What is the voltage reading on the 2430 battery? The watchdog circuit on these boards will prevent booting if the battery voltage is low.

IMG_20200927_193502496.jpg

I was hoping it would be an easy fix to swap a battery, but at 3.2v it's obviously new. The 65C02 chip is a 5v chip and if there is a leg that is reading at 2.4v, it's highly suspect. If you follow pin 39 out on the schematic, it shows as needing a +5v.

Did you say it boots normally if you jumper it?

First warning: Do not disconnect the display with power turned on. This will kill U8G on the dot controller, a GAL that will cost your $20-25.

The CPU will (kind of) "boot" with a dead or missing dot controller, but it won't really do anything other than rapidly pulse it's LED, indicating that it is expecting to be talking to the dot controller, but for whatever reason, it can't.

Pin 2 and pin 38 on the 65C02 are tied high via the 3K pull-up resistor at R7. One of the chip select signals (CS1) for both 6522s is also pulled high with this resistor tied to 5V. Normal stuff.

Pin 39 is (phi 2), which is the clock signal that is phase shifted from phi 1 at pin 37. A clock signal will read somewhere between logic zero and logic one (0 to 5VDC....really between TTL logic 0 or about .8VDC and TTL logic 1 or about 2.7VDC) as it is pulsing constantly. A logic probe (or oscilloscope) is a better tool to test a clock signal. The clock signal at pin 37 depends on the external crystal circuit and the 74HC74. The clock signal at pin 39 is produced internally by the 65C02.

Shorting pin 39 to pin 40 didn't provide "extra voltage". When pin 39 dipped to logic 0, it pulled the reset line low with it. Then when you removed the probe, that allowed the reset signal to return high. The processor is reset upon a transition from logic 0 to logic 1.

What does a pull-up resistor do?
Google is your friend here, but in short, a pull-up resistor holds a signal high until a device in the circuit provides a path to ground, causing the signal to go low. Most everything in a digital computer works by pulling signals to ground. There is no such thing as pulling a signal that is grounded to high. If the circuit contains a path to ground, then the signal will always be low (logic zero). Only by removing this path to ground, will the pull-up resistor tied to 5V return the signal to logic high.

The message you are seeing is generally caused by a low battery or a problem with the 8 pin DS1210 immediately below it. Sometimes, that lithium battery leaks and corrodes the DS1210 causing it to fail. My preferred solution is to remove the battery and the DS1210 and install a PiniTech NVRAM module. Should you decide this is the way to go, send the board to a pro. System 3 boards are incredibly fragile. You won't regret paying a professional to repair your board correctly.

I think that does it. I hope this helps.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact ... for board repairs
http://www.PinWiki.com - The Place to go for Pinball Repair Info

Yes, it needs to be at 5v to work properly.

I think Chris nailed it with this:

But it is hard to tell if the chip is bad, or recognizing a short somewhere else and triggering the watchdog.

NO -- with a meter, you will not be reading +5V or 0V on pin 39.
It is a free clock running at 2 million cycles per second. About half of his time is spent low (near zero), about half of his time is spent high (near 5V). Your meter is taking the average of this and giving you 2.4V. Pin 39 does not seem to be the problem on this board.

The important part from Chris regarding pin 39:
"Pin 39 is (phi 2 AKA "Phase 2"), which is the clock signal that is phase shifted from phi 1 at pin 37. A clock signal will read somewhere between logic zero and logic one (0 to 5VDC....really between TTL logic 0 or about .8VDC and TTL logic 1 or about 2.7VDC) as it is pulsing constantly. A logic probe (or oscilloscope) is a better tool to test a clock signal. "

If you suspect the watchdog circuit - let us know what you get on pins 1, 2 and 3 of U16 and pins 4, 5 and 6 of U14.

Tail end of reset logic (U13B) is working. Seen as high at pin 4 of U14.
Watchdog timer combined with watchdog timer enable is working. Not trying to reset cpu, seen as high at pin 5 of U14.
CPU is not being held reset, seen as high at pin 6 of U14.

Watchdog timer is enabled (enabled when high, disabled when low). Seen as high at pin 1 of U16.
Watchdog timers are not trying to reset cpu, seen as low at pin 2 of U16.
Watchdog timer combined with watchdog enable is not trying to reset cpu as seen as high at pin 3 of U16 (same as pin 5 of U14).

So the above voltages are good and this says the cpu is being allowed to run.
It doesn't say the watchdog timer is fully working but it does appear to be.
CPU pins 39 and 40 of the CPU are fine. I recommend not shorting the two together in the future.

What these readings do not tell us - are manual reset and powerup-reset are fully working?
Starting with manual reset -- does pushing reset button make any difference?
Push reset button - while holding reset button, measure pin 4 of U13 - does it go low? If it goes low then manual reset (U13A/U13B) is fine
U21C is powerup-reset, We will test U21C next time.

Did you look at what Chris pointed out above in post #5?
Start looking around U6. Any trace of corrosion? Measure voltages around U6 when machine is turned on - pins 1, 2 and 8 for starters.
then measure pins 1 and 2 when machine is turned off.

Reset circuit is working.

U6 is fine as far as power control to the RAM. Just inspect him carefully for white corrosion particularly around seam.

From here - it would be plug and probe testing for me.
Could check display board's reset. Pin 6 of U7 to see if it is being allowed to run.

Would also be interesting to see which part of watch dog reset is triggering as mentioned in first post.
Is there activity on both pins 9 and 11 of U13 on CPU board?

Sounds like the CPU's reset is cycling due to the watch dog timers expiring. So something isn't running right. Could be due to a lot of things. Now it's time for some detailed looking - time to call for Chris' repair service. See post #5.

I'd try replacing C20 & C21 first. Those two caps can sometimes cause issues with the watchdog circuit.

I suppose if those two caps have diminished in value then the timeout period would be reduced. Yeah, I can see this happening.

The typical guidance is to replace those caps with 10uf caps. This changes the interval of the watchdog timer.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact ... for board repairs
http://www.PinWiki.com - The Place to go for Pinball Repair Info

It can't hurt. I just looked at one of my boards. I have a 10uf/50V (voltage doesn't matter as long as it's about 10V) across C20. It works. I can't recall the exact circumstances that caused me to add that cap.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact ... for board repairs
http://www.PinWiki.com - The Place to go for Pinball Repair Info

There is another "hack" you can try, but I don't know exactly how it works or if it does at all.
See this FB post...
https://www.facebook.com/chris.hibler.3/posts/3525932187468831:27
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact ... for board repairs
http://www.PinWiki.com - The Place to go for Pinball Repair Info

Can't get to that last link. Can you copy it here?

@g-p-e, I had posted a "what the heck is this" image on a FB group. I though a couple of the System 3 folks might have an idea. John Robertson chimed in and just last night he found an implementation note for the 74HC123 that isn't quite consistent with the "hack" that we've found on several Gottlieb system 3 MPU boards. The implementation note calls for a damping diode to be connected across the external resistor as shown in the following image.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact ... for board repairs
http://www.PinWiki.com - The Place to go for Pinball Repair Info

345DF376-029B-400E-843E-9AA1E8C63B83 (resized).png653D075C-732C-494D-9A2C-3C573A1C3666 (resized).jpegA50A7981-46F2-4D7F-8540-D09908E7817B (resized).jpegEDE0027B-B1A9-40D2-B168-44AF5605FC61 (resized).jpeg

OK - the use of clamping diodes make sense. 1uF in a timing circuit is kind of big... but I have seen them as high as 10uF.
Texas Instruments makes no mention of these clamping diodes. TI might have them built in, they often do.

But the pinout and photo don't make any sense -

U11 pin 2 -- The B trigger input on timer 1 = anode
U14 pin 4 -- Final reset combination = cathode?

U11 pin 10 -- the B trigger input on timer 2 = anode
U14 pin 4 -- Final reset combination = cathode?

Huh?

And if you look at the photo - it seems there is a cut trace near pin 4 of U14???

I would think anode to U11 pin 15 and cathode to U14 pin 11
And anode to U11 pin 7 and cathode to U14 pin 11.
Maybe somebody got the IC pins mirror imaged when working on the bottom of the board?

I had that thought too about a "mirror" issue.
But I've seen this exact mod made to at least two boards.
And I agree that it doesn't match the implementation note.
Still a mystery.

What looks like a cut trace is just solder flux that wasn't cleaned up when this mod was installed. It's not the worst flux leftover I've seen...I see a LOT of plumbers solder being used.

If I get another booger board that is having a boot issue with the message above, I'll give it a try.

Thanks for doing that analysis @g-p-e. I always appreciate you lending your expertise.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact ... for board repairs
http://www.PinWiki.com - The Place to go for Pinball Repair Info

I made a (lousy) sketch of the 1N4148 added diodes - it is clear what the intent is from that! On power-up the HC123s are clamped until /Reset is over.

John :-#)#

System3_MPU_Diodes (resized).jpg

Thanks John!
—
Chris Hibler - CARGPB #31
Http://chrishiblerpinball.com/contact
http://www.PinWiki.com/ - The new place for pinball repair info

Would that solve this issue?