Over in the Taxi Club thread, someone mentions that he sometimes has weird issues when turning the game on for the first time in at least a few minutes. I'm also having the same issue. When I turn it on after it's been sitting for longer than a few minutes, there's a 70% chance that the only thing that happens is the bell starts going off.

I'm assuming that the way the bell works is that one of the transistors sends a signal to the bell solenoid, but it's wired in such a way that when the arm goes out to strike the bell, the circuit is broken until the arm comes back, which closes the circuit again and sends the arm out to hit the bell again. This continues until power to the solenoid is cut. This leads me to believe that either the transistor that handles the bell is starting to go out, or the chip that tells the transistor to fire the bell is going bad.

Anyone how know I can go about testing this? I suppose I could power the game on, let the solenoid lock on, and test the chip that drives the solenoid. Perhaps I can cut the wiring to the solenoid to reduce the risk of it burning up, and I can just test the voltage to see if it's otherwise locked on.

It not just a single 'ding' of the bell. It's the full-on bell ringing that happens when someone wins a jackpot or starts multiball. If this were any other coil, the plunger would presumably be locked in position until I turn the machine off and back on.

In fact, the fact that turning it off and back on makes it work properly definitely makes me think it's some kind of logic issue rather than a simple bad transistor.

Yep. This was a home-use machine for a large part of its life and its previous owner was savvy enough to replace the batteries regularly. Not a spec of corrosion on them. I mean, unless there's something really miniscule going on.

Along with the bell-ringing, the displays will also show gibberish with it.

I haven't left it on for a long period of time, but the game will sit like that for at least five seconds before I lose my nerve and turn it off.

This was happening before and still happens after a capacitor and high-voltage rebuild on the power supply. The fact that the display freaks out, the three red LEDs don't light up in the same configuration as they do when they boot normally, and that everything works just fine after a quick power off and back on makes me think that there's something going on with a logic or ROM chip somewhere.

I wish I knew if it was safe to leave on while I test things out with my logic probe and DMM.

I did some investigation around the blanking signal and learned that the chip at U43 is at least partly responsible for the signal itself.

"A way to test the blanking signal is to boot the CPU board with only +5 volts and ground (no +12 volts). This will keep the reset line low, and hence the blanking signal should also be low. If blanking is high there is a problem in the blanking circuit. Check the U23 555 timer chip and Q50 2N4403 transistor - these are the guts to the blanking circuit. The signal itself can be seen at U43 pin 3."

I looked at my Taxi and found that there is no chip at U43. Whaaaaa? So I looked at my Pinbot. Definitely a chip there.

Sometimes the system boots right up without any issues. Other times I have to turn it right off and back on to get it to work. How is any of this possible with a missing chip at U43?

IMG_1785.JPG
IMG_1786.JPG

Taxi has a knocker. I'm pretty sure this isn't some kind of alarm to tell you that something's wrong. And if it's firing on purpose to let me know that something's wrong, why does it only sometimes ring and not others?

It's been doing this since I first bought it at an estate sale a few weeks ago. A large majority of the time, powering on the machine for the first time in a while causes the the bell solenoid to lock on. Additionally, the displays may freak out and show gibberish characters. The diagnostic light never comes on, but the +5 and Blanking LEDs stay solid. I know that the blanking light is supposed to start low and then go high really quickly, but from what I can tell, when the machine powers up in this state, it's high all the time. The machine is unplayable in this state. It doesn't boot, you can't play a game - nothing.

When the machine is in this state, a quick power cycle almost always puts it into a working state. It boots, and everything is just peachy.

Now, I know that Pinbot is System 11A and Taxi is System 11B. I also have a Road Kings which is a plain System 11. All three of these machines list a 555 timer IC as being in spot U43. Whether or not it's socketted is not the question. The questions is why Taxi is able to boot at all with that chip completely missing. Perhaps the machine is able to luck into a stable enough state for the machine to boot on first try. Most of the time, though, it takes a quick power cycle for the CPU to get to a stable enough state to boot right away.

I'll pick up a 555 chip and see if the problem goes away.

I should add that it's been a few weeks since I completely stripped the playfield out, so the bell solenoid is the only one that's actually plugged in. It's possible that were everything to be plugged back in, similar solenoids would also lock up. I don't recall the pop bumpers or kickers locking on at the same time as the bell (that is, none of the special solenoids) but at the time, I wouldn't have thought to look at the upper right kickout. I certainly don't remember the catapult locking at the same time as the bell. It's weird.

OK, I went out and bought the NTE equivalent of a 555 timer chip, the NTE 955MC. I plugged that bad boy in, and while it does stop the bell solenoid from locking on power-on and allows the machine to boot straightaway, it kills my display, and the blanking LED never comes fully on.

Now, the playfield still isn't wired in, so all I'm really able to do is do the music and lamp tests. This works both with and without the 955MC chip in U43, but the display is rendered almost completely blank. A few digits come on, but they're not necessarily the right ones, and they're really dim.

I've checked to make sure that the socket the previous owner put in was put in properly. No shorted pins in the socket.

Can anyone give some insight into why adding the (presumably correctly functioning) chip would cause issues with the display? In the meantime, I'll take the chip out and try working through the schematics.

Well, I'm officially out of ideas. I removed the socket from the equation, cleaned up the trace pads and soldered a brand new 555 chip in there. It's an NTE 955M. I've also tried a different NTE, the 955MC. I've never had any problems with NTE's components in the past, so I don't have any reason to doubt these ones.

The machine no longer locks the bell solenoid on start up. But with the chip in, the machine sometimes doesn't boot at all. The +5 LED comes on, but the blanking LED never lights up and the diagnostic lamp never comes on, let alone starts blinking.

On the times when the machine is able to get past that point, the displays are dim, and only two digits in the top and bottom rows come on. The display test shows that they can cycle from 1 to 9 and when it's their turn to light up all the segments in their digit, they do so just fine. I'll have to check more thoroughly, but I'm not entirely sure they show the characters they're supposed to in other situations, though (such as during the music test).

Moving around in the diagnostic menu seems to work. I can trigger the music test and cycle through the songs. The lamp test, though, make all the lamps that are currently plugged in (just in the backbox) blink, but really, really dimly. I'm not sure the solenoid test actually produces any pulses at all, but since the bell is the only one I have attached, I can't really say for sure.

This is, by far, the most complicated problem I've had to fix on a machine. Is there anyone who can give me some insight as to why adding a 555 timer chip would cause such a collection of failures? Why would the blanking LED light on fully with no chip in U43, but not at all (and a host of other problems) when there's one there? Could I be using the wrong chip? Could there be some other centralized component causing the problem that gets bypassed when the 555 isn't present?

Agreed. I suspect that unless I suddenly get smarter, this will go unresolved by me as well.

I've worked through a cached copy of Clay's Pin Repair guide for System 11 that I found on the Interwebs. I diode tested all the relevant pins on all the PIAs. I've checked the voltages on the power supply and all around the CPU board. There's nothing out of the ordinary at any of these pins or test points.

I'm not sure what WAG means - but from what I can tell this board is exactly what's it's supposed to be. I haven't gone through and checked the rating of every single resistor compared to what's in the spec, but that would literally take hours.

So to recap, what we have is a machine that boots pretty consistently, but with the following symptoms.

1) The blanking LED never lights up, but the +5 LED does and the Diagnostic LED blinks steadily.

2) The score display is dim, and only four digits show up. I don't remember the exact digits, but it's something like 5 and 11 on both the top and bottom rows. Running through the display test works on those four digits. They cycle from 0 to 9, and when it's their turn to light up in the digit-by-digit part of the test, they come up just fine. All the while, these digits are very dim compared to what they should be.

3) The lamp flash test results in really weakly flashing lamps.

4) While there is currently only a single coil attached to the machine, that coil never fires during the coil test.

5) This whole thing started when I added a 555 timer chip (specifically, the NTE 955M) to U43 on the CPU board to replace one that was missing when I bought it. Prior to adding in this chip, the system worked fine, but would lock the bell coil on first power-on. The coil would unlock and the game would boot fine after a quick power off and back on.

So the question is this: How can a single 555 timer chip added to try to remedy the coil-locking problem cause such a chain of other failures.

I guess this machine is down for the count unless I'm willing to permanently remove the 555 timer chip and deal with having to quickly power cycle the machine when I want to play it. Any help anyone can provide would be great.

I'll take a look. But keep in mind that a malfunctioning display is only part of the problem. The lamps also flash really weakly and the solenoid test never fires the sole coil that is currently attached to the machine. All three of these issues go away when the 555 is removed.

It's like there's some kind of power drain happening when a 555 is put into spot U43.

What does this entail? I don't have any test ROMs. Do I need to buy some?

I totally want to learn how to do it. I'm just wondering what tools I need. I can get a power supply. Where can I get the "special eprom"? Also, what do you mean by lifting the leg of a PIA? Do you mean unsoldering it and somehow prying it out of the hole while leaving the other 39 legs in? Which leg do I lift?

I appreciate the help, but it sounds like you're using shorthand intended for someone who knows way more about working with electronics than I do.

OK, pulling out the 555 gives the machine a chance to enter one of two states. I'll refer to them as "broken" and "working".

The "working" state has the appearance of just that - everything's working. The +5 and Blanking LEDs are solid, and the diagnostic light is flashing. Everything works. Before I stripped the machine, I was able to play a complete game and everything was peachy.

The "broken" state is one where the diagnostic light never flashes, the bell solenoid locks on and the +5 and Blanking LED are solid. (There's also a "Broken State B" that I've only encountered once, which has a minor variation. I'll cover those variations at the end.)

I checked all the empty pin spots when the machine is in the broken state.

Pin 1 (Ground) has zero voltage and read low on the logic probe. This is presumably because it goes straight to ground.
Pin 2 (Trigger) has 3.3 volts and reads high.
Pin 3 (Output) has 4.5 volts and reads high.
Pin 4 (Reset) has 4.9 volts and reads high.
Pin 5 (Control) has zero volts and doesn't read at all on the logic probe. This could be because it goes through a capacitor first, according to the schematic.
Pin 6 (Threshold) has 3.8 volts and reads high.
Pin 7 (Discharge) has 4.5 volts and reads high.
Pin 8 (Power In) has 4.9 volts and reads high.

Here's the info for when the machine is in the working state.

Pin 1 (Ground) has zero voltage and read low on the logic probe. This is presumably because it goes straight to ground.
Pin 2 (Trigger) has 3.3 volts and reads as pulsing.
Pin 3 (Output) has 4.5 volts and reads high.
Pin 4 (Reset) has 4.9 volts and reads high.
Pin 5 (Control) has zero volts and doesn't read at all on the logic probe. This could be because it goes through a capacitor first, according to the schematic.
Pin 6 (Threshold) has 3.8 volts and reads as pulsing.
Pin 7 (Discharge) has 4.5 volts and reads high.
Pin 8 (Power In) has 4.9 volts and reads high.

From the broken state, grounding pin 40 on U15 will reset the machine and put it into the "working" state. Grounding pin 4 on U43 (the spot for the chip in question) will have the same effect. This is the first confirmation I have of spot U43 actually having some kind of positive effect on the game.

I think I understand why grounding pin 4 of the empty spot has the same effect as grounding ping 40 of U15 (as well as U10 and presumably others). If I follow the schematic, I see that the two are connected, so grounding one also grounds the other. But based on what I know the basic usage of the 555 timer chip is, I don't understand why the reset pin is the one that controls the grounding of other chips' reset pins.

The only thing I can think of is that it's not so much as the grounding of pin 4 that drives the reset of the other chips - it's that pin 4 is part of a larger circuit that gets grounded at some point along with all the others. What is the condition under which all these joined pins would get grounded at the same time?

And with that figured out, I move back to the point of the 555 timer within the scope of the whole system. If I understand how a 555 timer is most commonly used, it's designed to keep the output pin in a certain state until it detects a change in voltage of an attached capacitor. Depending on how its wired, it can either function as a "one-shot" pulse generator, or be used like a flasher circuit, alternating pin 3 between two states. There's some kind of third use that I still don't understand, as well.

How does a System 11 make use of the timer chip? Perhaps that will help me figure out why putting one into the board starts making things act weird.

(NOTE: There's also a second broken state where all three LEDs are solid. The voltages are the same as the working and primary broken state, but pins 2 and 6 are pulsing rather than high, just like the working state. I've only seen this state once, so it might have been a fluke.)

The first socket that was there was a round-holed machine-pin socket that I never put in there, so I can't speak to the quality of the work done to install it or how good a grip it had on any chips I put in there. After removing the first socket, I soldered the chip in directly. Perhaps some of the heat from the soldering iron toasted the chip. I pulled that out. I've put a dual-wipe socket back into the board, and it looks like things have improved. I can confirm that the chip is well-seated and each pin has connectivity to all the down-the-line components.

Where previously I was getting an all-around mess with the 555 chip in, I'm now getting close the same two states I was getting with no chip in there at all. This is better than bad, but it's still not good.

This new version of the "broken" state is the +5 LED on, the diagnostic and blanking LEDs are out, and the game never boots. I suspect it's because nothing ever sends the "reset" signal down the line. Grounding pin 4 of the timer chip (which also grounds the reset pin of several other chips) boots the machine just fine. Displays work well, sound works well, and everything works. Now, I need to find out if the initial reset signal ever gets sent. If not, why doesn't it?

I'm also still looking for details on how the system actually uses this 555 timer chip.

Spoke with a local pinball guy at PAX Prime in Seattle today, and in conjunction with my internet research, I think the problem is with the Power-On Reset circuit. Grounding one of the reset pins on one of the main chips (U10, for example) has never failed to properly reboot the game, which seems like a pretty clear indicator. I've read that I should test transistors Q34, Q36 and Q38 as well as resistors R55 and R56, Diodes ZR1 and ZR2 and capacitor C30.

The resistors test out fine, and the diodes are definitely only allowing small amounts amounts of voltage in one direction. However, since these are Zender diodes, I can't say for sure if they're in spec without some kind of variable power supply. The three transistors are also a bit weird in that Q34 and Q36 both read the same voltage for any given combination of red and black leads from the DMM to certain pins, but Q38 gives slightly different numbers. This could be because Q38 is in a different enough spot in the circuit to give these different numbers, or that Q38 is busted and needs replacing. The capacitor is also kind of untestable without a real power supply attached to it out of the circuit.

I think I'm going to tack on the three bucks worth of replacement components for the entire POR circuit to my next GPE or Marco Spec order and see what happens.

When you guys talk about "lifting a leg", do you mean de-soldering the entire chip, bending the specified leg up, and putting the chip back in?

You'd have to pull that leg up and then bend it at close to 90 degree angle to get it clear of the hole. Doesn't that kind of stress on a single leg run the risk of snapping it?