Arcade PCB Repair Tips

This document describes tips for general debug and repair of older "classic" arcade PCBs. It assumes the reader has some familarity with basic electronics and some repair skills (such as soldering). I created these notes based upon my own personal experiences of repairing and testing older classic arcade PCBs over the last year. Much of this content is general common sense, but it never hurts to hear it again.

My primary website links:

Braze Technologies (Main Arcade page)
Technical Info (Arcade Tech Info)
Braze Technologies Arcade Products (High Score Save Kits, Multigames)

By far, the two most common failures I see with older boards are (1) broken traces on the PCB and (2) bad connections on sockets and edge connectors. When a trace is broken on the board or there is a bad connection, the failure symptom can be just about anything from the board being totally dead to just some intermittent flaky video problem. These types of problems can be very difficult to debug and at first sight can be very discourging. Hang in there.

The best first step in debugging a boardset is to first FULLY inspect both sides of the PCB(s). First look for scratches on the board that might have the possibility of being deep enough that the foil trace is broken. Most boards are manufactured with a solder mask (thin green film) that covers the foil except in the places where they want solder. Boards that have solder film are a little easier to locate scratches, as the scratched area will show the shinny foil through. Some of the older boards (eg. classic Atari boards) do not have a solder mask and it can be more difficult to locate scratches. The time spent inspecting the board will well pay for itself down the road, so take your time and be thorough. Use a maginfying glass for close up inspection of suspect areas.

For inspection, make sure you work in a well lit area. I like to sit at our kitchen table to inspect boards. I usually lay the board on a piece of cardboard to prevent scratching the table. I make sure I have the following tools handy for inspection.

If you locate a scratch or what appears to be a broken trace, the next step is to use your meter to check the continuity of the trace. Hopefully you can find the two ends of the suspect trace in order to probe with your meter to check the connection. If the scratch is on the component side of the board, locating the other ends of the trace can sometimes be very difficult as traces can "dissappear" underneath chips. If the scratch is on the solder side of the board you should be able to locate either the component pin the trace is attached to, or a via (a hole filled with solder that is used to connect a trace from one side of the PCB to the other).

If you really get stuck and cannot locate one end or the other end of the trace, you still might be able to make an electrical connection to it in order to test it. With a sharp exacto knife, very carefully scrape away some of the solder mask on the trace that you cannot locate the end. Scrape only enough so that you can get your meter probe to touch it (say 1/16"). Be very careful not to scrape so hard that you break the foil trace. Once you have scraped away the film, you should be able to probe the newly exposed area to test with (however, if the trace is indeed broken, you may still need to locate spots that you can solder to).

Once the two sides of the suspect broken connection are located, put your meter on continuity mode (I like to use the audio beep feature if your meter has it), and check to see if the trace is broken. If the continuity is good, then nothing needs to be done. But, if the continuity is bad, first thing to do is double check that you are probing the right connection. If you have the schematic try and locate the trace on the schematic and see what other components it might be attached too. Then with your meter probe on the component side to double and triple verify the trace is indeed broken.

Lets assume you found a broken trace. It must now be repaired. I prefer to repair these by soldering on a new wire that in parallel jumpers across the broken foil. Note the new patch wire may or may not follow the same physical path of the old broken foil and the solder connections of the new wire may well be several inches away from the damaged section of the foil.

For repairing the trace, I prefer to use 30guage wire-wrap wire (available lots of places including radio shack - one spool should last a very long time). This wire is very fine and is a single solid strand. The best place I like to to solder in the patch wire is at a via. If a via is not available, the next choice is to solder the wire to a component that is aready soldered to the board. Note you do not want to solder the patch wire to a removable part such as a socketed IC, instead, solder to the bottom side of the socket instead. I generally prefer to install the patch wire on the same side as the broken trace, but sometimes you cannot do this, especially if the broken trace is on the component side of the board. if the broken trace is on the solder side of the board, you should always be able to also patch it on the solder side.

Cut the patch wire (30ga) to be almost the exact length needed. You do not want the wire to be much longer than needed as the extra wire will get in the way. The patch wire will later need to be either glued to taped to the board. If the wire is going on the component side, it might need to be a little longer in case it needs to bend around components. Remember the patch wire will eventually be placed flat to the board.

Strip both ends of the wire only a very short amount, like 1/16". Prior to soldering the wire, heat and suck the solder out of the hole of which you are soldering too. This applies no matter where you are soldering to. If the connection is to a via, this will leave plenty of room to get your patch wire inserted. If connecting to a existing component, hopefully with the solder removed you will be able to insert the wire into the space around the component and the hole. If not, you will need to hold the wire in place touching the connection while you solder it. The solder should be enough to hold the connetion physically and once the patch wire is later secured it should not move or come lose.

Repeat same step for other side of the patch wire. Now flaten the patch wire to the board curving it and/or bending it around components so it sits flat to the board. Then secure it to the board either with tape or glue. I prefer to use my cheapo hot glue gun and put a glob of clear glue at several places along the wire. Make sure you hold the wire flat to the board until the glue dries.

Finally, you need to check your work. With your meter, again probe the connection, it now should have good continuity. Next, inspect the area where the original trace was broken. It is common when a large gouge or scratch occurs that the foil of one trace might short with another trace. I like to inspect these with a magnifying glass and also check for short circuits with my meter. If one trace has shorted with another trace you will need to clean this up. To unshort them, I use my exacto knife and score the board between the traces to make sure they do not short against each other. Then retest with the meter. Also don't forget to inspect your solder joints. The re-soldered connections should not be shorting against any adajacent connections and the hole should be filled with solder.

The second most common failure is bad connections, primarily IC's not making good contact with the chip sockets, or chips that have been re-inserted and indavertantly have a pin that got bent. Inspecting the socketed components for bent pins that are not properly inserted in the sockets in pretty easy and straight forward. If you find one, remove the chip, straighten the pin and reinsert.

Anytime you are removing socketed components such as eproms or microprocessors, there are a couple of things to watch out for. First, be static conscious. I must admit, this is one area I probably could be safer at, however, to my knowledge I've not damaged anything... well not yet... By static conscious, I mean don't work on your board in a area that is susceptible to static charges, such as carpet floors or plastic chairs (like those white patio chairs). Probably wearing a sweater or other static type clothing is not advisable either.

When removing the chip, say for example by prying out with a screw driver, watch out! There are a few things can go wrong. First, if the socket is one of the cheaper quality ones, it may have hollow openings underneath the chip. Your screw driver can get in there and when you pry with it you will very likely scratch the PCB and damage the traces, which will put you back to the top of this document to repair them. With those same type sockets, the little plasic support bars that separate the hollow spaces can easily break. Another thing to watch out for, is if the chip is really stuck in the socket, you might actually lift the socket from the PCB rather than lifting the chip. This can damage the board and also can break the foil trace around the socket. Note, I learned all of these problems the hard way :(

Many times when I have a board that is just acting flaky and inspection has not proven anything, my next step is to reseat ALL of the socketed chips and then retest the board. Sometimes the board will magically work, other times it might get slightly better. In any case we want to rule out bad connections before we get into the dirty details of debugging the specific board.

If the board is part of a multiboard set, disconnecting and reconnecting the boards is also a good idea in order to make sure there is good contact. Also, the edge connectors can also be cleaned at the same time. I just use a small piece of scotch bright. Again be static conscious. I have heard that some people use steel wool, but I think that can leave very small wires on your board which could cause a short (or maybe they will burn up?), anyways, I personally stay away from using steel wool.

Other things to check for while you are inspecting the board include burnt edge connectors, burnt components, broken components such as capacitors and resistors, and of course missing components. Note it is very common for boards to have unpopulated chips, so don't assume any empty socket means a missing component. Check the schematic. Also, TTL designs require that there be a capacitor (usually a disc capacitor) every so often (sometimes as often as each chip), and they need to be nearby each chip. Anyways, it is common to see one or two of these capacitors broken. I usually don't worry about this right now, and usually don't replace them. However, for broken disc capacitors, I usually just remove the entire capacitor and mark the area to remind me that I removed it. This prevents a possible short-circuit which can occur if the broken disc capacitor happens to have the metal plates inside it touching each other.

Another good habit, is to mark or flag any suspect components you locate, so that when you come back to the board later in time you can easily pick up where you left off. I use small pieces of masking tape to identify suspect areas or components I think might be broken, or missing components, etc.

Hooking up power to the board. Now that you have fully inspected the board, repaired any broken traces, reseated the socketed chips, and replaced any obvious broken components, we are ready to powerup the board (assuming you have the pinouts). Two final steps I do just for completeness before I powerup a board is to (1) make sure GND and +5V are not shorted, and (2), double check that GND is GND and +5V is +5V. I do this by checking continuity between GND and several GND pins of ICs (GND is most often the bottom left corner of the TTL chip - pin 7 on a 14 pin device, pin 8 on a 16 pin device), and repeat for +5V which is most often pin 14 or pin 16 (upper right corner). This is just a double safety check.

I usually hookup just GND and +5V initially. I want to make sure there is no shorts and if the board happens to have an LED, check to see that it is on. This also gives a chance to check for any abnormally hot components. I might leave it up for a minute or so and then touch the chips. My (switching) power supply will automatically shut off if there is a short. I always wait until later when I'm ready to test sound before I hook up +12V and/or -5V.

Assuming no hot chips, next I use my logic probe to check out a few things. The video sync output should be pulsing and various pins on the processor should be pulsing as well as most pins on the eproms should be pulsing. This just gives you the sense that the game is minimally running.

Next I hook up video sync, and RGB (make sure you also have your video ground hooked up to your monitor. If there is no video displayed, make sure you have things hooked up right. My most common mistake is counting incorrectly on the edge connector, or counting from the wrong end, or hooking it to the wrong side of the board. Another gotcha is that several of the published pinouts, say for example in the spies wiretap data base, are wrong, or possibly the boards them selves are mislabeled. Make sure pin 1 is pin 1, and if there is a key in the edge connector, make sure it matches with your documentation. I have seen many times where the documentation will say side-A or side-B, only to have my board be labeled backwards or not labeled at all. I have also seen pin 1 in the documentation be lableled pin 28 (for example), on the board. Thi s is why I like to probe the video sync signal to make sure it is pulsing. Another video problem can be making sure the game is providing composite sync, as opposed to separate horizontal and verical sync lines, and of course making sure these match with your monitor connections. A few games will have switches or jumpers for negative versus positive video sync, but by far the most common is for the game PCB to provide negative composite sync.

Assuming the video is good, next step is to hook up sound. This usually requires hooking up +12V and sometimes in addition -5V, as well as hooking up the speaker ;-)

Things to watch out for. Some games have the volume control as an external component to the PCB (typically so it can be wired near the coin box for easy access). If the board does not have a volume control, you probably will not get any audio output until you wire in the control. Also, some of the classic video games, such as many of the old Atari and Nintendo games, require an external audio amplifier. This means the board itself cannot directly drive a speaker.

To allow me to test boards that need an external amplifier, I bought a simple audio amp kit from Jameco and assembled it. I then wire it in series between the game and the speaker, It also needs power connections. The sound quality might not be the greatest, but it is sufficient for testing purposes. Make sure you get one with a volume control.