Yet another use for the Radica Space Invaders joystick :)

[Paranoid]

I was feeling pretty confident about understanding it last night before bed, (although I'm pretty sure I dreamt about this damn project last night). This confirms that I've got a pretty good handle on what I need to do (hopefully I'll be able to actually DO it, not just understand HOW to do it... )

 

I had one more minor question. According to your schematic, Pin 2 has no connection, but you've got it wired through to the 9pin pass-thru anyhow. Pin 2 is the only key used *only* for keypad numeric entry?

 

I can't believe we're up to 3 pages on this thread and it is mostly me asking question after question. I've got it all now, and I'll be finishing up today and maybe tomorrow. I'll let you know my results once I get it set up and tested. Thanks once again for all your help and guidance on this.

Edited July 14, 2006 by Paranoid

[PacManPlus]

Yep - Pin 2 (the keypad center column) is the only pin not mapped to a function on the stick itself.

 

Cool and good luck!

Bob

[Shawn]

@paranoid&Bob - thanks for having the conversation publicly cause it's answering ALL the Q's anyone else would ask if they where making one of these sticks. Hell, I'm getting all my Q's answered just from reading and I have not had to ask a damn thing...yet.

 

Paranoid - Will you please post a pic of your work when its done? Just a pic of the guts would do great. Oh, there's my 1st Q

Edited July 14, 2006 by Shawn Sr.

[Paranoid]

Heh. I will, if I succeed. I've hit a major wall here. I've got the passthru working fine, I've tested it with a Atari keypad and everything works great. Even guessed the code on Countermeasure.

 

But I cannot get it to see the joystick, no matter what I do. I actually pulled apart the entire circuit I did, and rewired all the resistors onto 20ga. from the 18ga. wire (is that right? 20 ga. is thinner?) Whichever. I got rid of the solid copper core wire and used thinner braided core wire, then rehooked it all up. It is much cleaner inside now, but it still won't work.

 

Is it possible that I've burnt out the resistors by being too heavy-handed with the soldering iron, or blew them out with the Multi-meter? Once I hook up the sticks to 9, 10 and 11, it still acts exactly like it acts with no joystick attached at all. In Battlezone, for example, I just spin in circles to the right...

 

I *guess* I could have a bad cable. I pulled it from a stick that I've never confirmed the directionals on (which is why it was that one I decided to sacrafice. But I'd like to be certain before I part out another stick for this project.

 

And I *still* haven't gotten to wiring the start, pause, reset, * and # on the stick.

 

I'm attaching my latest schematic. This is what I'm following. I've done it twice, so I'm pretty sure I'm doing it *exactly* as it appears here. Bob, could you look it over and see if there are any mistakes in my wiring? I'm getting a real good feel for how the circuit works, which circuits tie into which other circuits. If everything looks good, I guess I'll try making the circuits again using new resistors, wire those in, and see if I get any better results. If it fails again, I guess I could buy a new 5200 cable from Best Electronics.

 

 

Edited July 15, 2006 by Paranoid

[Paranoid]

There is SOMETHING wrong with this schematic, too. I just can't put my thumb on it. I can SEE it... the left/right up/down are just a LITTLE different, and I can't figure out where I went wrong. GAH.

 

Nope... I drew them different, but the top and bottom is connected the same way.

 

I'm stumped.

Edited July 15, 2006 by Paranoid

[Paranoid]

 

 

So, I just cut all the resistors out, carefully, and I noticed something. Once I had them out, they're all in a sequential chain. Basically, I could solder all 6 of the resistors in sequence, then solder them to the appropriate contacts on the switches. Is that right? Or have a seriously screwed up the wiring here?

[Paranoid]

Well... being that I'm officially obsessing over this project... I couldn't put it down and walk away from it, and I finally pulled out the multimeter and confirmed the resistance on both circuits that I pulled from the stick. They're within 5k ohms of what they should be... each individual circuit tests right within it's range, and all of them together come in around 750k Ohms... so, that isn't it.

 

I checked my solder points on the switches... those are good too.

 

I started checking out the cable... and sure enough... Pin 9 isn't showing up at all. 10 and 11 show up from end to end, but 9 is dead. Must be a short in it somewhere.

 

Bah.

[eegt97]

 

 

So, I just cut all the resistors out, carefully, and I noticed something. Once I had them out, they're all in a sequential chain. Basically, I could solder all 6 of the resistors in sequence, then solder them to the appropriate contacts on the switches. Is that right? Or have a seriously screwed up the wiring here?

 

Umm, take out that short you have around the 250k? Looks like your making good progress, just some troubleshooting to go. A second thanks for the online reporting guys!

 

I think I'm going to head out and see if I can pick up a couple of the RSI sticks. Hopefully, they can still be found.

 

-Brian

[Paranoid]

Heh. On the last schematic, with the short AROUND the 250k, I'm not sure if this is how I really have it wired. It is a different way of looking at the circuit, with all the resistors in series. Once you assemble it using PMP's or my schematic, you end up with that, all the resistors in series... It is all a matter of where leads are tied from the resistors back to the switches. It was kind of confusing to look at it this way, so that diagram is more of a guess, and the short around the 250k was something I drew in AFTER I had drawn the rest, because it seems like there needs to be a path from NO/UP through the 2.2k to NC/Down. I think what I'll do is re-attach the circuit with all the resistors in series that I pulled, and document where I tie in to the circuit, as I go. That should give an accurate picture of the above schematic.

 

I pulled that entire 5200 cable apart last night, and tested it pretty extensively, and I'm positive there is an open short on Pin 9. It doesn't matter at this point, because the "exploratory surgery" pretty much destroyed the cable... but I was curious. Everything else I can pass current through from the db-15 connector to the wire-end of the cable, but pin 9 never even made the LCD move.

 

So... once I get done cleaning up the opposum that my dog killed last night, I'm gonna go find another stick to sacrafice to this project. I think I'm gonna get this stick cleaned up and tested before I pull the cable, though.

 

I figure by the time I get this all figured out, I'll be able to through together a second one pretty fast. I don't know that I'd want to do it as a service the way JayBird does the 7800 RSI stick, though, as it is WAY more invasive on the joystick, adding a slot for the DB-15 and holes for the 5 new buttons. It would be fairly easy to make a mistake and end up with at LEAST a very ugly stick, with this project.

 

Good luck with the sticks. I wish I had been able to find about 4 more when they dropped to $5 at Game Spot. I ended up driving all over town for the 2 I have. There were others around, but they were all too long of a drive. I've got a feeling I'll end up regretting the decision not to make the drive.

 

Alternatively, most of the Jakks sticks can be used in place of the RSI stick on this project. I've converted a Jakks Ms. Pac Man to the 7800 RSI... it should work for this one too. You might want to forego the on-stick buttons and just use the 5200 keypad and start/pause/reset if you went that way, though.

[Paranoid]

Update on the schematics for the stick that makes it much clearer how the circuit works. With this in hand, re-attaching the resistors is going to be a cinch, and I'm pretty sure I can get the two circuits into the battery compartment case, making it much easier to finish up the wiring inside the case of the stick itself. I'm even considering attaching the resistors to the switch contacts with quick disconnects, so that it'll be easy to open up the case and detatch and reattach everything.

 

 

[eegt97]

Good luck with the sticks. I wish I had been able to find about 4 more when they dropped to $5 at Game Spot. I ended up driving all over town for the 2 I have. There were others around, but they were all too long of a drive. I've got a feeling I'll end up regretting the decision not to make the drive.

 

Alternatively, most of the Jakks sticks can be used in place of the RSI stick on this project. I've converted a Jakks Ms. Pac Man to the 7800 RSI... it should work for this one too. You might want to forego the on-stick buttons and just use the 5200 keypad and start/pause/reset if you went that way, though.

It was 50 miles to the nearest Game Spot that had these (Mmm, about $10 in gas these days). So I ordered from their site. They have a coupon for free value shipping, so it will take a few days, but at least it's still 5 bucks a pop. Seems well worth it, since everyone has said this is perfect for these kinds of projects. Having all the *major* buttons on the controller itself will be nice.

[Paranoid]

Ok... so... this morning I woke up, ready to tear into the stick and fix it, and my wife comes in and goes, "The heat is too much in the front room since the fan broke. I'm going to fix it."

 

Which, I've already tried to do (read: Blew compressed air into the engine). So, to make a long story short, I spent the whole day installing a !@$*(*@!#(ing ceiling fan.

 

So I just *finally* finished the stick. I'll have pictures up tomorrow. I've got the resistor circuits in the battery compartment. At some point, though, I must have gotten confused, because UP is down and down is UP, and Left is Right and Right is Left.

 

Which works out OK, because you just have to turn the joystick upside down. That makes the Start, Pause and Reset Reset and Pause, and also reverses # and *... but it still works, and it seems like more trouble than it is worth to actually crack the thing back open and fix it.

 

I've played a couple of games. Berzerk was completely unplayable. BCs Quest was marginally improved... Donkey Kong was far better.

 

The PROBLEM, and what is making Berzerk and BC Quest not so much fun, is that it is getting almost constant false DOWN inputs. It isn't absolutely constant. I can get it to level off and stay center. But in general, it'll spit out false downs pretty consistently.

 

Any idea what is causing this? If I can nail this problem, it'll be a GREAT stick. I don't understand enough about the circuit to understand what makes it think it is getting a DOWN directional input. Which resistor is controlling the down? And could it be that I don't have ENOUGH resistance, or that I have too much, that is causing it to see false input?

[Paranoid]

I opened it up, and I had forgotten to cover the red and brown wires with electrical tape where they were connected to the stick. I was hoping that maybe they were grounding out on something, causing the false inputs. I wrapped them, and also found a small sliver of wire in the joystick housing. I removed that, too. The problem is *better*... but it still seems to interpet the odd UP input as a down input. I think there must be some more wire inside the stick itself causing a intermittent short.

[Paranoid]

Ok... here you go. Finished pictures. Because I wired it backwards/upside down, the orientation is correct for how you would hold it during play, which works ok. I've checked it out on Star Raiders, and the keypad passthrough works fine. I'm still getting occassional ghost DOWN inputs when I press directional up, and I have no idea what the cause is. I suppose if I had wired it correctly, I would expect occasional ghost UP inputs when I was pressing down. It is still very servicable, and for games like Popeye and PacMan it makes a big difference, probably Pitfall too... I'll have to try that. It doesn't work well for BCs Quest, because the ghost inputs are too hard to recover from. Any game that requires REAL precise timing is going to suffer from the ghost downs. Bob, are you seeing the same thing on your stick?

 

Otherwise, it works excellent.

 

[PacManPlus]

Hey man:

 

Sorry I haven't been around too much this weekend (except for the quick in and out).

 

I actually do not see that issue - but keep in mind that 'down' (and right) is the direction you get when you have an 'out of range' value (or when the joystick is not plugged in). It sounds like you may still have a 'leak' in the left/right connections.

 

Use the 'PAM Diagnostics' (if you have either a 128 in 1 cart or if you can burn your own carts). Option 4 (I believe) tests the joystick inputs. If memory serveds me correctly, the values you get should be 102-106 for center, 1-3 for one direction, and 209-216 for the opposite (I forget which is which at the moment).

 

Hope this helps...

Bob

[Paranoid]

Sure does. And it is good to know that it ISN'T an issue I should expect. All of the resistances were a little low on the multimeter, so I wouldn't be surprised if a leak was the problem. I'll check it out with the PAM Diagnostics and see what I'm getting.

 

Thanks for the help, the whole way along. It really is a killer stick. I was thinking about it, and the basic design to a db 9 pin wired for joystick input would make an inexpensive DIY Redemption adapter, too, wouldn't it? Basically a 15 pin passthru for the keypad wired to pins 1 to 8, and a 15 pin passthru wired to pins 9 through 15, with the circuits on 9, 10 and 11... the other end goes to an Atari 15 pin cable... You could plug in your 7800 RSI stick and have basically the same thing...

 

I was kinda thinking this may be my next project. It actually seems like this would be easier than this project.

[Paranoid]

So... that is right within the values I am getting, and I didn't see any anomolous blips indicating that it was getting ghost down inputs.

 

So if the Pokey Adjust isn't showing it... how come it is happening in games, in the multi-cart menu?

[Paranoid]

Ok. So, here is a result that was odd.

 

In Montuzuma's revenge, moving left or up is instantaneous, but moving right or down isn't. Moving right will pause for a second, then start slowly (quick tiny erratic steps), then move at normal speed. Moving down stalls out, unless I tap the joystick repeatedly in the down direction, and even then, it is painfully slow.

 

Can you check out this title and see what your results are like?

 

On the test, I'm getting a little bit higher than the range you stated... sitting idle at "center" I was getting 107-108 horizontal, and 109-112 vertical. When I pressed up, it was 1-2, down was 213-214... left was 1-2 and right was 215-216.

 

Could I need to adjust the pot inside? I've only tried the stick on one of my consoles. I can give the other one a try, too... see what happens there.

[Paranoid]

Ok. So, my results on the second console. It works much better. No noticable issues with Montezuma's Revenge at all... BCs Quest still doesn't work well at all, although I still noticed some false DOWN inputs.

 

So... I guess it is a nature of the beast kinda thing with the analog to digital on old Atari 5200 consoles where my resistors were a couple of hundred ohms off the figure quoted, and my guess is that I *could* dial it in using the pot on the logic card and still have it work within the tolerances of a regular 5200 stick. I might have to look into that.

 

I didn't check the PAM diagnostics on this console, though. Maybe I'll go do that. I imagine it will look about the same, just closer to the figures you quoted.

[Paranoid]

I've been thinking about it, and if I did it again, I think I'd want to do the circuits resting in the battery case, but attached to something like a DB9 or even a CAT-5 cable, with the female side attached to the wires inside the joystick housing. That way you could easily open it up, disconnect the connector, work on it, and then plug it back in when you were done.

 

Any ideas what kind of connector would be ideal for this? You've got 8 wires coming off the two circuits, so DB-9 would work. Would a RJ-45 cat-5 cable work too? THAT would actually be kinda spiffy.

 

Ok... so I had some Cat 5 wall mount jacks left over from when I wired my house, and RJ-45 cat 5 is 8 wire... so, theoretically that should work, right? I mean, wire is wire... So, how do I figure out which wires go to which pins on the RJ-45 jack, short of testing each short with the multimeter? I'm guessing someone knows this info already, right?> The jack says 2135 on one side, and 4687 on the other. As I'm looking at an RJ-45 jack from the front, do I count pins from the left or from the right?

 

I think I'm going to do this. It'll make it so much easier to work on the resistor circuits *and* to open the stick itself if I need to. I want this project to be over so bad, and I wish I had thought of this BEFORE I had finished, but, it seems like it will be an easy fix to reverse engineer into the design.

Edited July 18, 2006 by Paranoid

[PacManPlus]

Hi:

 

I didn't have a chance to test MR (I have to go to work now), but I can tell you that I was having similar issues at the beginning (and funny enough I was saying the same thing to myself, "I just want to be *done* with this now"). Although I would be getting incorrect readings, no matter what I tried. I had actually given up once, but came back to it, and found the schematic that I had posted on page 1 of this thread. The only thing I can tell you, is that the current will always look for the path of least resistance, so check it over. You might want to set it up exactly like I have in my schematic, only because I know that works well (it's the one I am using).

 

But if you can adjust your values, that '109-112' vertical center you might want to bring down to about 105. That might be causing some of your issues. The slight variances in resistance might "once in a while" make it go to 113, which *could* be the reason you are seeing some 'down' input every so often. Not that 113 is some magic number, just an example that some applications might have more of a 'dead zone' than others.

 

Bob

Edited July 18, 2006 by PacManPlus

[Paranoid]

So, by EXACT schematic on page 1, for example, you mean the 2.2k and the 250k should be joined to DOWN/N.C. right at the connector on the switch, not a bit further up on a wire coming FROM the connector on the switch?

 

Otherwise, although I've drawn it a little differently (and evidently applied it BACKWARDS... oh, and upside down), I have used your schematic with no real innovation in the circuit design. I mean, I don't really understand what this circuit does, and couldn't make it myself, so, I'm pretty much just following directions, here.

 

I can actually look at your schematic now and see that if you start at 2.2k N.O. and go down to N.C., then back across the 250k to N.C., then back UP across the 497k to N.O., that really it is a single line of resistors in series, connected through the connectors on each switch.

 

Maybe if I go ahead and do this RJ-45 connector for the circuit, I'll go ahead and orient it correctly while I've got the circuits cut out to splice them into the connector.

 

One "last" question... I take it I want to INCREASE resistance or reduce current. I can either ADD some resistors to the circuit or dial DOWN the numbers showing up on the Pokey Adjust using the pot on the circuit board, right? Adding resistors insures that my regular 5200 joysticks will work properly (in as much as they are "working properly" now). By dialing down the voltage on the pokey, I might find that I fix the RSI joystick, but my regular analog sticks don't behave properly. Do I understand this much correctly?

 

The fact that it WORKS for MontRev on on one system and not on the others makes me think it is a combo of issues. I think I might have a slight amount too little resistance (I suspect my 497k circuit), and that the adjustment of the pokey current on my one console is a little high, and that thirdly, the games are designed without a real standard way to deal with the analog nature of the sticks and therefore have a range of values for dead-zone, etc. Which of course, makes it terribly difficult to troubleshoot the issue.

 

In fact, the joystick is very servicable, and is already a better choice for certain games than the stock sticks. Best of all, it was cheap. I'm just being a perfectionist. I want it to behave like an absolute digital stick, where an up input is NEVER going to result in a ghost down input. Which is kind of a tall order.

 

I think I WAS actually able to see the ghost input in the Pokey Adjust diag, too. Sometimes when I press up, so briefely you can barely see it flash on the screen, the OPPOSITE direction will jump to a range between 70-90 for just a fraction of a second. Because this happens FIRST, however briefely, you go down, instead of up. In the menu screen, usually when I press UP to scroll up the menu, it moves DOWN once first, before moving up. So it is consistent. I'm just not sure what is causing this surge in the opposite direction.

[+remowilliams]

One "last" question... I take it I want to INCREASE resistance or reduce current. I can either ADD some resistors to the circuit or dial DOWN the numbers showing up on the Pokey Adjust using the pot on the circuit board, right? Adding resistors insures that my regular 5200 joysticks will work properly (in as much as they are "working properly" now). By dialing down the voltage on the pokey, I might find that I fix the RSI joystick, but my regular analog sticks don't behave properly. Do I understand this much correctly?

I don't think you want to be playing with the POT adjust on the 5200 board itself. You should calibrate a 'good' known 5200 joystick first using the diags - and only then adjust the 5200 itself if needed. And yes if you mess with the 5200 to make your RSI work, your 5200 sticks will no longer work correctly.

 

Some games are much pickier than others about the deadzone values from movement I've noticed. BBSB was particularly picky if I recall.

[Paranoid]

Heh. The problem is, none of my joysticks are "known GOOD". They're all, "known tolerable" for 30 year old sticks.

 

Even my Best Electronics Gold Dot acts erratic... oddly enough, often with "Down". On some games, it will NOT go down. On others, it will. K-Razy Shootout is the one that comes readily to mind. Maybe I'll go test results with that.

 

So, K-razy Shoot Out played FLAWLESSLY, and was actually maybe a little more fun than Berzerk. I know I got further than I ever seem to get on Berzerk. The controls felt VERY crisp and responsive with the RSI stick, and I didn't notice *any* ghost movement at all.

 

Gah. I expected this game to be HORRIBLE. I'm going to have to try it on the other console, the one that wasn't working right with MontRev.

Edited July 18, 2006 by Paranoid

[Paranoid]

Ok... so I rewired it today, using a RJ-45 between the circuits and the stick. Now, I'm going to have to go confirm this... but I get Zero Resistance in one of the vert and Zero Resistance in one of the verts and "nothing" in the other directions (although it will input a single input in the other two directions).

 

Down works. Up doesn't. I think Right works, Left doesn't.

 

I've got a question about this circuit. Is it using TWO of each resistor? That is, is there a 2.2k between Up and Down, and a separate 2.2k between left and right? Or is there only a SINGLE 2.2k resistor shared between all four directional inputs? I'm using two different resistors, one for the horizontal circuit, one for the vertical circuit. So, each circuit is 6 resistors, for 12 resistors total.

 

Is that the problem?