74LS76 J-K Flip Flop used as a 2-bit counter: Strange output of high-order output bit

Thread Starter

aroetter

Joined Sep 24, 2017
27
Hi there,

I'm building an 8-bit breadboard computer, based on Ben Eater's excellent 8-bit computer project but with some of my own extensions. Goal is just to have some fun and learn a bit about HW (I'm a SW guy by training/profession)

I'm having trouble building an output register, and while I've found a fix for my bug (involving a random 220 Ohm resistor, see below), I have no idea why it works (or even why it should), and I'm trying to fully understand the system and not just place the randomly discovered band-aid and paper over it.

Before I post a schematic below, let me explain at a high level the breadboard I'm debugging:

1. A 555 timer generates a clock frequency (probably not relevant so I'm omitting details).
2. A 74LS76 Dual J-K Flip Flop is used to build a 2 bit timer, iterating forever over (00, 01, 10, 11). Those 2 bits drive a decoder (see below). Those 2 bits also serve as input address lines to an EEPROM. Let's call these 2 bits the high-order bit and the low-order bit. The low-order bit output also serves as the clock for the high-order bit.
3. A 74LS139 decoder (see above) takes the 2 bits and outputs a repeating pattern of (1110, 1101, 1011, 0111). Those 4 bits are connected to the common cathode of 4 adjacent 7-segment LCD displays.
4. An EEPROM is addressed by the 2-bit counter, and sends "what segments of the displays to light up" to each of the 4 7-segment displays.

In this way, I can turn on 1 of 4 LCD displays, and drive "what segments to light up" from the EEPROM. When I run the clock fast enough, it looks like all 4 displays are driven at once and I have a nice output for an 8 bit register (sign bit and then 3 decimal digits).

I am following the design from: https://eater.net/bbcpu8-output-register/

Here is the problem: The Dual J-K flip flop should output 2 square waves, with the high-order bit at half the frequency of the low-order bit. It doesn't. The low-order bit looks great, but the high-order bit is just flat in the low position (~0.15V):

<<See Attachment: System_Broken_Without_Resistor.png>>

Here is the weird part (which I randomly discovered while inserting a bunch of debugging LEDs... before I just borrowed an oscilloscope so now I don't use LEDs for debugging!). Even though the problem is with the output of the high-order bit from the JK flip flop, if I connect the "low-order" bit output to ground via a 220 Ohm resistor, everything works fine!

It changes the high-order bit output and I get the desired result:

<<See Attachment: System_Working_With_Resistor.png>>

[In both these diagrams, Channel 0 is the low-order output of the J-K flip flop (pin 10), and Channel 1 is the high-order bit (pin 14)]. You can see that logical high for channel 0 is only 2.5 volts, which I believe is due to the resistor to ground. Logical high for the high order bit is 3.6V. If I remove the resistor, I get a good 3.6V on Channel 0 for the low-order bit, but of course then Channel 1 is just DC at 0.1V or so

So, given that something is weird with the output of the high order bit, I started to think maybe one of the "consumers" of that output was behaving badly. As you can see from the schematic below, this high order bit output goes both to the 139 chip, and also to an address input of the EEPROM (2 green lines in my hand drawn schematic). Here is the weird thing: when both of those output lines are connected, the signal is wrong (the low, noisy image above). However, if I disconnect *either* the line into the 139, OR, the line into the 9th address bit of the EEPROM, the output signal from the JK flip-flop becomes fine again, and I get the 2nd oscilloscope screenshot. Disconnecting *either* of them is sufficient to get the output signal from the JK flip-flop correct again (but of course the LCDs don't work properly with either line disconnected, so this, unlike the random 220 Ohm resistor, isn't really a functional fix).

So, this would lead me to believe something is weird with either of those chips. FWIW, I've replaced all 3 chips (JK, decoder, and EEPROM) with spares but that doesn't fix anything. Also, I'm totally confused by why a 220 Ohm resistor to ground, on the LOW ORDER output, fixes the HIGH ORDER output bit of the JK flip-flop chip.

One more note on power: The breadboard is connected to a 5V DC bench power supply. Across +%5V and GND I have a 0.1uF decoupling capacitor and a 330uF power storage cap (thanks to others on this forum who taught me about that in previous threads!)

In case you can't tell by the above, I'm very new to electronics. I've tried to include all the relevant details above, but also omit what I think are extraneous details for clarity. However, if you read this and think "how can I help when he didn't say X", that's certainly b/c I don't know enough to know what is relevant to post. Please just ask and I'll provide any info that is useful. I greatly appreciate any help!

Hand-drawn Schematic Attached as: Output_Register_Schematic.pdf
The outputs from the high-order bit of the JK Flip Flop are labelled green. Disconnecting either one of these two wires gets the waveform on this output back to normal. The outputs of the low-order bit are red/pink. This is where we can add a 220 Ohm resistor to ground to make everything "work"

I'm totally perplexed at this point. Thank you so much in advance!
 

Attachments

Thread Starter

aroetter

Joined Sep 24, 2017
27
Follow up with one more piece of info from my debugging. (sent as a separate reply to keep size down).

You may notice in the schematic above outputs 1Q and 2Q are both tied to nothing. Even though they are outputs, I know that is not best practice, so, I tied both to ground through 1K resistors, but, that didn't change anything. I still get the flat @~0V "high order" output signal.

Then, just to see what would happen, I removed those resistors (so they are floating free again), and tied 1Q and 2Q together directly. This gave me a super bizarre 3-state output (as in I get 3 of the possible 4 2-bit outputs). And the signal looks as attached.

Connecting 2 outputs (one of which may be high and one may be low) together directly seemed bizarre to me, so then, I decided to connect them via a 1K resistor instead. That gets me back to the original broken world where the high-order output is always low.

At this point, I am randomly connecting resistors to things which doesn't seem very principled, so, I'm going to stop now. I just wanted to pass along that other piece of info.
 

Attachments

ebeowulf17

Joined Aug 12, 2014
2,942
Follow up with one more piece of info from my debugging. (sent as a separate reply to keep size down).

You may notice in the schematic above outputs 1Q and 2Q are both tied to nothing. Even though they are outputs, I know that is not best practice, so, I tied both to ground through 1K resistors, but, that didn't change anything. I still get the flat @~0V "high order" output signal.

Then, just to see what would happen, I removed those resistors (so they are floating free again), and tied 1Q and 2Q together directly. This gave me a super bizarre 3-state output (as in I get 3 of the possible 4 2-bit outputs). And the signal looks as attached.

Connecting 2 outputs (one of which may be high and one may be low) together directly seemed bizarre to me, so then, I decided to connect them via a 1K resistor instead. That gets me back to the original broken world where the high-order output is always low.

At this point, I am randomly connecting resistors to things which doesn't seem very principled, so, I'm going to stop now. I just wanted to pass along that other piece of info.
I'm really intrigued by your problem. I have very limited experience with logic chips like you're using, so take my thoughts with a grain of salt!

I have a hunch, or gut feeling, that pin 10, when outputting "low," isn't sinking enough current, or simply isn't pulling down to a low enough voltage and/or pin 1 is unusually insensitive, requiring an especially low voltage to register as logic low. If this were true, adding more load on pin 10, which would pull the output just a little higher, could be enough to push you over the threshold and make pin 1 fail to recognize low states.

Unfortunately, the datasheets don't seem to support this idea. When all chips are behaving normally, it looks like pin 10 should be able to sink more current than all three inputs that it's linked to need. So, I'm probably not on the right track.

Nevertheless, one thing you might check as an extra clue to work with is the voltage reading at pin 1, during logic low states, in each of the three configurations (no external chips connected, one external chip, both external chips.) If you do see a rising pattern in the voltage at pin 1 as you increase the number of external chips sharing pin 10's output, that might guide you to a better solution.

Two possible solutions that occur to me would be using either a MOSFET or an op amp in between pin 10 and all of the inputs it's driving in order to get a lower voltage, capable of sinking more current, for logic low states.

Other thoughts that occurred to me would be:
  • Providing power supply bypass (decoupling?) caps at each chip - it sounds like now you have it only at the power input of your board.
  • Investigating the effects of parasitic capacitance and/or unexpected voltage drops as a result of using breadboard vs. a cleaner pcb/soldered assembly. Unfortunately, it's beyond my skill set to assess the severity or meaning of these issues.
 

MrChips

Joined Oct 2, 2009
19,280
Take some notes:

1) It is ok to leave output pins unconnected.

2) It is not ok to tie output pins together.

3) It is not ok to tie outputs to GND via a resistor without a particular reason.

4) It is not ok to take logic signals from an output pin that is already driving an LED (just saying because of a recent thread).

5) Replace all 555-timer ICs with CMOS versions, LMC555, TLC555, or ICM7555.
 

Thread Starter

aroetter

Joined Sep 24, 2017
27
Mr. Chips,

Thanks for the tips. Understand all (and good to confirm #1), but not quite sure where that leaves me.

Can you describe #5 to me? Do you think that might have any bearing on this issue? I would imagine not since I'm getting the JK flip-flop activated properly, meaning the incoming clock from a 555 timer is working okay?

#3: Totally agree. It's just the only thing I've found that patches up the symptom, but I agree it's not correct which is why I'm posting.

If you have any ideas on what I could do to debug this further, I'd love them.
 

Thread Starter

aroetter

Joined Sep 24, 2017
27
I'm really intrigued by your problem. I have very limited experience with logic chips like you're using, so take my thoughts with a grain of salt!

I have a hunch, or gut feeling, that pin 10, when outputting "low," isn't sinking enough current, or simply isn't pulling down to a low enough voltage and/or pin 1 is unusually insensitive, requiring an especially low voltage to register as logic low. If this were true, adding more load on pin 10, which would pull the output just a little higher, could be enough to push you over the threshold and make pin 1 fail to recognize low states.

Unfortunately, the datasheets don't seem to support this idea. When all chips are behaving normally, it looks like pin 10 should be able to sink more current than all three inputs that it's linked to need. So, I'm probably not on the right track.

Nevertheless, one thing you might check as an extra clue to work with is the voltage reading at pin 1, during logic low states, in each of the three configurations (no external chips connected, one external chip, both external chips.) If you do see a rising pattern in the voltage at pin 1 as you increase the number of external chips sharing pin 10's output, that might guide you to a better solution.

Two possible solutions that occur to me would be using either a MOSFET or an op amp in between pin 10 and all of the inputs it's driving in order to get a lower voltage, capable of sinking more current, for logic low states.

Other thoughts that occurred to me would be:
  • Providing power supply bypass (decoupling?) caps at each chip - it sounds like now you have it only at the power input of your board.
  • Investigating the effects of parasitic capacitance and/or unexpected voltage drops as a result of using breadboard vs. a cleaner pcb/soldered assembly. Unfortunately, it's beyond my skill set to assess the severity or meaning of these issues.
Thanks ebeowolf17 for the detailed and thoughtful post. Sadly, none of it panned out and I'm still stuck, but, it gave me a lot to think about.

Let me describe what I did in case it's helpful to you.

First, I hooked up another set of leads for my scope to pin 1 of the JKFlipFlop. Since this is connected via wire to the pin 10 output, I would expect every trace to show identical signal on pin 1 as pin 10, which it did (I guess good to rule out some crazy issue with my lead wire).

In my original post, I had removed jumpers from pin 14 (the "high order" output of the JK) to the downstream chips, and showed that removing either one gets the correct 2 waveforms back out of the "low order" and "high order" bits.

This morning, based on your post, I left all those in, pulled out the resister, but tried removing the 2 outputs from the "low-order" chip, per your advice.

I see the same thing. When the low-order output, pin 10, is connect to both the 139Decoder chip AND the EEPROM, the system is broken as described. However, removing either one of them makes the system behave again, the high order and low order outputs make nice square waveforms as I'd want.

I looked at the voltages, and, nothing stands out to me (I have all the scope screenshots if you'd like but don't think they are helpful).
The waves, when working, are always at 3.4-3.5V high, 0.16-0.18 volts low. There isn't really a pattern of them climbing as more things are connected.

I know that "what worked for someone else won't necessarily work for you" but, it does seem like this circuit did work for the original author, and other "copycats" such as myself. So, I'd love to learn a way to debug this without having to add other chips, replace chips with diff types, etc, etc.

But, at this point honestly I'm not sure how to debug further, and make progress on this issue when more things are connected the "high-order" bit goes flat. So, I'd be open to ideas.

Thanks again. Even though we haven't found a solution yet your post gave me some good ideas, caused me to think, and try a few new things, so, thanks!
 

AnalogKid

Joined Aug 1, 2013
8,108
For the unused half of the 74LS139 you have the unused outputs tied directly to GND. This will eventually destroy the device. For ALL standard TTL chips, unused outputs should be left floating. This is the lowest stress and lowest noise condition.

ak
 

Thread Starter

aroetter

Joined Sep 24, 2017
27
Thanks AnalogKid. To be honest, I only did that just before writing this post (b/c I didn't want to leave things floating). However, based on your advice and that of Mr. Chips I've removed all that now, the 4 unused outputs are floating.
 

Thread Starter

aroetter

Joined Sep 24, 2017
27
Alright, i did a bit more debugging, I am thinking it might be a problem in one of the columns of my breadboard.

Here is why I think that:
-- Circuit is fully connected, with "random resistor" removed. So the system is not working.
-- I now pop the EEPROM out of the breadboard, but leave all wires going into the lines that would have connected to the EEPROM. System is still not working.
-- There is a wire (pink in my schematic) that runs from pin 10 of the JK flip flop into the a8 input of the EEPROM. I remove the end of the wire going into a8 (where the EEPROM would connect to it if it were present). Simply by removing that wire into the breadboard, the problem is fixed!
-- I replaced that wires with another jumper cable. When it's connected to the JK and then to a8, even when the EEPOM is missing, the circuit is broken. If I pull out just the end that would go into the EEPROM, and leave the wire dangling off pin10, it works fine. And yes, there is nothing else connected to that column ont he breadboard, just the incoming pink wire and then the EEPROM.

This makes me think it must be some issue with the column of the breadboard that a8 goes into? It's going to be a *huge* pain to swap out the breadboard, as I have so many all connected to each other tightly, so, any way to verify this is much appreciated.

So my question:

Assume that is is a bad breadboard column. Why on earth do the above fixes (pulling certain lines, adding a resistor) mask the problem / have the behavior outlined above? This is an area where my knowledge is outstripped by the realities of what's actually happening, so, any ideas appreciated!
 

dl324

Joined Mar 30, 2015
8,952
Can you describe #5 to me? Do you think that might have any bearing on this issue? I would imagine not since I'm getting the JK flip-flop activated properly, meaning the incoming clock from a 555 timer is working okay?
Some seem think that there are issues with bipolar 555 timers. I've been using them for 4 decades and have never had a problem with the so called shoot through on the output. Use good supply filtering/decoupling and it should be fine.
 

dl324

Joined Mar 30, 2015
8,952
This makes me think it must be some issue with the column of the breadboard that a8 goes into? It's going to be a *huge* pain to swap out the breadboard, as I have so many all connected to each other tightly, so, any way to verify this is much appreciated.
If that row of the breadboard was ever abused (by inserting leads that were too large), it may no longer be able to make good contact.

You can test this by removing power from the circuit and using a continuity checker to verify correct connectivity. Wiggle the suspected connections while monitoring connectivity to make sure it isn't intermittent.

What gauge is the wire for that connection? You could try something a little larger.

How many tie points are available on that pin? Can you shift the EEPROM towards the other side of the breadboard to give more tie points on the bad side?


A number of people have complained that breadboards are useless because they can't provide reliable connections. I've never found that to be the case and I've used a half dozen different brands over the past 4 decades. IF there are any bad connections, you can find them using the same techniques that would be used to troubleshoot a circuit on a PCB.

I have had a few wires break, but they were straightforward to find using typical troubleshooting techniques.
 

ebeowulf17

Joined Aug 12, 2014
2,942
Alright, i did a bit more debugging, I am thinking it might be a problem in one of the columns of my breadboard.

Here is why I think that:
-- Circuit is fully connected, with "random resistor" removed. So the system is not working.
-- I now pop the EEPROM out of the breadboard, but leave all wires going into the lines that would have connected to the EEPROM. System is still not working.
-- There is a wire (pink in my schematic) that runs from pin 10 of the JK flip flop into the a8 input of the EEPROM. I remove the end of the wire going into a8 (where the EEPROM would connect to it if it were present). Simply by removing that wire into the breadboard, the problem is fixed!
-- I replaced that wires with another jumper cable. When it's connected to the JK and then to a8, even when the EEPOM is missing, the circuit is broken. If I pull out just the end that would go into the EEPROM, and leave the wire dangling off pin10, it works fine. And yes, there is nothing else connected to that column ont he breadboard, just the incoming pink wire and then the EEPROM.

This makes me think it must be some issue with the column of the breadboard that a8 goes into? It's going to be a *huge* pain to swap out the breadboard, as I have so many all connected to each other tightly, so, any way to verify this is much appreciated.

So my question:

Assume that is is a bad breadboard column. Why on earth do the above fixes (pulling certain lines, adding a resistor) mask the problem / have the behavior outlined above? This is an area where my knowledge is outstripped by the realities of what's actually happening, so, any ideas appreciated!
Here's a fun experiment: with the EEPROM disconnected, you say connecting the "pink" wire to one particular column causes the malfunction. In this particular case, I believe you have 5V in the column next to it. What happens if you disconnect the 5V supply from the column next to it? Could the capacitance of the breadboard columns be your issue - not necessarily a problem with any particular column, but just the relationship between one column (pink wire's target) and the adjacent column (5V supply.)

Also, this feels like a long shot, but I'll ask anyway: have you tried decoupling caps per chip (instead of per breadboard) yet? I don't understand the mechanisms involved, but I've read countless stories on these forums of truly inexplicable behavior that was solved with proper decoupling caps.
 

MrChips

Joined Oct 2, 2009
19,280
Are you still using bipolar 555-timer chips such as NE555 or LM555?

These are notorious for injecting switching noise into the power supply rails.
Sure, you can get your circuit to work properly with proper power supply filtering.

When you are having unexplained counter problems it takes very little effort to switch to LMC555 or TLC555 and see if the problem goes away. Of course, if the problem stays the same then you have to keep looking.
 

Thread Starter

aroetter

Joined Sep 24, 2017
27
Alright, so I think I'm starting to go a bit crazy :) Or rather, at least I've found some weird non-deterministic behavior. (The random "insert 220 ohm resistor to ground and move on" is starting to sound pretty good, sadly)

First, what works reliably:

I pulled the EEPROM out of the board. I've also pulled the 2 connections that used to go from pin 14 (the high order bit) and pin 10 (the low order bit) to that EEPROM. So they aren't connected to anything. Now I see the 2 square waves I expect out of pin 10 and pin 14.

Then I try to connect pin 10, and/or pin 14 to a breadboard column (choose a random unused column elsewhere in the system), things start to go south, but, it looks somewhat random.

Generally, I've found when pin 10 is connected to an "idle" breadboard column, and pin 14 is not connected to anything, it still works.
But if pin 14 is also connected, whether it still works varies a lot. If i find an "idle" breadboard column elsewhere in my computer (where the new board is in-use, but not the column), and connect pin 14, that tends to still be okay. When I connect to a brand new breadboard otherwise totally isolated from the system (whole board not in use, no power), it tends to kill it (the 'high order' bit goes flat). This is "fairly" repeatable. It's as if it likes breadboard columns elsewhere in my computer, but not brand new ones. Very strange.

Sometimes when pin 14 has a wire on it, but that's just floating in air, I see the right signal, sometimes not. Touching the end of the wire between my fingers affects things too. Generally, touching the 14 pin wire with my fingers gets "more" correct signal from the 'high-order' bit, whereas letting the pin 14 wire just float in air tends to make the 14 wire output totally flat. See the screenshot for what I mean by "more correct signal" works part of the time, then dies.

So it's as if just plugging in a loose wire to the pin 14 output causes it to fail. Unless I plug that wire into "some" breadboard columns, it comes back. Or I squeeze the wire with my fingers.

There must be a set of things to debug when you start seeing bizarre non-deterministic things like this, but, I'm sort of at a loss. I've got a basic multimeter and this software/USB scope I've been sharing screenshots of, but at this point I'm sort of struggling on how to make more progress. Maybe someone with more wisdom/experience has some ideas?
 

Attachments

Thread Starter

aroetter

Joined Sep 24, 2017
27
Here's a fun experiment: with the EEPROM disconnected, you say connecting the "pink" wire to one particular column causes the malfunction. In this particular case, I believe you have 5V in the column next to it. What happens if you disconnect the 5V supply from the column next to it? Could the capacitance of the breadboard columns be your issue - not necessarily a problem with any particular column, but just the relationship between one column (pink wire's target) and the adjacent column (5V supply.)

Also, this feels like a long shot, but I'll ask anyway: have you tried decoupling caps per chip (instead of per breadboard) yet? I don't understand the mechanisms involved, but I've read countless stories on these forums of truly inexplicable behavior that was solved with proper decoupling caps.
Sorry somehow I missed this reply.

I tried your experiment.
EEPROM chip removed. Pin 14 and Pin 10 both connected to what would be the A8 and A9 inputs of the EEPROM. high-order bit signal is dead. Per your suggestion, pulled the 5V connection, but that didn't help things.
When I remove the Pin 14 wire from the a9 board though, and leave it free floating, it behaves correctly.

I have a decoupling cap across the timer. But, I will now go add to every single chip on this board and see what happens.

I will also go try to order those other 555 timers and see if swapping them out helps at all.

I suppose, to be honest, I can just pop the 555 timer out entirely, and manually create a clock signal myself with wires by hand, and see if at least it works (albeit extremely slowly....)

Thanks for all the ideas, I'll try this later tonight (also see my recent post on bizarre non-deterministic behavior)...
 

ebeowulf17

Joined Aug 12, 2014
2,942
Sorry somehow I missed this reply.

I tried your experiment.
EEPROM chip removed. Pin 14 and Pin 10 both connected to what would be the A8 and A9 inputs of the EEPROM. high-order bit signal is dead. Per your suggestion, pulled the 5V connection, but that didn't help things.
When I remove the Pin 14 wire from the a9 board though, and leave it free floating, it behaves correctly.

I have a decoupling cap across the timer. But, I will now go add to every single chip on this board and see what happens.

I will also go try to order those other 555 timers and see if swapping them out helps at all.

I suppose, to be honest, I can just pop the 555 timer out entirely, and manually create a clock signal myself with wires by hand, and see if at least it works (albeit extremely slowly....)

Thanks for all the ideas, I'll try this later tonight (also see my recent post on bizarre non-deterministic behavior)...
The more you describe these crazy behaviors, the more it sounds to me like stray capacitance issues. Sadly, even if I'm right, I don't know how you defeat it (other than getting off of breadboards and actually removing the source of the capacitance.) A small bit of googling led me to this:

https://en.m.wikipedia.org/wiki/Parasitic_oscillation

The only mitigation advice that applies to your situation seems to be better decoupling. If that doesn't help, I'm out of ideas.
 

Thread Starter

aroetter

Joined Sep 24, 2017
27
Hi there,

Thanks for the thoughts. Even though I'm still stuck, having insight like "The more you describe these crazy behaviors, the more it sounds to me like stray capacitance issues" is super helpful.

So, here is what I did: I put a 0.1uF capacitor across every IC on this breadboard. I also put them across 3 other 555 timers I have on another breadboard (the computer clock, a few breadboards away).

Sadly that didn't work. When someone says "better decoupling" what does that mean? Should I literally put one of these across every single IC i have in my system? (49 at current count). I could try that of course, though am somewhat skeptical (not sure why).

I could also replace the 555 timers in my system (one on this board, 3 on the clock module) with the CMOS version peple mentioned.

However, I did an experiment: I unplugged the 555 in the schematic I've shared. And, I built a "poor man's clock" with a pushbutton, goes to +5v when depressed, gnd otherwise. I now used that as a clock, and I see the same behavior (albeit at lower frequency b/c I'm limited by my ability to press the button).

When pin 10 and pin 14 aren't connected to anything, i get the 2 square waves I want.
When the both are connected to dangling wires (other end in the air), the low-order square wave works (pin 10), but pin 14's output is flat.

I guess one other thing I could do is rip out the Dual JK flip-flop chip (74LS76) and build this 2-bit counter out of a 4-bit counter (74LS161). No idea why that would work, unless just some random difference in that chip means it's less susceptible to this problem.

But that doesn't seem that satisfying either. Feels like the most principled answer is get really good at decoupling, so I suppose if there is a question in this post, that's it. How do I do "best" decoupling? Is it as simple as 0.1uF capacitor right at the pwd/gnd inputs of every single IC?

Thanks again for all the help on this. Just having another set of eyes & opinions has been super helpful to me today as I've poked at this on and off.
 

Thread Starter

aroetter

Joined Sep 24, 2017
27
Final update for the day: I just ripped out the JK Flip Flop chip (74LS76) and rebuilt this with a 4-bit counter, just using the bottom 2 bits. It works great.

I've left in all the decoupling caps from the previous trials (which didn't fix that system) just to be safe.

Tomorrow I'll pull my mess of jumper cables (did this fast and dirty), replace with custom flush wires, etc, and see what happens!

Still curious about the decoupling question, but at least I have a working system now that doesn't rely on that weird resistor hack

(Though still a bummer I never could debug the 76 JK chip).

Thank you again!
 
Top