Time to call it a night!

The purpose of capturing the CLK signal was to be able to determine exactly when the scope wanted to read the incremental encoders.
If we can't do this effectively then maybe need to try another crude approach.

We know that when the "C" control signal at U2 pin 9 goes high, the scope is in the phase of reading the incremental encoders. So using U2 pin 9 going high as the trigger, what happens on the FPMUX signal at U2 pin 3?
You're going to need to connect both oscilloscope probes. Ch1 on U2 pin 9 and Ch2 on U2 pin 3.

The problem with this approach is we don't know which encoders data we're looking at.

Yes both at the same time, because when U2 pin 9 goes high that's when one of the encoders are being read (with this approach we don't know which encoder). So we're looking at what info is coming out of U2 at that time. The encoders should be reporting a mixture of open and closed switches going into U2.

What about a third probe? (I'm not kidding ). How gymnastic are your fingers?

Ext Trigger input connected to the FPCLR pin 2 of U1. This tells the scope and us when the read process is beginning and to start tracing.
Ch1 input connected to the FPCLK pin 1 of U1 which we can use to tell us which switch is being read at a certain point in time.
Ch2 input connected to the FPMUX pin 3 of U2 which tells us the state of the switch (or encoder we are interested in) being read.
Somebody rotates the first encoder (Volts/Div CH2) and we watch if the switch state out of U2 is changing.

Getting complicated huh?

Answering your previous question, voltage on channel 2 make it the same as channel 1, time base probably around 24us.

Let me book a super jet now and I might make it in time..

Sorry, I'm confused...

Bob is trying to see 63 or 64 clock cycles, correct? Then, ideally, maybe see a delay. Bob's plot below is in the nS range, so the plot he captured is a single clock cycle, no? Why doesn't he increase the sweep time in order to count the pulses?

Might be a moot question if you are try to capture a specific event.

Regardless, finding a proper trigger would be the way to go if he can somehow get a probe on it.
pasted_image (resized).png

Hey,
Have a look at the diagram I drew on the following post which shows all the switches configured in the matrix. I left off the U1 chip that controls both 74HC4051 chips.
https://pinside.com/pinball/forum/topic/mpu-100-don-t-boot-on-first-try/page/3#post-5284901

If you download the schematics Bob linked and mentally put U1 from page 7 onto my diagram, you should get an idea how the switches are read and that it takes 63 clock pulses to go through the matrix.

https://www.mikrocontroller.net/articles/Datei:Hantek_Tekway_DSO_MSO_hw1007.pdf

The assumption has been that the CLK signal we're trying to measure, pauses at the end of each 63 cycles which may or may not be true.
If yes we have some way of knowing when the process starts and when each switch is being read.
If not we need to trigger the scope on the "CLR" signal going to U1 to know when the circuit is being reset to read the first switch state.

Bob needs more hands to hold all the probes in place or find a way to connect them to the board by soldering wires to the pins being probed.

At the end of the day, there's 3 surface mount chips controlling these switches/encoders and U2 on the board is the first suspect with its "C" input pin potentially not working.
One way or another Bob probably needs to replace it and if unsuccessful maybe the other chips. If replacing the 3 chips doesn't fix it then it's potentially an issue at the FPGA 1 chip on the main board which could mean it's game over.

Kind of, you're getting external activity on U2 pin 9 coming in, but internally, U2 pin 9 might be open circuit - this is the assumption since U2 pin 9 needs to go high to pass the rotary encoder signals through.

The waveforms you posted for U2 pins 11, 10 and 9 all look good to me. You should notice that pin 11 is switching twice as fast as pin 10 who is also switching twice as fast as pin 9.

Essentially, the way U1 is wired yes. But your U3 pin 11, 10 and 9 pin waveforms don't look right. They're not uniform like those from U2. Now the U3 pins 11, 10 and 9 are controlled by the same U1 signal (pin 5) driving the suspect pin on U2 pin 9.

Download the U1 "74HC393" datasheet. Have a quick read of the description and then look at the function table which will tell you what happens on the outputs with each active CLK signal change. This chip is designed to change outputs on the falling voltage edge of the CLK signal. We call this a falling edge triggered clock. A logic high voltage on the CLR signal on this chip resets all outputs to logic low voltage state. i.e. when this happens both U3 and U2 are connected to the very first switch in the matrix.

https://www.ti.com/lit/gpn/sn74hc393

Hot air guns aren't that expensive but you can do this with a soldering iron. My method back in the early 90's was to solder the corner pins so the chip is held in place, then flood solder over all other pins. Using the thinest solder wick, remove any solder bridging the pins. It actually worked quite well with nice results. Flux helps a lot with this because it makes the solder less likely to bridge when you're cleaning up excess with the solder wick.

Of course you need to remove the chip first, probably cutting each pin with a sharp pointy nose cutter then remove the remaining pin debris with your soldering iron and solder wick.

Pins 1 and 2 of U2 connect to the rotary encoders which are reporting their bit positions. While the activity looks unusual, they should be changing.
Pin 5 goes to all the push button switches on the encoders. Since you're not pressing any of them, they are all responding with open circuit. Press a few of the encoders push button switches while analysing pin 5. Does the waveform change?

Bob, see the two diagrams below.
I've configured both the U3 and U2 A,B,C input pins - refer to the function table.

First picture shows U3 C,B,A pins configured as Lo, Hi, Lo respectively. This makes U3 internally switch the X2 pin to the X pin which is grounded. So in the switch matrix, you can see the column of switches on X2 are now connected to ground.

U2 C,B,A pins are configured as Hi, Hi, Lo respectively. This makes U2 internally switch the X1 pin to the X pin which is then sent to the mainboard via the FPMUX signal. So in the switch matrix you can see the row of switches which are now connected to U2 X1.

The junction point of these two selected column and row lines is the front panel switch "F0" (Function 0).
If the F0 switch is being pressed, then the FPMUX signal reports logic low zero volts to the mainboard that the switch is closed.
If the F0 switch is NOT being pressed, then the FPMUX signal reports logic high 3.3 volts (via the 10k pullup resistor) to the mainboard that the switch is open.

U1 controls the state of the A,B,C input pins to both the U3 and U2 chips. U1 is controlled by the mainboard which via the CLK signal steps through the switch matrix.

The second picture I've configured the U3 and U2 A,B,C input pins to look at one of the rotary encoder lines of the "V0" control.

Oscilloscope_FrontPanelSwitches_1.png
Oscilloscope_FrontPanelSwitches_2.jpg

Can you post a more closeup clear picture of U2? It almost looks like some of the pins on the left side aren't soldered properly.

When you're not sure of the package type, download the datasheet - you'll find details there:
Your dimensions roughly match the "TSSOP-16" package.

.
74HC4051.jpg

Bob,

Do you have a micrometer? If you cant read the full part number on the IC, then just measure.

I'm a big Mouser fan. Digi-Key is good also.

It's a TSSOP-14 package. The "PW" letters is Texas Instruments way of including TSSOP-14 to their part number which for them becomes 74HC393PW.

On a quick look I like the Nexperia brand parts which have low capacitance. You should be fine with the two Texas Instruments parts you linked from Mouser.

But gee they're pretty small. I hope you've got a decent magnifying glass.

I'm listing these 4051 grouped by type:

These two are 74HCT where the 'T' indicates they are TTL version and only run at 5 volts - you don't want these.
https://www.mouser.com/ProductDetail/771-74HCT4051PW
https://www.mouser.com/ProductDetail/771-74HCT4051PW-T

These two are 74LV where the 'LV' indicates they are low voltage version (between 1 and 6 volts) you don't want these.
https://www.mouser.com/ProductDetail/771-LV4051PW112
https://www.mouser.com/ProductDetail/771-74LV4051PW-T

These two are 74HC (same spec your board has) which is what you want. They are identical, the only difference is how they come packaged from the supplier. The first one comes in tubes and the second one '-T' comes in tape reels. Choose whichever. The tube ones will probably be easier to store away.
https://www.mouser.com/ProductDetail/771-74HC4051PW
https://www.mouser.com/ProductDetail/771-74HC4051PW-T
.
.
Ditto for the 393 parts, you want 74HC only - first one is tube, second one is tape reel.

https://www.mouser.com/ProductDetail/771-HC393PW112
https://www.mouser.com/ProductDetail/771-HC393PW118

Since their cost is peanuts get some Texas Instrument ones too.

Funny I was looking at buying those a month ago but got something else and not sure I made the right decision. If you get them I'll be interested on your thoughts. It's got more than enough zoom I think.

By the way, I find soldering these small things outdoors in daylight helps a lot with visibility - best light conditions..

Bob,

Removing/replacing the 4051 should be rather painless.

1) First off, note the pin 1 mark, dot, etc.

2) The simplest way to remove is to use two irons with a wider (chisel) tips, or maybe just a normal tip placed at an angle. But not high wattage! It won't need much heat to flow the solder - it's just a little part. You can actually add solder so that it bridges the pins on each side to help transfer the heat. Plus use some 63/37 or 60/40 (lead) solder since the scope is likely RHOS. Lead free solder melts at a high temp is somewhat of a PIA.

3) With an iron on each side and solder melted, the part will easily slide free. Shouldn't damage any traces. Don't force it, it should slide off like butter. Watch out for the bypass chip cap right next to the IC. Thank goodness there's no thermal pad to unsolder on the bottom of the part - that would make it a lot harder.

4) Use solder wick to clean up excess solder on the pads. It will probably be a mess due to the extra solder, but the solder wick will take care of it.

5) Place the new part on the pads, use just a small amount of solder on opposing corners to anchor and center up the part. Solder the remaining leads, and then correctly solder the first two joints used to anchor the part. Too much solder? - just wick some off. Check the pin 1 mark.

Easy peasy. I'm an old fart and have been using surface mounted parts for many decades. For the most part, I'd much rather work on surface mounted parts than thru-hole. At work we even use 0201 (0.02 x 0.01") parts and we are going smaller, but I admit I use bigger parts due to my line of work. You might want to get some thin solder.

You might be surprised how easy it is to replace/install if you have steady hands.

Got an old board to practice on?

Quickly tinning the part's lead prior to installing will help a lot. Don't need much, just a light coat. You can also add a little solder to the opposing pads so it's all set to go. Again, not much is needed - just enough to hold it in place and solder properly afterwards. Just take the eraser end of a pencil (or something similar) and hold the part in place, hit the opposing leads, part will settle in place because of you holding it in place - secured. No need to overthink it. I would think the tape would be a big PIA and is not needed.

0.031" solder - give it a try and make sure you don't short between pins. Smaller might be better, but 0.031 is definitely OK.

You'll get the hang if it - it really is not hard if you have halfway decent soldering skills (and you do). Too many post on Pinside about SMT being this weird and hard thing to work with since it's new to them. It's just not.

If you're really worried about it, just buy this and practice first: A few dollars, and money well spent to calm your nerves.

https://www.amazon.com/s
pasted_image (resized).png

ChipQuik seems to be a brand of a range of products. Which one are you referring to?

If it's not too late get the thinnest solder wick you can. My personal experience is I always found it performed better for dealing with surface mount stuff - it robs less heat from the iron and is easier to manage. The wick I've always used is about 1mm thick.

I put the chip into position, push down on the chip say with a screw driver so it doesn't move, put some solder on the iron and then iron onto a corner leg so it attaches to the board. Inspect the chip positioning to the PCB pads and once I'm 100% happy it's centered, do the opposite corner leg. If positioning is still good then solder the rest. Otherwise adjust if necessary til it's right before doing all pins.

Great idea!

Interesting.
You'll be an expert in no time

That is some interesting stuff. I don't think you need that since your part is so small. In the demo the chip was huge - so that does look useful in that instance.

On a side note, he did make a very good comment...don't drag the iron and solder wick across the pins when cleaning.. Just down and up, then move over, repeat. Also, that was a pretty good size iron for the little pads to use w/the solder wick. If I was removing excess solder, I'd use a smaller iron.

Also, not all solder wick is created equal. Some has very little flux, and it can be a little bit of a pain to get it to wick up solder. Add some rosin flux to it if it doesn't take solder well. Good solder wick will really pull solder away.

Yeah, this is the problem with some of them.

Yes! Flux is your friend with surface mounts.

I thought about the residual low temp solder, but once you wick it up, it will be gone. Any minuscule amount will simply mix with a lot more fresh 63/37 and be a moot point.

344C seems VERY hot. 63/37 should melt around 180 - 190C. I'm not opposed to applying a fair amount of heat to make the job go quickly. To me, 'not much heat and taking forever' to get the job done is by far an easier way to damage a board rather than 'hot and fast' with a little more heat - withing reason. 344C seems way too hot. WAG, but 250C'ish should be enough, depending on the tip. Hence, my comment about globing solder over all the pins to heat them up all at once with a bigger tip, and sliding the part off. Fast and done...

Well at least it looks like you got the same brand 4051 as the original.

After you get the hang of this, you will officially be known as "NewSchoolBob".

You can do it! I built a few pieces of test gear for myself and the chips only came in surface mount. I managed and it was not too hard. Take your time and flux is your friend.

to remove a smallish SMT chip with pins on two sides like shown in the picture using just a soldering iron wand style I would blob some solder across all the pins on one side. This will help transfer the heat to all the pins. Heat up one entire side until all pins are loose and the slide a wedge under the chip to lift the hot side up off of the board then let it cool off off the pcb. Then do the same on the other side and it will come off. Use solder wick or a vacuum desolder to clean up the pads.

Its a bit of a brute force method but works for me and I can do it without damaging the board but the chip being desoldered will usually get ruined. Most dangerous part is sliding the wedge in under the chip when the first side is hot. If you lift to hard you can rip the tracks off of the cold side.

When they are the standard outline, biggest size of SMT, flush cutters can usually snip the chip out then just clean off the left over legs stuck on the board with a de/soldering iron.