Joystick multiplexer help

[+johnnywc]

I had a couple hours free and wired up the design above with the 74HC03 and 74HC05 chips with the inductor and capacitors.  Functionally it works great!

 

The interference is much much less than the previous test with the x245 and 74HC04 inverters but still small traces at the top of the screen.  Interestingly, the interference is worse with the capacitors; if I pull them out it's barely visible but with them a bright short line of interference shows at the top.  The inductor doesn't seem to have much effect either as the interference seems the same whether or not the 5V vcc is connected to the chips directly or thru the inductor.  As with the previous design, if I disable switching (DUMP_PORTS is always on or off), there is zero interference (but of course just 1 joystick works and controls both players  ).  I hope I'm connecting the capacitors up correctly!   

 

FYI The 74HC05 has 6 inputs/outputs so I was able to use 5 of them for the output prior to the Atari for the 4 directions/button and the other to invert the select line (the design above has the up direction hooked up to a NAND gate to save a chip but I was able to use the 05 instead since it really doesn't save a chip - you still have 3 74HC03's and 1 74HC05).

 

For the select logic, I have the select pin 9 connected to a 10K pull up resistor before connecting it to the 74HC03's for joystick 1 inputs, and the output from the inverter (the inverted select line) I have hooked up to a 100K pull up resistor prior to connecting it to the 74HC03's for joystick 2.  Does that sound correct?

 

So it looks like we're getting really close!    Any other suggestions?  I can try a version with the LS641 open collector bus transceivers, plus I have the 240 octal buffers too.

 

 

[ChildOfCv]

1 hour ago, johnnywc said:

I had a couple hours free and wired up the design above with the 74HC03 and 74HC05 chips with the inductor and capacitors.  Functionally it works great!

 

The interference is much much less than the previous test with the x245 and 74HC04 inverters but still small traces at the top of the screen.  Interestingly, the interference is worse with the capacitors; if I pull them out it's barely visible but with them a bright short line of interference shows at the top.  The inductor doesn't seem to have much effect either as the interference seems the same whether or not the 5V vcc is connected to the chips directly or thru the inductor.  As with the previous design, if I disable switching (DUMP_PORTS is always on or off), there is zero interference (but of course just 1 joystick works and controls both players  ).  I hope I'm connecting the capacitors up correctly!   

 

FYI The 74HC05 has 6 inputs/outputs so I was able to use 5 of them for the output prior to the Atari for the 4 directions/button and the other to invert the select line (the design above has the up direction hooked up to a NAND gate to save a chip but I was able to use the 05 instead since it really doesn't save a chip - you still have 3 74HC03's and 1 74HC05).

 

For the select logic, I have the select pin 9 connected to a 10K pull up resistor before connecting it to the 74HC03's for joystick 1 inputs, and the output from the inverter (the inverted select line) I have hooked up to a 100K pull up resistor prior to connecting it to the 74HC03's for joystick 2.  Does that sound correct?

 

So it looks like we're getting really close!    Any other suggestions?  I can try a version with the LS641 open collector bus transceivers, plus I have the 240 octal buffers too.

 

 

Try the capacitors one at a time to see which one hurts more.  On the large capacitor, are you connecting the stripe side to ground, or to power?  It should be ground.  But the fact that it's worse with the capacitor would seem to indicate that it's drawing power.

 

Does the interference still only happen when a direction is pushed, or is it constant?  Does it happen even without joysticks plugged in?

 

Finally, to make sure we're correct on the power situation, it should be Atari pin 7 -> inductor -> ( all capacitors and VCCs and resistor pull-ups ), and ( all capacitors and GNDs ) -> pin 8

 

I may be able to test some things on my 7800, but it looks like even by the 2600Jr days they already noticed the folly of their power distribution scheme and modified it.  So the 7800 may be immune to these issues.  I do have a SECAM 2600 board that uses this stupid power scheme, but it seems to have a bad RIOT chip, at the least.  I'll have to socket the chips in my 7800, then maybe I can swap the RIOT from it and see if it works then.

Edited September 10, 2019 by ChildOfCv

[+johnnywc]

1 hour ago, ChildOfCv said:

Try the capacitors one at a time to see which one hurts more.  On the large capacitor, are you connecting the stripe side to ground, or to power?  It should be ground.  But the fact that it's worse with the capacitor would seem to indicate that it's drawing power.

The large capacitor causes more interference when connected.  The small one doesn't seem to have any effect.  I connected the large capacitor with the short pin (negative) to ground; the small capacitor doesn't seem to have polarity.

1 hour ago, ChildOfCv said:

 

Does the interference still only happen when a direction is pushed, or is it constant? 

The interference gets slightly worse when you press the joystick in one direction, but seems to almost go away when you press the joystick down and to the left or up and to the right.

1 hour ago, ChildOfCv said:

 

Does it happen even without joysticks plugged in?

The interference seems to go away when you unplug both joysticks.

1 hour ago, ChildOfCv said:

 

Finally, to make sure we're correct on the power situation, it should be Atari pin 7 -> inductor -> ( all capacitors and VCCs and resistor pull-ups ), and ( all capacitors and GNDs ) -> pin 8

Yes, I think that's how I have it hooked up.  I originally had the resistor pull-ups connected to straight 5v as it wasn't clear in the original diagram, but I made the change and it didn't seem to have any effect either way.

1 hour ago, ChildOfCv said:

 

I may be able to test some things on my 7800, but it looks like even by the 2600Jr days they already noticed the folly of their power distribution scheme and modified it.  So the 7800 may be immune to these issues.  I do have a SECAM 2600 board that uses this stupid power scheme, but it seems to have a bad RIOT chip, at the least.  I'll have to socket the chips in my 7800, then maybe I can swap the RIOT from it and see if it works then.

I can always ship you a unmodded 2600 if that helps and if you have the time to work on something like this.   

 

I want to also mention that the interference does go away completely if I disable switching or use the QuadTari with a standard Atari game (I use Basketball to test).  The controls work great (it always uses joystick 1 since Basketball isn't setting DUMP_PORTS) but there is no interference.

 

One last thing I want to mention; if I disconnect VCC (pin 7 from the Atari) completely, the circuit still works and there is no interference.  I measured vcc at the chips and it's getting 4 volts somewhere, probably from the other Atari pins; I measured the select pin and the directional/button pins and they measure 4.6 volts).  I'm sure it's not good for the overall circuit or health of the Atari/chips but thought it might be useful.

 

I should stress that the interference is almost down to nothing compared to what it was before.  Here is a pic and a short video so you get a visual.  The interference are the white dashes:

 

[ChildOfCv]

1 hour ago, johnnywc said:

The large capacitor causes more interference when connected.  The small one doesn't seem to have any effect.

Well at least the world isn't too insane then.  If the smaller capacitor was the cause, I'd be giving up on life.  But still, this is a DC filtering circuit, so it shouldn't be increasing the interference.  At best, it should be the same.

 

1 hour ago, johnnywc said:

One last thing I want to mention; if I disconnect VCC (pin 7 from the Atari) completely, the circuit still works and there is no interference.  I measured vcc at the chips and it's getting 4 volts somewhere, probably from the other Atari pins; I measured the select pin and the directional/button pins and they measure 4.6 volts).  I'm sure it's not good for the overall circuit or health of the Atari/chips but thought it might be useful.

This at least seems to confirm that it's the power draw that causes the problems.  Oh, and are you using 1M resistors for the pull-ups now?

 

Well, you can try building a minimal circuit and increasing it until the interference kicks in.

 

For starters, just do the filtered DC circuit (including the large capacitor) between pin 7 and 8 to make sure there's no interference when you try to input-select.

 

Next, add in the 75HC05 and its pull-up resistors on the inputs.  Also wire the input selector since it's part of this circuit, but you can leave its output disconnected for this test.  This will mean no real inputs, but at least some electronics.

 

Now try wiring it up so that one joystick can use the set of inputs of the 74HC05.  This still leaves the input select out of it, but now you have joystick input.

 

 

Is there a way that you can set the input select to toggle at a more noticeable rate?  If so, you could even use your multimeter to see whether it's the up transition or the down transition that causes the static (or both).

 

1 hour ago, johnnywc said:

Here is a pic and a short video so you get a visual.

Hmmm, are you sure that's not a UFO sighting?  ?

[+johnnywc]

16 hours ago, ChildOfCv said:

Well at least the world isn't too insane then.  If the smaller capacitor was the cause, I'd be giving up on life.  But still, this is a DC filtering circuit, so it shouldn't be increasing the interference.  At best, it should be the same.

I hope I'm hooking the DC circuit up correctly.   I have pin 7 going to a pin on the board, then the inductor between this pin and another on the board.  I then connect the output of the inductor to my common vcc pin strip on the side.  I connect pin 8 to the common ground strip on the board.  I then add the capacitors with one pin on the common vcc and one pin on the common ground strip.  I then pull off of vcc after the capacitors to feed the chips and resistor networks.  Hope that makes sense; I can take a pic to clarify.  Is it possible I fried the large capacitor when I connected it the wrong way (a couple times   )?  For sure the interference increases slightly but noticeably when I add the large capacitor and reduces when I remove it.

16 hours ago, ChildOfCv said:

 

This at least seems to confirm that it's the power draw that causes the problems.  Oh, and are you using 1M resistors for the pull-ups now?

Yes, I have 1M resistor networks on the 5 inputs from joystick 1, 5 inputs from joystick 2, and the combined output from the NAND gates prior to the 74HC05 which then goes directly into the Atari.  I am using a 10K resistor pull up on the select line input and a 100K pull up on it's output from the 74HC05.  Should I try a 1M on this too?

 

16 hours ago, ChildOfCv said:

 

Well, you can try building a minimal circuit and increasing it until the interference kicks in.

 

For starters, just do the filtered DC circuit (including the large capacitor) between pin 7 and 8 to make sure there's no interference when you try to input-select.

 

Next, add in the 75HC05 and its pull-up resistors on the inputs.  Also wire the input selector since it's part of this circuit, but you can leave its output disconnected for this test.  This will mean no real inputs, but at least some electronics.

 

Now try wiring it up so that one joystick can use the set of inputs of the 74HC05.  This still leaves the input select out of it, but now you have joystick input.

Okay, I will give this a try.    

16 hours ago, ChildOfCv said:

 

Is there a way that you can set the input select to toggle at a more noticeable rate?  If so, you could even use your multimeter to see whether it's the up transition or the down transition that causes the static (or both).

I changed my test program so it toggle the input select every 2 seconds (instead of twice a second).  In this case there is no static, you just can only use 1 joystick every 2 seconds.   I measured pin 9 with my multimeter and alternates between 4.98 volts and 0.16 volts every 2 seconds as expected.  I think it was suggested above that there may be a situation where all the chips are active for a split second during the transition that may be drawing more power which is causing the interference?  Maybe it's more noticeable when it's happening 120 times a second vs. once every 2 seconds.

 

I'll keep you posted on my progress.

 

Thanks!

John

 

16 hours ago, ChildOfCv said:

 

Hmmm, are you sure that's not a UFO sighting?  ?

 

[ChildOfCv]

3 hours ago, johnnywc said:

Is it possible I fried the large capacitor when I connected it the wrong way (a couple times   )?

Well, it's possible.  But you should notice a definite change in voltage when plugging the capacitor in in that case.  You can hook up the capacitor through an amp meter, that is, "- stripe" to ground, + side to meter black wire, meter red wire to VCC, and meter in amps mode (if the meter requires a different plug for current measurements, make sure you use it), and see how much current is flowing through the capacitor/meter.

 

3 hours ago, johnnywc said:

Should I try a 1M on this too?

It probably won't have enough impact on the circuit to matter.  So that would be one of those desperation moves

 

3 hours ago, johnnywc said:

I changed my test program so it toggle the input select every 2 seconds (instead of twice a second).  In this case there is no static, you just can only use 1 joystick every 2 seconds.

Not even a momentary dot?  :S

3 hours ago, johnnywc said:

I think it was suggested above that there may be a situation where all the chips are active for a split second during the transition that may be drawing more power which is causing the interference?

Yeah, but the open-collector outputs defeat that problem.  Since nobody is supplying a strong positive voltage for someone else to fight against, there should not be any especially strong power draw for this case even when two or more devices think they own the output.

 

[+johnnywc]

12 minutes ago, ChildOfCv said:

Well, it's possible.  But you should notice a definite change in voltage when plugging the capacitor in in that case.  You can hook up the capacitor through an amp meter, that is, "- stripe" to ground, + side to meter black wire, meter red wire to VCC, and meter in amps mode (if the meter requires a different plug for current measurements, make sure you use it), and see how much current is flowing through the capacitor/meter.

Okay, I do have a multimeter so I'll see if I can get a measurement.

12 minutes ago, ChildOfCv said:

 

It probably won't have enough impact on the circuit to matter.  So that would be one of those desperation moves

Let's hope it doesn't come to that.  

12 minutes ago, ChildOfCv said:

Not even a momentary dot?  :S

It's hard to tell since this is an RF connection and the picture isn't clear to begin with, but I was eagle-eyed on the area where it usually appears and didn't see anything.

12 minutes ago, ChildOfCv said:

Yeah, but the open-collector outputs defeat that problem.  Since nobody is supplying a strong positive voltage for someone else to fight against, there should not be any especially strong power draw for this case even when two or more devices think they own the output.

 

Okay, that makes sense.  

 

Thanks!

John

 

[ChildOfCv]

On 9/10/2019 at 12:07 PM, ChildOfCv said:

I may be able to test some things on my 7800, but it looks like even by the 2600Jr days they already noticed the folly of their power distribution scheme and modified it.  So the 7800 may be immune to these issues.  I do have a SECAM 2600 board that uses this stupid power scheme, but it seems to have a bad RIOT chip, at the least.  I'll have to socket the chips in my 7800, then maybe I can swap the RIOT from it and see if it works then.

Well, I socketed the chips in my 7800, then tried the old swap with the RIOT chip from the SECAM board.  It failed big time.  So it does indeed appear that the chip is dead.  I tried the good RIOT in the SECAM system, but still no joy.  I probed around on the CPU and found that only the clock pins showed life.  So it's probably dead too.  Joys.  I wonder how many other chips are dead?  TIA at least has the 900KHz and SYNC outputs, which is promising.  But that means I still can't test anything RF-related.  I found a 4-switch 2600 that claims to work and seems unmodified, so we'll see.

[ChildOfCv]

2 hours ago, ChildOfCv said:

Well, I socketed the chips in my 7800, then tried the old swap with the RIOT chip from the SECAM board.  It failed big time.  So it does indeed appear that the chip is dead.  I tried the good RIOT in the SECAM system, but still no joy.  I probed around on the CPU and found that only the clock pins showed life.  So it's probably dead too.  Joys.  I wonder how many other chips are dead?  TIA at least has the 900KHz and SYNC outputs, which is promising.  But that means I still can't test anything RF-related.  I found a 4-switch 2600 that claims to work and seems unmodified, so we'll see.

And then with another test to make sure the processor reset and ready pins are correct... and the processor is working.  WTF...

 

Now I just have to figure out why the color circuit is borked.  And it would also help to have a PAL version of the game.

 

[ChildOfCv]

 

On 9/11/2019 at 11:23 AM, johnnywc said:

Okay, that makes sense. 

I thought I'd post an update for those still playing along:

 

I built the circuit and tried it out on my Atari with the test program, and managed to see a little bit of the same interference.  In this case, the orange player's joystick must be held up and left with the fire button pressed, then there's a small short white line on the bottom left of the screen.  I used jumper wires to hold the joystick in that direction.  Using the other joystick with those connections in place, if I press in any direction or press the fire button, the interference disappears.  Bizarre.

 

I also tested an external power supply on the circuit instead of using pin 7.  No change.  Apparently, it's not the power usage that is causing this.  Well, at least not for the weird line.

 

 

[+johnnywc]

Thanks for the update!

 

It's weird that this only happens with non-modded Ataris using RF and there is absolutely zero interference with a modded Atari, even with the circuits that use 7404 and bus transceivers where the interference is much more prominent vs. the circuit with the NAND gates.  Could that be a clue? 

[ChildOfCv]

Modded Ataris aren't affected as much by small voltage shifts since they output thousands of times the voltage of the RF modulator.  1V vs <1mV.  The <1mV RF signal also presents an easier target for outside interference, not just distortion from the modulator itself.

 

Anyway, I discovered another head-scratching point while watching the inputs on a scope.  The voltage for a particular direction falls to 0 when you press in that direction (Duh) with some of the expected switch bounce.  But what I didn't expect is for the other directions to be bounced around by it too, certainly not to this degree!

 

Expected waveform for the direction pressed:

 

Unexpected waveform for other directions which were not pressed:

 

Also, the power filter seems to be working quite well:  The filter is keeping the breadboard's electronics nice and glitch-free, even when the Atari has weird spikes.

 

This seems to suggest that the RF's voltage spikes are not coming from the multitap.

[ChildOfCv]

So, don't get out the confetti just yet, but I may be on to something.

 

I've been researching what might cause power draw on these chips, and the common answer is state changes.  Obviously true on the final output, as standard logic chips have to pull the signal in both directions, and the faster they have to flip the state, the more power they draw.  We did reduce that a lot by using open collector, since now it only has to pull down.

 

But there is still remaining interference, and it's probably caused by the 6532 and the TIA.  After all, they're also logic chips and they can also see rapid state changes.

 

As noted in the above scope pictures, you see a huge amount of ringing whenever a direction is pressed, and less so when a direction is released.  This is caused by the switch contacts, which will have a period of bouncing just as a connection is made.  The logic chips react to this with an attempt to turn that bouncing into a square wave.  This takes power internally in the amplification stages.  Also, when a particular chip is turned on for output, it takes power to affect the output state.

 

The solution, probably, will be to eliminate the bounce.  For that, we use a debouncing circuit and a Schmitt trigger.  This is a model for the circuit, though there are a few choices for implementation.

 

[+johnnywc]

On 10/9/2019 at 7:09 PM, ChildOfCv said:

So, don't get out the confetti just yet, but I may be on to something.

 

I've been researching what might cause power draw on these chips, and the common answer is state changes.  Obviously true on the final output, as standard logic chips have to pull the signal in both directions, and the faster they have to flip the state, the more power they draw.  We did reduce that a lot by using open collector, since now it only has to pull down.

 

But there is still remaining interference, and it's probably caused by the 6532 and the TIA.  After all, they're also logic chips and they can also see rapid state changes.

 

As noted in the above scope pictures, you see a huge amount of ringing whenever a direction is pressed, and less so when a direction is released.  This is caused by the switch contacts, which will have a period of bouncing just as a connection is made.  The logic chips react to this with an attempt to turn that bouncing into a square wave.  This takes power internally in the amplification stages.  Also, when a particular chip is turned on for output, it takes power to affect the output state.

 

The solution, probably, will be to eliminate the bounce.  For that, we use a debouncing circuit and a Schmitt trigger.  This is a model for the circuit, though there are a few choices for implementation.

 

This sounds very promising!  Anything I can do to help?  Should I look into buying some of these chips?

 

Thanks!

John

 

[Danjovic]

You may Try adding 100 ohm to 220 ohm resistors in series with each input as they will dissipate part of the energy that is causing the ringing of the signals during commutation.

[ChildOfCv]

17 minutes ago, Danjovic said:

You may Try adding 100 ohm to 220 ohm resistors in series with each input as they will dissipate part of the energy that is causing the ringing of the signals during commutation.

That's also been in the back of my mind.  The only thing I don't like about it, though, is that the 6532 has pull-up resistors, so you need to have small enough resistors to get a reliable low.  The TIA's trigger input has a 10K pull-up resistor, a 220 ohm series resistor, and a 470pF capacitor, probably for a little debounce in case you don't want to use the input latch, so in any case that pin should be left alone.

 

Also, in my last experiment I used 74LCX760's.  These are open-collector, 5V tolerant, but low Vcc preferred.  So naturally I used a quick 10K/15K voltage divider to power them from the +5V.  With that, the ringing still caused multiple-volt power fluctuations.  I even disconnected Vmod entirely and used an external +5V source, and still saw ringing on Vmod, even though the breadboard's power was smooth.  This leads me to believe that the 6532 is also consuming a lot of extra power just reacting to the switch bounce, which was unexpected.  So I want to try eliminating the ringing at all rather than merely attempting to reduce its effects.

 

On the other hand, a joystick connected directly will directly throw that switch bounce full force at both chips, so why doesn't it cause problems?  Well, if you furiously jackhammer the joystick around, you can occasionally get a spark or two to appear on screen.  But definitely something you'd have to be watching for to notice.  The open-collector circuits should be approximating the exact same action:  Only pulling to ground.  Yet they can reliably cause snow.  So my hypothesis sounds good, but I admit that it has holes.

 

Well, the one other thing I've seen somewhat, is that the port connector needs to make good contact.  I got some cheap DB9 connectors that have screw terminals for easy wire attachment.  But they have way too much slop when plugged in, and sometimes things can be improved or worsened simply by shifting the connector around.  So this does leave one other item of total importance:  A solid ground wire.  Maybe there is some resistance on the ground wire, and that's raising the floor whenever power is required.  If so, then simply finding a GOOD plug might go a long ways.  So that will be something to consider on the final product.

 

I guess this is why the professional engineers get the big bucks.

 

[ChildOfCv]

5 hours ago, johnnywc said:

This sounds very promising!  Anything I can do to help?  Should I look into buying some of these chips?

Well, the circuit model there was just because Kicad didn't have the components I was planning on, so I used what it had.  I have some Schmitt trigger buffer ICs on order and a CD4019B data selector for final output.  Yeah it's back to full drive output, but if the switch bounce proves to be the issue, then this old low-power CMOS chip should be just fine.  It's also of note that the 2600 trackball uses the 4019 for output, so it's field-tested.

[ChildOfCv]

Oh right, another observation:

 

As noted, the joystick hardly makes any static when plugged directly in.  But if I plug in my screw terminal connector and clamp a wire into pin 8 and touch the joystick input pins with it, it reliably causes static.  So I can make the screen sparkle even without electronics.

[Danjovic]

Have you tried to simulate an open collector using a diode? 

Looking ob the opposite direction, maybe the 10k internal pullups are way too high, letting the 6532 inputs kinda float so they ring;  then external pullups like 1k to 2k2 could get you a more solid pullup voltage. 

[ChildOfCv]

Yeah, my current setup has 3.3K pull-ups on the board.  I noticed that the 7800 trackball uses open-collector outputs with 2.7K pull-ups, and I used my closest on-hand equivalents.  Anyway, it changed the personality of the static somewhat, but didn't reduce it any.

[ChildOfCv]

This setup seems to be the most successful so far.  It's two 74AHCV17's (non-inverting hex buffer with schmitt trigger inputs) and two CD4019B's (quad data selectors).  You might be able to make something work using the chips you've already acquired, but the point is that 1) it's a schmitt trigger input with resistor pullups and capacitor debounce.  2) There is no output contention.  The 4019B's use OR gates to combine the separate inputs, so there isn't a problem even when both are activated.

 

Since I only needed one NOT gate, I just used a 2N3904 transistor and a few resistors to make it.  Why waste an entire chip for one gate?

 

The capacitor values are just a guess.  I used 0.1uF, but they hold active for a noticeable time before returning, so that's not good for action games.  So I cut it by 10x in the schematic.  A good approach might be to use the 0.1uF capacitor with a 1M potentiometer, then keep adjusting the potentiometer until the response looks good to you.  Then the values can be normalized through chosen resistor and capacitor values.

 

Alternative ICs would be 74ls14 for the buffers and 74ls158 for the multiplexer.  The 158 has the additional benefit of not needing the transistor.

 

 

temp.pdf

Edited October 21, 2019 by ChildOfCv

[+johnnywc]

This looks pretty promising - thanks!  I was away at the Portland Retro Gaming Expo this weekend and we demoed the finished Wizard of Wor Arcade with a prototype QuadTari using 74HC245 transceivers and 74HC04 inverters using the DUMP_PORTS as the select line.  Functionally it worked great and since we were using a modded Atari there were no interference issues (the interference on an RF system is really bad as we know).

 

I would like to try to build this circuit over here so I can see how the interference looks and try a few different capacitors to see how the response is.  Would you happen to have any links to Mouser to the parts I should pick up?

 

How does the interference look on your end?

 

Thanks!

John

 

[Danjovic]

Do you thinhk that analog mux worth a try? It might be possible to add pin 9 on the muxing

 

Jmux.pdf

[ChildOfCv]

48 minutes ago, johnnywc said:

This looks pretty promising - thanks!  I was away at the Portland Retro Gaming Expo this weekend and we demoed the finished Wizard of Wor Arcade with a prototype QuadTari using 74HC245 transceivers and 74HC04 inverters using the DUMP_PORTS as the select line.  Functionally it worked great and since we were using a modded Atari there were no interference issues (the interference on an RF system is really bad as we know).

 

I would like to try to build this circuit over here so I can see how the interference looks and try a few different capacitors to see how the response is.  Would you happen to have any links to Mouser to the parts I should pick up?

 

How does the interference look on your end?

 

Thanks!

John

 

 

The static is gone, from what I could see.  Even removing the debounce capacitors to improve response time, it was still quiet except for an occasional tiny white dot on maybe 50 joystick/button presses.

 

 

Okay, so option 1 is the inverting buffer to inverting data selector:

74hc14: https://www.mouser.com/ProductDetail/Texas-Instruments/SN74HC14N?qs=sGAEpiMZZMuyBeSSR239Ias%2B1yN%2FIiBNq5xI0bbHe7g%3D

74ls158: https://www.mouser.com/ProductDetail/Texas-Instruments/SN74LS158N?qs=sGAEpiMZZMtxONTBFIcRfk9GpaUoLd0cKrV198MuqdA%3D

 

Another option is the non-inverting buffer to non-inverting data selector.  You'd have your choice of NOT gate plus 4019, or a 74ls157.  74AHC8541 is also a usable choice.  Actually, you can set it for inverting or not, so it could be made to work with either data selector.

*74AHCV17 https://www.mouser.com/ProductDetail/Texas-Instruments/SN74AHC8541N?qs=%2Fha2pyFaduildg0HBe%2FQufJuLq6%2FoFr3geyhxP%2Fr1Co%3D

(but also needs https://www.mouser.com/ProductDetail/Aries-Electronics/LCQT-TSSOP14?qs=%2Fha2pyFaduh3s%2F01dvn%2Fga6dGKUFOQKWguYfMiqGSvvV40Tsjgf1HA%3D%3D for use with a breadboard)

74AHC8541 https://www.mouser.com/ProductDetail/Texas-Instruments/SN74AHC8541N?qs=%2Fha2pyFaduildg0HBe%2FQufJuLq6%2FoFr3geyhxP%2Fr1Co%3D

 

74ls157: https://www.mouser.com/ProductDetail/Texas-Instruments/SN74LS157N?qs=sGAEpiMZZMtxONTBFIcRfpKvrGTByBdiIFx8m8XZjlw%3D

*CD4019B https://www.mouser.com/ProductDetail/Texas-Instruments/CD4019BE?qs=sGAEpiMZZMtMa9lbYwD6ZHl%2Bb36l%2BC4RxjVZZsInI%2BQ%3D

and 2N3904 

 

I marked the ones I actually used with an *

 

1M potentiometer (good enough to dial in a response time) https://www.mouser.com/ProductDetail/BI-Technologies-TT-Electronics/P160KN-0QD15B1MEG?qs=sGAEpiMZZMtC25l1F4XBU7XpbFdDDzhr3l49V2gsqx8%3D

 

When I looked at the 2600 trackball schematic again (the one that's part of the 7800 trackball schematic) I noticed that the less-official schematic is off by an order of magnitude on the inductor.  So here's a 150uH one that matches what the 2600 uses:  https://www.mouser.com/ProductDetail/Bourns/78F151J-RC?qs=sGAEpiMZZMsg%2By3WlYCkU0AkrlqHXBx5xpBcD11zi1c%3D

 

Finally, with the trip point of the schmitt triggers at around 3V, it may be useful to reduce the line voltage to the joysticks by a bit.  So instead of a single 1M pull-up, maybe it needs a 330K pull-up and 680K pull-down.  Something like this:

[ChildOfCv]

6 minutes ago, Danjovic said:

Do you thinhk that analog mux worth a try? It might be possible to add pin 9 on the muxing

 

Jmux.pdf 12.78 kB · 1 download

That's an interesting potential solution, and theoretically it should work, since analog states shouldn't pull any extra power from the multiplexer.  Might be fun to try.  Heck, it may even work better than the digital solution.