O. OK. Now I think I know what you mean.
But first, what do you mean by "errors".
Are the switches reported to have an error when you enter diagnostics?

You can try this.
Get a jumper with alligator clips on both ends.
Game OFF.
Connect one end to the white/gray wire on the rocket kicker switch.
Connect the other end to the white/gray wire on the lower right 5 million target.
Make sure those connections are correct and not touching anything else.
Power on.
Test the switches.
Report back.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact
http://www.PinWiki.com - The Place to go for Pinball Repair Info

Can you snap a picture of the wiring on this ramp switch? It bound to be wired to the wrong side of the diode, diode backwards, etc.

Im a noob, can you explain why you suggested this as a test. Talk to me like im an idiot.
Thanks.

We all start somewhere Rick. And we learn by asking questions. Good for you.

The test I prescribed will connect a known working on a switch in that row to a problematic switch in that same row. If the problem has been caused by a break in the daisy chain wiring for that switch row, this test will tell us. If the switches (or at least more of them) begin to work, then we know that there is a discontinuity. This is an old operator hack.

An equivalent test would be to set a DMM to continuity and "buzz" from a known working switch to a problematic switch. If no "buzz" the wiring chain is broken.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact
http://www.PinWiki.com - The Place to go for Pinball Repair Info

Thanks. I guess i missed parts where we knew some of the switches worked.

I would have expected
18 to 28 test
18 to 38 test
18 to 48 test

Etc.

Glad for you taking the time to explain. Happy to learn from the debugging.

OP, do the tests that Chris suggests before rushing to replace anything. He had me do the isolated CPU board switch test and I found that the column of switches that was out on my game was due to the ULN2803 on the board and not a broken wire somewhere(that I never found, like your situation).

Systematic testing first, leads you to what to replace at the end.

Your switch is wired wrong compared to mine.

Banded side of diode is on the right most lug and has a green wire on that lug. White wire is in the middle. Other side of diode is on the left.

20161109_210439 (resized).jpg

It is not necessary 'wrong' but it is different.
Switch on picture in post #57 has standard configuration GRN-column on 'Normally Open' where switch in post #58 has WHT-row on 'Normally Open'.
Both are correct as long as the diode is pointing to GRN when switch-circuit is closed.

Hey mamemaster, I just read this whole thread. What a PITA! Anyway, if it were me I would check continuity between adjacent switch lugs along the entire switch row that you identified as having the issue. I suspect you have a broken wire somewhere underneath the sheathing that you cannot visually see. I would use alligator clips so you can clamp to the switch lugs and wiggle the wires between the adjacent switches and see if the audible continuity test on your DMM intermittently cuts out.
My other thought is that you pinched a wire under a ramp when replacing them, creating a short. Did you already try removing the ramps and reinstalling them? Test for the switch errors with the ramps out.
Good luck and please keep us posted!

Do the test Chris Hibler suggested.

I just see, that one of the clock-optos is in that row - SW98 clock hour 4.

An often overseen problem is, that if a phototransistor is defective in the way that it conducts permanently, phantoms switches MUST occur. This is because then it behaves like a closed mechanical switch WITHOUT a decoupling diode.

If it were my machine, I would disconnect the clock board and look, if the fault stays or has gone. If it is gone, you have found the fault, a defective opto on your clock board.

Good suggestion, especially since the clock was probably monkeyed around with when removing/installing the ramps.

A "normal" switch matrix status for TZ is shown here.
Compare to yours. Analyze diffs to see if they make sense.
http://www.pinwiki.com/wiki/index.php?title=Williams_WPC#T.1_Switch_Edges_Test
--
Chris Hibler - CARGPB #31
http://ChrisHiblerPinball.com/contact/
http://www.PinWiki.com - The new place for pinball repair info

Did you try buzzing between like colored wire solder tabs in the suspect switch row?
The problem almost HAS to be the wiring.
I've seen instances where a switch matrix wire will connect to a ramp switch via a .062 molex connector. Two wires are crimped into the same .062 pin. One wire makes, the other wire doesn't. Buzzing from solder tab to solder tab of the white/whatever wire will tell the tale.

--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact
http://www.PinWiki.com - The Place to go for Pinball Repair Info

Any updates, Mamemaster? I have been following this thread due to a deep interest in understanding what went wrong. Wish I had some possible test or fix to contribute, but everyone has hit the things I'd suggest. This is a good mystery, but I am sorry that your TZ has been down for so long.

Chris is right, it's a wiring problem, back away from the boards.... You've got something not right somewhere on a switch on the playfield...
It may be a backwards diode, a loose wire, a misconnected wire, only one wire connected where two should be, etc. But it is a wiring problem since both the column and row work on the WPC board.

I think we've all missed something. Let's go back to the beginning...

Power up the machine and tell us if there is ANY diagnostic messages upon power up and what they are (even if it is clock not set, we need to know, gumball disabled would be really important, but list every single one to us). There is also the possibility that you've set the incorrect number of gumballs on the machine, in firmware 9.4 you can change the number of balls installed in the machine to 5, 6, or 7 but it as listed as 2, 3, or 4 gumballs in the menu, if the setting does not match the number of balls you put in then the machine gets very confused.

Let's get some information...

(SEE NEXT POST FIRST, ROW 8 OBVIOUSLY HAS A BROKEN WIRE, NO SENSE TESTING GUMBALL YET)

Using the test menu, empty all balls from the machine.
Start filming with your camera. It will be important to see the DMD and the playfield when testing, and the machine must be EMPTY of ALL balls before starting the gumball test.
Go into gumball machine test and then put all 6 balls, 1 at a time SLOWLY, pausing between each ball, with the DMD in view, into the drain on the playfield, start the gumball test and let it run for 7 to 8 minutes ( it may take 3 or 4 cycles for it to mark a failure) filming it and the DMD the entire time...
Post video, If it fails, note what the failure message is, if it doesn't fail, post video and we can spot any abnormal behavior...

I just watched this video, YOUR OUTHOLE SWITCH IS BROKEN.... All the balls you see below the flippers should NOT be there, it doesn't know they are there, so it does not move them to the trough, and since they are not in the trough it is not counting them.

The reason it did start the game, and moved a couple over to the trough and then to the shooter lane is the ball search you started upon starting a game fired the outhole kicker, which move two balls (due to the way they had landed) into to trough.

The first thing you must do is fix the outhole switch. NOTHING ELSE WILL WORK CORRECTLY UNTIL THIS IS DONE! The gumball test would have eventually shown this as a fault on the 3rd run...

There is no sense troubleshooting any other part until we fix your outhole switch....

P.S. When I say "Broke", I imply that the machine does not know the switch is closed, this may be a bad switch, but more likely it is a wiring problem and the two wires from the trough switch do not make it back to the board. Since this is the farthest away switch on the playfield the two wires that go to it are connected to 7 other switches each to make it to this switch, each one of those jumps must be good for this switch to work.

Watching more you rocket kicker switch is dead too, your left inlane is reporting an error in diag, and the slot switch is not working either, you've disconnected the white-gray wire somewhere.

Use switch EDGE test and test each of the row 8 switches, I bet you half of them don't work... Your problem is on one that DOES work; on one that DOES work, there should be two white wires, and only one is connected....

This is where the alligator clips and jumpering come into play...

tzrow 8 (resized).PNG

tzswithc (resized).PNG

It's probably still not a MPU board problem, it's a diode or wiring problem that is intermittent (the funnest to track down).

If I'm guessing what you did correctly your J206-8 to J208-6 should NOT have produced anything, since those are both the connector keys, meaning they are not wired in. You can look at the switch matrix I posted above and see that those two pins are not listed in the row or column labels telling you where the connector ends and what wire color it is.

Clay is right to hit the playfield to see if there is an loose wire, the problem is all of row 8 are normally open switches, that are only closed normally during play. You actually did make a change when you hit the playfield, you've got at least the outhole switch and slot switch to work now, you just didn't see it, since there were no balls there when you hit the playfield.

The reason the gumball wouldn't load is the outhole switch, so let's keep focused there.

We have to figure out why it now works, and then sometimes doesn't. Weather there is a loose white/grey wire on one of the row 8 switches or it's broken going back to the MPU at J208-9. Since we want to outhole to work, Doing what you did between J206 and J209 with the jumper, short J206-1 to J206-9 and the machine should say the outhole switch is closed in the edge test. If it does that and you can not make it not say that as long as that jumper is there, then it is 100% a wiring problem.

Now you just have to look at all 8 switches in row 8, check, pull, tug, etc. on both white/grey wires on each switch in row 8, if you're lucky you'll find one that come off, and that could be your only problem, fix it and then check all of row 8 works and your done and working...

The other possibility is that you have a diode backwards on a switch somewhere, and when that switch is closed it causes row 8 not to work. This is a cause and effect problem, the switches in row 8 don't work ONLY when some other event is or is not happening. This could be any switch changing state, it may be open that causes it, it may be closed that causes it, but that other switch shorts out row 8 when something happens to it. This would be a good reason to disconnect the clock (do this while doing the test in the next paragraph, to make sure the clock doesn't effect the outhole switch), since it does have a row 8 switch in it, and the connectors on the clock have been known to do some flaky things...

We need to know what state ALL the other switches were in when the outhole switch is working. If it is currently working, go to edge test, and make sure it closes with a ball on it, then remove every other ball by hand from the machine but the one in the outhole. If the outhole switch changes state when removing another ball, we know we have a suspect switch under that other ball. If you can remove every ball, then close all 54 other switches by hand (well ignore the 8 in the clock for now). If one of those cause the outhole to change state, we have a suspect switch. We then have to figure out how that switch is effecting the outhole (it could be because of wrong wiring or a bad diode on an entirely different switch then the one that cause the symptom, but we at least have a place to start).

We have to figure out what you did to make the outhole (and slot, and rocket kicker) switches work again. Was it the hit to the table (meaning probably a loose wire), or is it because you have the balls resting in certain places which is bypassing a wrongly wired diode.

column (resized).PNG

My guess at this point is it's the diode that was wired wrong on that ramp switch shown in post # 57 & 58 or a faulty connector pin on the switch connectors on the cpu at j206-j208.

Thanks for all the advise you guys are giving. I have had some weird issues to deal with as well and your ideas are great.

Did you already check your +12Vpower and +12Vdigital on Power-Driver-Board at TP 1 and TP3?
You know, especially the +12Vdigital is the important voltage for the comparators for the switch matrix rows. We all know the problems with weak +12Vdigital, if these are weak, you switch matrix can become crazy.

At least its a try, to first check the +12Vdigital. Best is to measure it under load, means mesauring the voltages when the all lamp test is running. Imho my few cents ...

I am also suspect of this, this game has many mods jumped from various sources, I would unplug as many as I could and see if it makes any difference, or check the voltage with a meter if he has one.

If all these issues happened after you installed the ramp have you tried just unplugging the ramp switches under the playfeild and see how everthing else acts? Maybe the switch got messed up inside from heat when you soldered the wires on.

I just don't see how a single switch or diode will cause other switches in the row to NOT operate. The only thing on the playfield that can take out all our part of row 8 is wiring, or a ground short, and you would know if there was a ground short.

If you were to hook up the wires incorrectly, say with the diode backwards between the row and column and put the row wires from the MPU to this switch and the other row wire that continues onto the next switch in the row in the right configuration (and I've seen it done and had to troubleshoot it), on a switch in the right way where the column is on Normally closed of the switch and the row wires were split between common and Normally open, then the downstream switches in the row would only work when the switch wired incorrectly was closed, otherwise they would always show open, and you would get phantom closes on other switches on the same column when the switch was open.

By incorrectly splitting the bridged wires in the row between common and either normally open or normally closed, you now have made a switch, that dependent on weather there is a ball on it or not, now causes the rest of the row to function or not to function. The same can happen for the columns too, or can also cause it to be even weirder when the diode is on the wrong two legs of the switch or backwards; combine a backwards diode, and a split row wire, and you really get a hard to troubleshoot mess.

It is easily possible, for one switch to effect the entire row, and to cause phantom problems all over the matrix. That's why when you get really strange random closures in the matrix, you start looking at the box patterns in the matrix. The switch with the wrong wiring usually forms a "box" on the matrix with the switches it is causing to malfunction.

You need to NOT do this in game mode, but run all the tests, very methodically, in switch EDGE test mode...

Since we know something changes state to make the switches start working, note where the balls are when the switches were working in game play. Remove ALL the balls from the machine and while in switch Edge place one ball in the outhole. If it doesn't work, add balls back by hand to the places they were, always keeping your eye on the box for the outhole switch #18. When you see it change state to working then we have something to look at. Your going to have to do one step at a time and in a test mode that doesn't allow many different things to change at once. When it game play your fighting ball searches, ball movement, a different number of balls in the trough and the gumball, and a number of different switches closing and opening, all of which can cause the issue.

Since the rest of the rows seem to work, we need to stop changing other things, the replacement of the diodes on every switch (meaning you've changed the wiring on every switch, and even if you think you got the all right open the possibility of change), the driver board, the ramp, and the mods, all add a second or third or forth thing that can be stacked up and going wrong, the jumper test also proved the MPU is fine, and the driver board doesn't effect switches, so there's no point changing it.

You need to focus on one thing and that one thing is row 8, and specifically the outhole switch. Lets find its problem first... You need to know what does and doesn't make it work. Shotgunning the machine and changing boards is only adding to the questions and not narrowing down the issue.

Can you provide a close-up, in-focus picture of the row connector?

This sounds more like a switch matrix error than a game compensation. The only way to award a skill shot is to actually make the skill shot. In switch edge test, try this same scenario. You'll see more than one switch reported.
--
Chris Hibler - CARGPB #31
http://ChrisHiblerPinball.com/contact
http://www.PinWiki.com - The Place to go for Pinball Repair Info

Mine does the same thing. Extra dot when pressing the 5M in test and clears when pressing the inlane.

image (resized).jpeg

You can ignore that extra dot on the very left, it's a screen draw issue with an out of place pixel, that code flaw is in most WPC games. It means nothing.

The extra rocket kicks could be either the rocket switch didn't open (which would match your theory, but may not be true), a full ball search, an after shoot ball search (there is a mini ball search that happens after a ball is shot, to clear the rocket and slot), or a couple of other things.

Also to clear a previous post up, a orange (lowest) skill shot is awarded if the rocket switch, slingshots, or hitchhiker switch goes active after the ball shooter switch is deactivated AND the slot switch or slot prox is NOT activated first (the machine compensates the skill shot since the ball didn't make it to the slot).

I've reattached a switch matrix that is color coded to help out... What is inside the red boxes are what shows on the switch matrix in edge test (Column 9 does NOT show on the Edge test screen, but you need to know it is there).
Rows are in gray and are white wires on switches, columns are in green and are green wires on switches.
Dark Yellow boxes are Opto switches, and are normally closed when NOT active, meaning they will show as a box when there is NO ball upon them, and as only a dot when there IS a ball upon them (meaning the ones in YELLOW show backwards from the rest of the matrix).
Light Yellows are Flipper button opto switches.
The switches with Blue on them are the switches that are normally closed when the machine is in switch test and there are ZERO (0) balls in the machine. If there are at least three balls in the machine, and the next ball would NOT be a powerball, then the switches in purple would also be closed (If the next ball is a powerball or the trough is empty then 26 would be open.)

The last pic of your switch matrix shows it in really bad shape. I really need to know where the balls were and what switches were taped down... To get the matrix looking like that you would need:
4 balls to the right of the outhole in the trough, the rightmost one being metal.
Al least 1 ball in the outhole
A ball in the shooter lane
a ball in the rocket kicker
a ball in the slot machine
a ball stuck in the left outlane,
a ball stuck in the right outlane,
two balls stuck in left inlane,
a ball stuck in the right inlane.

So there needed to be AT LEAST 13 balls in the machine for that kind of display on the switch matrix, and at least half of them stuck in the inlanes and outlanes (not likely).

I'm also curious, if you really taped down all the row 8 switches, then the picture does NOT show 4 of the row 8 switches as active, but shows a lot of switches that shouldn't have been active as active (in ORANGE on attachment), if you can tell us the state of the balls and which switches were taped down when you took that picture, we can start narrowing down the issue (take a look at the TZswitchwrong attachment, if it is in Orange, it is an active switch in your switch matrix that should not have been, and that's assuming you had 3 balls in the trough with the right most being metal.)

We also need to know if the clock is or is not hooked up currently, and where it is disconnected if not hooked up.

You need to stay in switch test, especially if that is what your matrix looks like currently. We can find the problem if you stay OUT OF GAME MODE, and we methodically change the state of ONE switch at a time, and look at the ENTIRE matrix for changes (since all of those switches could not have really been closed when you took the picture).

If the matrix still shows what it shows in the picture, and you can tell us where every single ball is, and what switches you have physically closed (by tape), we can form some ideas from there of where to look for the issue.

P.S. if you have the third magnet mod, let us know that now too.

Also what has the clock been doing, and are there ANY ERRORS WHEN YOU GO INTO TEST MODE?

tzswithc (resized).PNG
tzswithcwrong (resized).png

This isn't my game and I don't have this problem, but you deserve a major thanks for this post. Incredible amount of work and time to put the images together, describe them, and detail what to test.

Mamemaster: have you tried testing/playing the game with the clock unplugged AND disabled from the options menu?

What switches did you have taped down?
Were they all in row 8?
Also were any of the 6 balls you had loaded in the gumball machine? Or were they all out on the playfield, where exactly (down to the ball and position in the trough right to left?

I've taken your picture, and applied it to the switch matrix, just for reference...

The reason to stay out of game play, is that things like the gumball geneva switch, ball poppers, the shooter lane, etc, are changed during ball searches, and if one of those things is the problem, then allowing the machine to physically change the state of those switches means we have to start troubleshooting from a different place each time. Thus, too many variables are changing to troubleshoot easily.

tzswithc (resized).jpg