 9. The most unusual part of all this is that the jumper wire restores the ground to pin 21 across the EPROMs but with the jumper in place below U26, there is also 5V on pin 21. This board would never work configured like this and this is likely why U12 was bad.

After removing the jumper to ground, insuring the trace to pin 21 was cut and cutting and strapping U24, I discovered that pin 21 still had continuity to ground as well as the 5v line. While the cut looked good, to be sure I removed the 5v jumper below U26. When I did no more continuity to the ground trace, but what it means is that some other component is shorting to ground. Ugh.

The good thing about having so many components removed from the board is that it makes it easier to find these types of problems. You usually have to suspect previous work which after checking all of it, as well as removing any components that were socketed, showed the problem was still there.

I then decided to look at the remaining integrated circuits on the board. Many are 74XXs and can be easily tested with a DMM in diode setting. The procedure for this is outlined on PinWIKI:

http://pinwiki.net/index.php?title=How_to..._(solder,_desolder,_%22stitch%22,_test_transistors,_test_ICs,_etc)#Testing_an_integrated_circuit

This is a useful procedure to know for any board you’re looking to repair.
In trying to figure out where the problem was, I started at U16 – a 74LS138. This chip is the digit driver for the displays – and it stood out to me.

Seeing the legs of this chip blackened made it suspect and sure enough following the procedure above, I discovered that pin 14 was shorted. I cut it out and tested for continuity between TP2 & TP7 (5v & ground.) Still had continuity.

So time to check the other 74XX chips. U20, a 74LS379 used for solenoids was shorted at pin 15. Out it comes.

STILL continuity between 5v and ground. Test the 74154 at U14. That was shorted at pin 20 – out it comes.

FINALLY, no continuity at U12 pin 21 between 5v and ground – hooray! That makes the bad component count on this board 7: U9, U10, U12, U14, U16, U17 and U20. The good news is I believe that’s it.

Fixing Damaged Through Holes and Traces

One of the problems you're likely to run into when repairing any board is damaged through holes and traces. This board was no exception. Depending upon your skill set, tools and supplies, there are various ways to repair circuit board damage.

For through holes, the easiest way to repair damage is with a solder stitch. PinWIKI has a section on how to do this:

http://pinwiki.net/index.php?title=How_to..._(solder,_desolder,_%22stitch%22,_test_transistors,_test_ICs,_etc)#Repairing_traces_or_creating_a_.22solder_stitch.22

The method I use is called eyelet repair. It essentially replaces the through hole with a plated eyelet. Here is a video of how it is done:

Remember the 5V hole at J1 that I damaged:

j1 screw up.jpg

This is what it looks like after I inserted an eyelet. Front:

through hole after front.jpg

And back:

through hole after back.jpg

And after the connector is soldered in place, you almost wouldn't know there was an eyelet there (which is why I documented it so the next person who might work on this board knows what's there.)

through hole soldered.JPG

For traces, there are again several ways to do it.

You could use copper replacement trace and incorporate it into the board:

For shorter runs like this:

jumpers before.jpg

You can use cut off leads from a resistor.

jumpers after.jpg

Here for the jumper below U26 I used the through holes rather than solder to the traces on the top of the board. For the other damaged trace I use through holes on the board. Whenever possible you should try and use either through holes or solder pads for your jumpers.

For longer jumps like those for the diodes that span multiple connectors, using a solid core wire with insulation is the best way to go.

long jumpers after.jpg

You need to decide if you can handle these sorts of repairs or not before beginning any sort of board work. When in doubt, send it out (to a professional.)

Replacing Obsolete Components
You will find that some of the components on a GP MPU are no longer available/obsolete. To fix a board where some of these obsolete parts have failed will require substitution. Here is a partial list of replacement parts to use. I will try to update this as more parts become obsolete and suitable replacements are identified.

6551 RAMs at U6 & U7 - replace with 5101
This is probably going to be the most difficult part to replace if it is failed on your board. As mentioned earlier, Arcade Chips still lists these for sale on their site as of this writing but at $8/each. If it gets to the point where there are no 6551s available, you can replace them with 5101s by following the procedure that John Robertson has documented on this page:

http://www.flippers.com/gameplanTips.html

74C32 at U9 replace with 74HC32
This part is obsolete. It's designed to deselect the SRAMs during power cycling to prevent
memory corruption.There has been discussion about what to replace this with on RGP previously by some heavy hitters. Because it is in the battery-backed memory circuit, I subscribe to what Ed @ GPE told me after looking at the MPU-2 schematics:

"74HC32 draws the same power as a 74C32 and can be used. I would avoid using a 74LS32 (too much power) or the 74HCT32 (not designed for battery backup)."

Since I'm using a memory cap I don't want any part there that uses more power or isn't designed for battery backup.

74LS379 at U20 and U21 replace with 74F379
The LS379 is part of the solenoid address circuitry and is another obsolete part. The best available replacement for it is a 74F379 - Fairchild's equivalent to the 74AS. Both the F and LS are Schottky parts. More on 74XX differences can be found here: http://en.wikipedia.org/wiki/7400_series

74154 at U14 replace with 74HCT154
The 74154 is used here to drive the switch matrix, displays, solenoid select and reading the DIP switches.. The HCT154 is the best substitute both on the MPU board as well as the solenoid board, SDU-1 and any sound board that uses a 74154, SSU-2, 3 & 4.

1n959B Zener Diode replace with 1n4738
Almost forgot to mention this since I didn't have to replace it on this board. This diode is in the path of corrosion below U9. If it's damaged you can't get a replacement. It is an 8.2v, .5w zener diode. The most appropriate replacement is a 1n4738 which is also 8.2v but has 1 watt power dissipation which is fine.

I also previously discussed replacing the green disc decoupling caps with equivalent axial ceramics. And the two electrolytic caps on the board can be replaced with higher voltage equivalent value versions.

As of now, the other parts on the MPU board are pretty easy to come by and you shouldn't have a problem sourcing replacements.