Yet Another Masterplay Clone

[+DrVenkman]

Are you seeing good resistance changes between pin 15 with 10 and 11 and +5V between pin 12 and 15 on the 5200 connector?

 

I'm not sure I understand your question. Are you referring to the Masterplay clone board or the connector on the 5200 itself?

 

Like I said, my CX-52's work, my Trak-Ball works, and the thumbstick boards of Scott's I built two weeks ago work. I just don't get what's going on here.

 

I think, given Scott's comment earlier, that I'll build my second board tomorrow with 470K resistors at R3 and R4 and see how it goes.

[smbaker]

Are you seeing good resistance changes between pin 15 with 10 and 11 and +5V between pin 12 and 15 on the 5200 connector?

 

I'm not sure that you can directly read the resistance between pin 9 and 10 (or 9 and 11) while the 5200 is powered up. Maybe. I've just not tried it. This is an RC circuit and the 5200 is occasionally charging a capacitor through the resistor and letting it discharge. My expectation would be that would mess with a multimeter. Safest bet would be to disconnect pin 9 from the 5200 and then take the measurement. Alternatively, disconnect the whole board from the 5200 and power it up independently. Maybe I'll try this on mine and see what happens, for the sake of knowledge.

 

Scott

[Swami]

 

I'm not sure I understand your question. Are you referring to the Masterplay clone board or the connector on the 5200 itself?

 

Like I said, my CX-52's work, my Trak-Ball works, and the thumbstick boards of Scott's I built two weeks ago work. I just don't get what's going on here.

 

I think, given Scott's comment earlier, that I'll build my second board tomorrow with 470K resistors at R3 and R4 and see how it goes.

 

Sorry, I was referring to the 5200 connector on the MPC board, since your 5200 itself seems to be operating correctly.

[Swami]

 

I'm not sure I understand your question. Are you referring to the Masterplay clone board or the connector on the 5200 itself?

 

Like I said, my CX-52's work, my Trak-Ball works, and the thumbstick boards of Scott's I built two weeks ago work. I just don't get what's going on here.

 

I think, given Scott's comment earlier, that I'll build my second board tomorrow with 470K resistors at R3 and R4 and see how it goes.

 

Sorry, I was saying pin 15 for pots when I meant 9. You can check the resistance between 9 and 10 and 11 on the MPC 5200 connector with it not plugged in on the MPC 5200 connector and check the voltage swings between 9 and 10 and 11 with it plugged in, as well as +5V on pin 9 and 12. I would also check the resistance swings at the end of the extension cord where it connects to the 5200 to make sure you are getting good contact between the extension cord and the MPC connector and that the cord is good.

[+DrVenkman]

Well, the extension cord(s) are brand new from Console5 and work fine with the thumbstick controllers I built, so I think they're not the root cause of my problems, whatever else might be going on. Unfortunately, I don't have a handy way to power the controller independently from the 5200.

 

I just can't help thinking there's *GOT* to be something more fundamental going on here that I've just missed, given that all the rest of my controllers are working. I'm going to build out my second board tomorrow with 470K resistors in place of the 10M ohm resistors and we'll see how it goes.

 

Thanks Scott and Swami for the continued patience and suggestions.

[john_q_atari]

Other than building the other board all I can think of at the moment is verify the conductivity with a multimeter of every schematic net to try and eliminate bad traces, a broken wire, or perhaps a bad solder connection. If it's not the 5200 or the extention cable and unlikely the chip then that leaves the board or connections to it. If that is good then it's off to Petes and or scoping the lines.

[Swami]

Well, the extension cord(s) are brand new from Console5 and work fine with the thumbstick controllers I built, so I think they're not the root cause of my problems, whatever else might be going on. Unfortunately, I don't have a handy way to power the controller independently from the 5200.

 

I just can't help thinking there's *GOT* to be something more fundamental going on here that I've just missed, given that all the rest of my controllers are working. I'm going to build out my second board tomorrow with 470K resistors in place of the 10M ohm resistors and we'll see how it goes.

 

Thanks Scott and Swami for the continued patience and suggestions.

 

Were you seeing full swings in the resistance between pins 9 with 10 and 11? You don't need power for that. Also, do you have the two diagnostic cart roms or carts to check controller operation? Anyhow, good luck with the second build.

[+DrVenkman]

 

Were you seeing full swings in the resistance between pins 9 with 10 and 11? You don't need power for that. Also, do you have the two diagnostic cart roms or carts to check controller operation? Anyhow, good luck with the second build.

 

Without the board powered, I don't see any changes at all ("OL" on the meter), trying to measure between pins 9 - 10 and 9 - 11. Of course, it's awkward has hell to run solid-core jumper wires into the 15-pin female connector on the board and try to make contact with my meter probes (no alligator clips, sadly) and also manipulate the joystick to test. So it might just be I was too tired last night to come up with a better way to rig up a test. If I get super-motivated I might run leads from the connector to a breadboard, hold the meter lead probes to the breadboard, then use my other hand on the joystick … Eh, maybe a better idea will present itself.

 

Anyway, I tested more today and got some weird results with Missile Command: basically I can "center" the cursor into pretty small rectangular area in the upper-left of the screen. From there, a controller in the 9-pin port will move the cursor around the small box - up/down/left/right all work, within that small bounding box. And of course, in any game I test if I unplug the 15-pin cable, the cursor/ship/character will immediately head right and down, which makes sense - infinite resistance (open loop) on both controller axes.

 

So it almost looks like I've got a scaling problem? Does that make sense?

 

Edited: typos. Bah.

Edited October 21, 2018 by DrVenkman

[Swami]

 

Without the board powered, I don't see any changes at all ("OL" on the meter), trying to measure between pins 9 - 10 and 9 - 11. Of course, it's awkward has hell to run solid-core jumper wires into the 15-pin female connector on the board and try to make contact with my meter probes (no alligator clips, sadly) and also manipulate the joystick to test. So it might just be I was too tired last night to come up with a better way to rig up a test. If I get super-motivated I might run leads from the connector to a breadboard, hold the meter lead probes to the breadboard, then use my other hand on the joystick … Eh, maybe a better idea will present itself.

 

Anyway, I tested more today and got some weird results with Missile Command: basically I can "center" the cursor into pretty small rectangular area in the upper-left of the screen. From there, a controller in the 9-pin port will move the cursor around the small box - up/down/left/right all work, within that small bounding box. And of course, in any game I test if I unplug the 15-pin cable, the cursor/ship/character will immediately head right and down, which makes sense - infinite resistance (open loop) on both controller axes.

 

So it almost looks like I've got a scaling problem? Does that make sense?

 

Edited: typos. Bah.

 

If you have any of these, they work really well for the difficult pin testing, male or female version, depending. Avoid the ones that have the large nuts next to the connector, although they can be forcibly removed, if that's what you happen to have.

 

https://www.ebay.com/itm/DB15-Female-D-SUB-2-Row-15Pin-Plug-Breakout-Terminal-Solderless-Connector-Screw/331950758687?epid=1183959383&hash=item4d49ce9b1f:g:3mYAAOSwiBJaDlA5:rk:2:pf:0

[+DrVenkman]

I don't, or I would've rigged one up already.

 

In any case, it's pretty clear that I've got continuity at least, all the way through. I mean, I can "center" in a very small range and the cursor (in Missile Command) responds as expected to Up/Down/Left/Right and FIRE commands. But I'd like to know what's going on with the scaling.

 

I just took a look at the 5200 Field Service Manual. I've not taken a look at it since I bought my 5200 about 15 months ago. When I got the machine, at least one of the 4052 chips had failed - keypad commands weren't registering on Port 1, though directions worked with my cleaned CX-52 stick and my Trak-Ball. So I replaced them (all 5 to be safe) with modern 74HC4052AP replacements (made in mid-2016). Does anyone think that might be part of the issue, even though my Trak-Ball and CX-52 work?

[Swami]

I don't, or I would've rigged one up already.

 

In any case, it's pretty clear that I've got continuity at least, all the way through. I mean, I can "center" in a very small range and the cursor (in Missile Command) responds as expected to Up/Down/Left/Right and FIRE commands. But I'd like to know what's going on with the scaling.

 

I just took a look at the 5200 Field Service Manual. I've not taken a look at it since I bought my 5200 about 15 months ago. When I got the machine, at least one of the 4052 chips had failed - keypad commands weren't registering on Port 1, though directions worked with my cleaned CX-52 stick and my Trak-Ball. So I replaced them (all 5 to be safe) with modern 74HC4052AP replacements (made in mid-2016). Does anyone think that might be part of the issue, even though my Trak-Ball and CX-52 work?

 

It doesn't sound like the 5200 console has an issue. If you can force the cursor to go right or down by turning the pot, it sounds like it can move R/D, but the joystick cannot make it move R/D. I'd say you can at least narrow it down to being between the 9 pin joystick port and before the pots, especially since when it is in an "always right" state, you can still move left.

[+DrVenkman]

 

It doesn't sound like the 5200 console has an issue. If you can force the cursor to go right or down by turning the pot, it sounds like it can move R/D, but the joystick cannot make it move R/D. I'd say you can at least narrow it down to being between the 9 pin joystick port and before the pots, especially since when it is in an "always right" state, you can still move left.

 

I don't think you understood my post (or perhaps I've misunderstood yours, which is distinctly possible). What I meant was, with Missile Command, I can trim the pots such that the cursor is centered in a "bounding box" (for lack of a better term) in the upper-left quadrant of the screen. Well, it's smaller than a quarter of the screen but you get the idea. Within that box, with the joystick untouched, the cursor doesn't move. From that centroid point, the cursor responds to joystick commands - right and down as well as up and left. If the entire game was played in that small "window" of the screen, I could play it.

 

Anyway, I built my second board this afternoon, this with 470K resistors at R3 and R4 in place of 10M. Sadly, it made essentially no difference at all. I just can't for the life of me figure out what's going on, unless it's something to do with the Console5 extension cable - perhaps it's not allowing sufficient contact with the female jack on the MPC connector; I tried the cord with one of my CX-52s and while it does indeed work, the CX-52 connector slides into the Console5 cable about a 1/8" or more deeply and it's DEFINITELY a firmer connection; plugged into the jacks on my MPC and thumbstick boards, I wiggle the connector and manipulate the cursor as resistance/timer readings from the 5200 change.

 

For now, I am at an impasse unless someone else has a major new idea; I'd suggest I just totally screwed up both builds but despite my lifelong habits of self-doubt at critical junctures, I've repaired numerous old Atari computers and consoles in the past, and managed to build an entire 1088XEL computer from parts last winter all by myself, along with two working examples of Scott's thumbstick controllers.

 

So I am at an impasse for this week at least. Next weekend I'll buy a copy of Pete's Test Cart and some larger 15-pin cable ends that I can screw into the jacks on the MPC and thumbstick boards. I'll be in a better position to figure out what's going on once I can see the on-screen measurements from the test cart, as well as ensure the cable isn't just unplugging itself from the board enough to affect readings.

[Swami]

 

I don't think you understood my post (or perhaps I've misunderstood yours, which is distinctly possible). What I meant was, with Missile Command, I can trim the pots such that the cursor is centered in a "bounding box" (for lack of a better term) in the upper-left quadrant of the screen. Well, it's smaller than a quarter of the screen but you get the idea. Within that box, with the joystick untouched, the cursor doesn't move. From that centroid point, the cursor responds to joystick commands - right and down as well as up and left. If the entire game was played in that small "window" of the screen, I could play it.

 

Anyway, I built my second board this afternoon, this with 470K resistors at R3 and R4 in place of 10M. Sadly, it made essentially no difference at all. I just can't for the life of me figure out what's going on, unless it's something to do with the Console5 extension cable - perhaps it's not allowing sufficient contact with the female jack on the MPC connector; I tried the cord with one of my CX-52s and while it does indeed work, the CX-52 connector slides into the Console5 cable about a 1/8" or more deeply and it's DEFINITELY a firmer connection; plugged into the jacks on my MPC and thumbstick boards, I wiggle the connector and manipulate the cursor as resistance/timer readings from the 5200 change.

 

For now, I am at an impasse unless someone else has a major new idea; I'd suggest I just totally screwed up both builds but despite my lifelong habits of self-doubt at critical junctures, I've repaired numerous old Atari computers and consoles in the past, and managed to build an entire 1088XEL computer from parts last winter all by myself, along with two working examples of Scott's thumbstick controllers.

 

So I am at an impasse for this week at least. Next weekend I'll buy a copy of Pete's Test Cart and some larger 15-pin cable ends that I can screw into the jacks on the MPC and thumbstick boards. I'll be in a better position to figure out what's going on once I can see the on-screen measurements from the test cart, as well as ensure the cable isn't just unplugging itself from the board enough to affect readings.

 

I was referring to what you'd said about Galaxian earlier: how you could turn the pot far enough that it would move right without the joystick and still move left with the joystick. That was with the pot on the MPC, not the one inside the console, right? I thought from your post about Missile Command it was moving just a couple pixels, but if it moves a lot, but only half of the expected distance in all directions, that is truly strange. It should be all or nothing in missile command. It sounds like instead of the voltage being delivered to the controller port being 0V for left, 2.5V for center and 5V for right, it is only going from 0V to 2.5V from left to right. Same for up to down. I don't know if that explains the Galaxian behavior you described, though, or if you see this half-range motion in the other parts of the missile command screen. It almost sounds like a too small of a capacitance issue or the pots are half range, but it seemed to work for Scott's, so that seems like a stretch unless something could be throwing it off. Sorry I can't say anything more than that. Hopefully, the test cart will explain something.

[smbaker]

So I am at an impasse for this week at least. Next weekend I'll buy a copy of Pete's Test Cart and some larger 15-pin cable ends that I can screw into the jacks on the MPC and thumbstick boards. I'll be in a better position to figure out what's going on once I can see the on-screen measurements from the test cart, as well as ensure the cable isn't just unplugging itself from the board enough to affect readings.

 

Another thing you could try jumpering across each diode. This is another one of those "I can't possibly see what difference it should make" situations, but the circuit seems to work fine with the diodes bypassed, and that would eliminate another potential source of component difference.

 

At this point, I can't imagine you've built the thing wrong. I'm worried there's some weird as-yet-not-understood compatibility issue. Hopefully we'll get some reports in from the other builders.

 

Having a Pete's test cartridge will definitely make it easier to see some real numbers, though it's unfortunate you have to spend the additional money to buy one. I wish this was going more smoothly.

 

There is a controller related pot you can adjust inside the console, though I would not recommend doing so at this time, considering that your other controllers all work fine.

 

Scott

[+DrVenkman]

 

I was referring to what you'd said about Galaxian earlier: how you could turn the pot far enough that it would move right without the joystick and still move left with the joystick. That was with the pot on the MPC, not the one inside the console, right? I thought from your post about Missile Command it was moving just a couple pixels, but if it moves a lot, but only half of the expected distance in all directions, that is truly strange. It should be all or nothing in missile command. It sounds like instead of the voltage being delivered to the controller port being 0V for left, 2.5V for center and 5V for right, it is only going from 0V to 2.5V from left to right. Same for up to down. I don't know if that explains the Galaxian behavior you described, though, or if you see this half-range motion in the other parts of the missile command screen. It almost sounds like a too small of a capacitance issue or the pots are half range, but it seemed to work for Scott's, so that seems like a stretch unless something could be throwing it off. Sorry I can't say anything more than that. Hopefully, the test cart will explain something.

 

In Missile Command, it's more than a few pixels but less than half the screen in any direction. I can get it to move around a square maybe 20% (?) of each axis. That said, that range of resistance sensitivity isn't enough to get any other game I tested to move down or right at all; their code just doesn't register that small a change. FWIW, when it works this way, when I unplug the board and pull the IC, the trim pots are set around 800K - 1M each.

 

 

Another thing you could try jumpering across each diode. This is another one of those "I can't possibly see what difference it should make" situations, but the circuit seems to work fine with the diodes bypassed, and that would eliminate another potential source of component difference.

 

At this point, I can't imagine you've built the thing wrong. I'm worried there's some weird as-yet-not-understood compatibility issue. Hopefully we'll get some reports in from the other builders.

 

Having a Pete's test cartridge will definitely make it easier to see some real numbers, though it's unfortunate you have to spend the additional money to buy one. I wish this was going more smoothly.

 

There is a controller related pot you can adjust inside the console, though I would not recommend doing so at this time, considering that your other controllers all work fine.

 

Scott

 

I was thinking of the diodes myself when you posted, so I jumpered them (soldered a cut-off resistor leg across each one). Absolutely no change. Oh, well.

 

And I also rather doubt anything is really up with the console internals - after I messed around with Missile Command on the MPC board, I powered down, plugged in a thumbstick controller and booted back up. The cursor was dead-center in the middle of the screen as expected and moved around perfectly. And I did manage to set a personal best in Defender today with the thumbstick at almost 77,000.

 

I truly do appreciate your efforts, both of you, to try to help figure out what's going on. I honestly have no clue. But no worries - I do enjoy a good challenge. I'll get a copy of Pete's Test Cart next week or maybe just suck it up and buy an AtariMax cart. Been meaning to get one anyway - I presume there's a rom image of the test cart I'll be able to load onto it?

[Swami]

 

In Missile Command, it's more than a few pixels but less than half the screen in any direction. I can get it to move around a square maybe 20% (?) of each axis. That said, that range of resistance sensitivity isn't enough to get any other game I tested to move down or right at all; their code just doesn't register that small a change. FWIW, when it works this way, when I unplug the board and pull the IC, the trim pots are set around 800K - 1M each.

 

 

I was thinking of the diodes myself when you posted, so I jumpered them (soldered a cut-off resistor leg across each one). Absolutely no change. Oh, well.

 

And I also rather doubt anything is really up with the console internals - after I messed around with Missile Command on the MPC board, I powered down, plugged in a thumbstick controller and booted back up. The cursor was dead-center in the middle of the screen as expected and moved around perfectly. And I did manage to set a personal best in Defender today with the thumbstick at almost 77,000.

 

I truly do appreciate your efforts, both of you, to try to help figure out what's going on. I honestly have no clue. But no worries - I do enjoy a good challenge. I'll get a copy of Pete's Test Cart next week or maybe just suck it up and buy an AtariMax cart. Been meaning to get one anyway - I presume there's a rom image of the test cart I'll be able to load onto it?

 

Yes there is a rom for it and also for the diagnostic cart. I also have a rom that just shows the stick position in hexadecimal for each axis.

[+DrVenkman]

Just wanted to let Scott and anyone else interested know that I ordered a copy of Pete's Test Cartridge from the AA Store today. I know Al is still recovering from the PRGE and no doubt will be catching up on backorders for a week or two, but hopefully I'll have another tool to help me try to figure out why my MPC boards aren't working as expected. I'll get this figured out eventually, I have no doubt.

[smbaker]

Just wanted to let Scott and anyone else interested know that I ordered a copy of Pete's Test Cartridge from the AA Store today. I know Al is still recovering from the PRGE and no doubt will be catching up on backorders for a week or two, but hopefully I'll have another tool to help me try to figure out why my MPC boards aren't working as expected. I'll get this figured out eventually, I have no doubt.

 

I'm still hoping this is just an isolated incident, and not an issue with the design itself.

 

There should be another 3 or 4 people with pcboards in hand by now, has anyone else built one yet?

 

Scott

[+DrVenkman]

 

I'm still hoping this is just an isolated incident, and not an issue with the design itself.

 

There should be another 3 or 4 people with pcboards in hand by now, has anyone else built one yet?

 

Scott

 

UPDATE! AND GOOD NEWS!(*)

 

I dug my second 5200 out of storage and whaddaya know? The MPC works GREAT! So the bottom line is no design flaw. I only tested it with one board (the one I built using 470K ohm resistors in place of the 10M ohm resistors at R3 and R4) but it works just fine with both Defender and Ms. Pac-Man. Playing Defender with the Smart Bomb control right there on the Genesis pad is awesome!

 

 

(*) Now the caveat - there's a reason this 5200 was in storage; after about 10 minutes it started glitching out all over the place in Defender. First the screen glitched, then it started leaving garbage graphics trails across the screen in Defender. So I'm guessing CPU, GTIA, ANTIC or possibly RAM. I can play swap-the-chips with most of those (though I'm running low on all of them in terms of true "spares." I'll probably just suck it up and buy spares for all of them at BEST in a couple weeks. In the meantime, this gives me something to think about with regard to the issues on my first system. I'd strongly suspect POKEY except that my thumbsticks, CX-52s and Trak-Ball still work fine. But I'll wait until I get my copy of Pet's Test Cart and then figure out what's going on.

[smbaker]

That is good news!

 

There is that potentiometer on the 5200 main board that adjusts "joystick centering". I'm not sure whether misadjustment/drift of that pot could be responsible for the issue or not, but it's something to take a look at once you have your Pete's. Service manual I think tells how to adjust the pot.

 

Scott

[+DrVenkman]

That is good news!

 

There is that potentiometer on the 5200 main board that adjusts "joystick centering". I'm not sure whether misadjustment/drift of that pot could be responsible for the issue or not, but it's something to take a look at once you have your Pete's. Service manual I think tells how to adjust the pot.

 

Scott

 

Okay, I got my copy of PETE'S TEST CART from the Store today. I was hoping it would show the POKEY resistor in this 5200 all out of whack. Unfortunately, that's not what I'm seeing.

 

The first is the Masterplay clone, the second is one of the thumbstick controllers I built, the third is one of my CX-53's.

 

 

Leftward motion of an attached stick or gamepad drops the value down to about 9, same for vertical movement. Right or down movement isn't being registered at all. Motion of my thumbsticks or CX-53 on the same port is normal - 0 to left/up, off-scale high for down/right.

Sadly, the second 5200, the one the MPC worked with a couple weeks ago, seems down for the count. I tried swapping some likely suspects like GTIA and CPU, and then I reseated all the socketed chips on the board to try to cure the glitching but now the unit won't boot at all. I started another thread about that issue, but haven't yet been able to really devote myself to that project yet.

[smbaker]

44/43 on center is too low. If you adjust the pots on the masterplay adapter, can you get it to read around 110 on center? Regardless, correcting center probably isn't going to affect the movement to the right.

 

Scott

[+DrVenkman]

44/43 on center is too low. If you adjust the pots on the masterplay adapter, can you get it to read around 110 on center? Regardless, correcting center probably isn't going to affect the movement to the right.

 

Scott

 

I can max out the trim pots to give me readings in the upper 50's/lower 60's range. Movement up and left still drops readings to about 9 on the measurement screen, but down and right each register, but only drop by one unit in the numeric screen, and almost imperceptible movement on the graphical screen. So it's almost as if the resistor values I've got are all wrong, except I measured each individual resistor as I installed it, so I don't think that could be the root cause.

 

I sure wish my flaky 5200 was booting up - I'd desperately like to know what readings the test cart show there for all my various controllers.

Edited November 11, 2018 by DrVenkman

[smbaker]

 

I can max out the trim pots to give me readings in the upper 50's/lower 60's range. Movement up and left still drops readings to about 9 on the measurement screen, but down and right each register, but only drop by one unit in the numeric screen, and almost imperceptible movement on the graphical screen. So it's almost as if the resistor values I've got are all wrong, except I measured each individual resistor as I installed it, so I don't think that could be the root cause.

 

I sure wish my flaky 5200 was booting up - I'd desperately like to know what readings the test cart show there for all my various controllers.

 

Well that's interesting, and shows the Pete's cartridge has provided some useful info. If you remove the DG413 IC, what happens?

 

Scott

[+DrVenkman]

 

Well that's interesting, and shows the Pete's cartridge has provided some useful info. If you remove the DG413 IC, what happens?

 

Scott

 

Offscale high readings on both axes.