Do you remember whether the other switches connected to U2 still worked? i.e. is anything hooked up to U2 working?

There's a function table in this datasheet for the 74HC4051 which can give you an idea what it does and if there's anything common to the signals not working.
https://www.onsemi.com/pub/Collateral/MC74HC4051-D.PDF

Bob,

The 4051 is pretty straightforward. Depending on the the level on A,B,C lines, it will select one of the X0-X7 inputs/outputs and connect it to the X (common) pin. Let's say you want the signal to go from X3 to the X (common), you would set Enable to low, set Select C, B, A to low, high, high (respectively). I highlighted this below in red and yellow. Want the X0 to X (common) connection? Set Enable to low, set C, B, A to low, low low - and so forth. Don't want any connections? Set the Enable to High to disable all the switches. You can have only one (or none) connection (i.e. you can't have X1, X2 on at the same time).

Think of it this way: It's just 8 switches (X0-X7) that go to a common connection (X). To pick what switch you want, set the the Select (i.e. the address) C, B, A appropriately along with having the Enable low.
pasted_image (resized).png

Thanks for that Mark, you saved me some typing
Bob, now with that info and the fact that X4, X5 and X6 at U2 are having problems (X7 isn't connected), do you see anything that's common to them based on the function table in Marks picture?

To answer your question Quench, it would be the C line is stuck low.

Maybe that was a test for Bob? LOL Sorry if it was...

pasted_image (resized).png

Do you remember if any of those push button switches on the pots still worked? because they have U2 X5 connected to them.
Actually they are not pots (potentiometers) at all. You'll notice you can freewheel them 360 degrees so they're not variable resistors with a start and end position. Those things are what's called "incremental encoders". Put simply, their pins digitally indicate which direction you're turning the rotary control.

In that case, U3 is probably ok.

Correct, U1 controls both U2 and U3. Your next mission is to download the datasheet for U1 "74HC393" and see what it does

Pin 9 being held low could be U1 pin 5, BUT U1 pin 5 also goes to U1 pin 13 which then controls U3. Since we presume U3 is working ok, the problem might potentially be the input pin 9 of U2.

In this circuit, pin 6 of U2 and U3 never go high. The circuit is designed so those two chips are always enabled by tying their active low "enable" pins to ground.

Nope, the terms LO and HI refer to logic levels. The red clip is just the logic probes power lead so connect it to the power rail of the logic circuit.
Have a read of the below link which explains logic levels (OFF or ON, 0 or 1, Low or High) and how voltages levels translate to logic levels for TTL and CMOS devices. Because logic levels are a little different between TTL vs CMOS chips in the middle voltage range, this is the reason for the switch on the logic probe so it can better determine what's "LO" or "HI" (low or high).

https://learn.sparkfun.com/tutorials/logic-levels/all

In the old days it was easy. Logic chips with part number starting with "74" (74 series) were TTL chips. Logic chips starting the "40" or "45" (4000 or 4500 series) were CMOS chips.
The funny thing is the part number for the U2 chip is a 74HC4051. This part number is an amalgamation of both 74 and 4000 series chips. The function of this part was originally sold as a "4051" back in the day.

https://en.wikipedia.org/wiki/List_of_7400-series_integrated_circuits
https://en.wikipedia.org/wiki/List_of_4000-series_integrated_circuits

That trace appears to be ground. Its likely for a RF connection to a shield or cover. EMI or similar type of connection. Look at where that board goes and check for fingestock, conductive foam, etc. Either it lowers an emission or grounds a RF hot spot on the board.

Sorry for the information overload..

Looking at the front panel schematic basically indicates the oscilloscope panel switches and incremental encoders are all hooked up to U2 and U3 in a switch matrix scenario. I'm not sure how familiar you are with the switch matrix on the early Bally/Stern games?

U3 sends the "Strobe" signals to the panel switches, U2 receives the state of the switches being strobed. The difference between the switch matrix on this panel and the switch matrix in those early pinballs is the oscilloscope panel reports one switch at a time (serial data) meanwhile the pinballs report 8 switches at a time (parallel data).
Besides power to U1, U2 and U3, only 3 other signals are used to interface all those switches/encoders to the main processor board in the oscilloscope.
The three signals are:
FPCLR to pin 2 and 12 of U1 which indicates to start reading from the beginning of the switch matrix.
FPCLK to pin 1 of U1 which moves to the next switch/encoder to read.
FPMUX from pin 3 of U2 goes back to the processor board with the state of the panel switch/encoder just read.

I don't think the logic probe will work on 3.3 volt logic circuits anyway. If you have one of the common 610 logic probes, minimum voltage is 4 volts.
You might need to use your other oscilloscope.

See if you analyse pin 3 of U2
Also pick one of the three rotary encoders and analyse the three pins near each other. Do the outer pins voltages change when you rotate its knob? I suspect you should see activity on the middle pin pulsing low which should be the strobe pin.

If it comes to attempting repairs, are you in any position to be able to replace U2? before shipping costs, 74HC4051 are less than a dollar.
Having said that, I think it might be too early to be attempting repairs in terms of replacing parts - what if they ask for the faulty unit back?

I'm impressed! that's pretty good.

I've drawn the switches in the matrix setup below:

The CLR signal from the mainboard forces the A, B, C signals to U2 and U3 to go low (zero volts) which switches X0 to the X pin on both chips.
At U3, this causes X0 pin to connect to ground via the "X" pin.
If the first switch is closed (top left corner), the X0 pin at U2 will be connected to ground so the mainboard will read logic low (zero volts) on the U2 X pin (FPMUX signal). If the switch is open the mainboard will read a logic high (the FPMUX signal has pull-up resistor (10k ohms) to 3.3 volts on the main board).

The mainboard then activates the CLK signal to switch U2 to the X1 pin (via the A, B, C pins) so it can read the state of the next switch. After a few CLK ratchets when U2 is at X7, the next CLK signal activation rolls U2 back to selecting the X0 pin. At the rollover, U3 then moves to the next pin which is X1.
So in the matrix, the top left switch is read at the beginning of the process and each CLK then goes down the column. At the end of the whole process which is 63 CLK signals later, the last switch on the bottom right has been read.

This process probably happens hundreds of times a second. FYI relating this to the Bally/Stern MPU boards they perform this process 120 times per second. Difference is they read eight switches at a time.

Do you have a soldering hot air gun? You'd be surprised how easy it is.

Oscilloscope_FrontPanelSwitches.png

Cue Pink Floyd?

U2 pin 4 (X7) isn't connected to anything that I can see on the schematic. So that waveform is probably residue you're seeing coming from the U2 pin 3 "X" pin back from the main board.

Can you take a snapshot of the CLK signal at U1 pin 1 showing the full cycle of 63 odd ratchet clicks?
Essentially what you need to do is use the CLK signal as your trigger reference, then with the other oscilloscope channel record the U2 pin 3 (FPMUX) signal.
At the 5th and 7th CLK ticks, the FPMUX signal will indicate the state of the two "Volts/Div CH2" encoder bits. Rotating this encoder should change the state of those bits during the read switch process.

However I understand this is going to be difficult to achieve with only 2 hands and limited access..

Even tho Quench has been steering this ship...I feel a little left out here. LOL!

But you're in good hands...

Please jump in if you have some extra insight or need to correct me

Nah, other than my simpleton comments about the 4051 and the EMI shielding, I haven't been following the thread very closely. You're doing a great job!

What we're looking for is the CLK signal being idle, then it pulses 63 times then goes idle again. That's the full cycle. This is the theory.
So you may have to adjust the Sec/Div to fit all those CLK pulses in on the screen.

I count 50 CLK ticks. Try taking the recording another few times. If there is no idle, then as soon as it finishes reading the last switch it goes straight back to read the first switch.
Can the scope go down to 4.00us SEC/Div? That should fit it all in - will be difficult to see the transitions though.

If there's no idle time then the only way to know when the read process is at the beginning is when the CLR signal pulses high (U1 pin 2).

Adjust the sweep time longer to catch all 63 cycles. Use the Single Seq button to catch the start of the 63 pulses. It will only trigger once (it won't keep sweeping). Make sure you set the trigger level to catch a real pulse (like you did above - that's probably OK other than you might have the x1 x10 setting wrong).

Sorry if I missed something, just butting in here for a few while I got a free sec.

FYI: The Single Seq button makes the scope only trigger once, it won't keep sweeping. So one it triggers at xx volts, it does a full sweep and stops. And it won't sweep until the input goes above the trigger level. If the time base isn't long enough (i.e to catch all 63 cycles), then you won't be able to count the pulses Quench is referring to. If you slow it down, then you should be be able to see all the pulses.

Make sense?

Thanks Mark, I actually don't have a digital scope so that info helps. I presume Bob will be able to zoom in time wise on the recording?

I have the same Hantek DSO5072P Bob has. I wasn't worried about Bob zooming in - thought that might muddy the waters. But yes, he could do that.

BTW, change the time base to 4.0us, not 400us.

At 4.0us Sec/Div, how many "screen fulls" of data could it record?

Bob, try hitting the F7 button. I think that allows opening up a window to zoom in. i.e. does a longer sweep and you can still zoom in to examine closely.

Not sure how much memory it has.

Bob,

F7 is hiding here:

You can see the wider sweep up top, but still zoom in below via the red brackets.
pasted_image (resized).png

Is the recording long enough to determine how many CLK ticks are occurring before it goes idle again?

Looks like you are getting there. Now increase the sweep time to catch the number of pulses you are interested in.

Is the scope probe ground connected? You have some noise on that trace.

Looking at your screen Bob, the complete cycle (on-off-on) is about 1.6uS. x64 = 102.4uS Set you scope to around that for a full sweep and you should see all the pulses Quench is talking about.

Set the sweep time longer to see more pulses if you are trying to count 63 pulses.