My homemade 5200 controller

[the Goat]

10 hours ago, DrVenkman said:

The short throw of the Sparkfun thumbstick is honestly very much like a D-pad already in 4-way games.

 

If you want to use a D-pad, the professionally-manufactured D-pad on the Genesis 3-button controller is way better than almost anything you'd be able to rig up yourself, and you can use one with one of Scott's Masterplay clone controllers, complete with 2-fire buttons.

 

 

The point of a d-pad is you can easily press left or right without accidentally pressing up or down.  Or vice versa.  The analog stick on these controllers is very difficult to do that with.

[+DrVenkman]

37 minutes ago, the Goat said:

The point of a d-pad is you can easily press left or right without accidentally pressing up or down.  Or vice versa.  The analog stick on these controllers is very difficult to do that with.

I’ve built over 60 of these to sell for others and use them routinely. Your opinion differs greatly from mine.

 

That said, as I also pointed out, if you prefer a D-pad, use a Genesis 3-button controller or compatible and plug it into an inexpensive Masterplay clone. I helpfully linked resources for you to build one above. 

 

You’re welcome. 

[the Goat]

4 minutes ago, DrVenkman said:

I’ve built over 60 of these to sell for others and use them routinely. Your opinion differs greatly from mine.

I built two of these controllers.  They are outstanding.  But they could be better.

Quote

That said, as I also pointed out, if you prefer a D-pad, use a Genesis 3-button controller or compatible and plug it into an inexpensive Masterplay clone. I helpfully linked resources for you to build one above.

I want a single controller with both analog and d-pad controls.

[+DrVenkman]

24 minutes ago, the Goat said:

I want a single controller with both analog and d-pad controls.

You’re going to have to design it. Scott’s PCB does not have room for both and the circuit is analog in nature, not digital (D-pad) converted to analog needed by the 5200. 

[the Goat]

29 minutes ago, DrVenkman said:

the circuit is analog in nature, not digital (D-pad) converted to analog needed by the 5200. 

No actually the controller uses digitally controlled potentiometers.  So it would be trivial to have the controller output a fixed value when pressing a direction on the d-pad.

[+DrVenkman]

15 minutes ago, the Goat said:

No actually the controller uses digitally controlled potentiometers.  So it would be trivial to have the controller output a fixed value when pressing a direction on the d-pad.

No, actually it uses an analog thumbstick dual-axis potentiometer to start the signal chain, so really it’s an analog-digital-analog circuit. This is all well-documented in Scott’s blog post and throughout the thread. 

 

And once more, this particular PCB does not support a D-pad. Just use a Masterplay clone or design your own hybrid combination.

[the Goat]

3 minutes ago, DrVenkman said:

No, actually it uses an analog thumbstick dual-axis potentiometer to start the signal chain, so really it’s an analog-digital-analog circuit. This is all well-documented in Scott’s blog post and throughout the thread. 

 

And once more, this particular PCB does not support a D-pad. Just use a Masterplay clone or design your own hybrid combination.

I'm not sure why you are being so obtuse.  Obviously the PCB would need to be redesigned to accommodate the d-pad.  Did you think I expected to just glue on a d-pad and it would magically work?

[+DrVenkman]

Just now, the Goat said:

I'm not sure why you are being so obtuse.  Obviously the PCB would need to be redesigned to accommodate the d-pad.  Did you think I expected to just glue on a d-pad and it would magically work?

Then why are you posting in this thread instead of starting a new one to propose your design? 

[the Goat]

3 minutes ago, DrVenkman said:

Then why are you posting in this thread instead of starting a new one to propose your design? 

Because I was proposing a new feature being added to this design.

Edited July 22, 2020 by the Goat

[MakerMatrix]

16 hours ago, the Goat said:

Because I was proposing a new feature being added to this design.

The microcontroller is already out of pins, effectively (there is the reset pin but there are a lot of gymnastics to go through in order to use it for data and still be able to program the chip and it would still only give you one direction).  So you'd need a bigger MC to wedge another four data signals into it and then update the FW so it can read the pad and tell the digital POTs what to do based on that.   And then decide how to code up which input wins between the analog stick and the d-pad in case both are activated at the same time.   Which brings up a less technical but maybe more important consideration which is that there are already 17 buttons and a 4-way stick on this thing.  Adding a four-way pad might put it over the amount of real estate that can be organized in an ergonomic way.

[lazzeri]

On 7/21/2020 at 2:57 PM, MakerMatrix said:

I have enough PCBs and parts on hand to build about 25 of these.  I've had enough requests trickling in over the last few months that it's probably time to go ahead and get started on a short run of these.  And if you or a friend are handy with a soldering iron, you can always get the board from Scott and order the BOM off Digikey, et. al to DIY one.

   Would you please let us know when they’re available for purchase?

 

  Those are awesome. ? 

[Lan Di]

I just got done assembling one and the analog stick will not work. I have in on the left side. The characters on screen (pac-man, berserk, etc.) automatically move to the left. The buttons work fine but the analog stick will not work. I tried rearranging the jumpers and it does not help. Maybe a bad analog stick?

 

[+DrVenkman]

16 hours ago, Lan Di said:

I just got done assembling one and the analog stick will not work. I have in on the left side. The characters on screen (pac-man, berserk, etc.) automatically move to the left. The buttons work fine but the analog stick will not work. I tried rearranging the jumpers and it does not help. Maybe a bad analog stick?

 

I've built over 100 of these and never, ever run into a bad stick. Do you have a copy of Pete's Test Cartridge or the PAM SALT (diagnostic) cartridge? There are three possible problems that cause what you're seeing.

 

1) Marginal or bad solder joints on the (very!) small pads for the 0.1 uF capacitors that serve to effectively "amplify" the apparent range of the modern stick to meet what the 5200 POKEY circuit expects. If those caps aren't properly soldered, the stick won't register properly.

 

2) The POKEY Adjust resistor on your 5200 has drifted out of spec over time - this is more common than most people realize with machines pushing 38 years old. You do have to open the console to adjust it, but it's a very minor thing, much like adjusting the color potentiometer. You really shouldn't adjust it though without a way to measure the effects of your change, so if you have Pete's Test Cart or the PAM SALT cart, you should use them when making any adjustment, otherwise you're flying blind.

 

3) The digital pot chip might be out of spec - we ran into a surprisingly high percentage of these that would read low on one or both axis measurements and had to be replaced. 

 

 

 

 

[Lan Di]

2 hours ago, DrVenkman said:

I've built over 100 of these and never, ever run into a bad stick. Do you have a copy of Pete's Test Cartridge or the PAM SALT (diagnostic) cartridge? There are three possible problems that cause what you're seeing.

 

1) Marginal or bad solder joints on the (very!) small pads for the 0.1 uF capacitors that serve to effectively "amplify" the apparent range of the modern stick to meet what the 5200 POKEY circuit expects. If those caps aren't properly soldered, the stick won't register properly.

 

2) The POKEY Adjust resistor on your 5200 has drifted out of spec over time - this is more common than most people realize with machines pushing 38 years old. You do have to open the console to adjust it, but it's a very minor thing, much like adjusting the color potentiometer. You really shouldn't adjust it though without a way to measure the effects of your change, so if you have Pete's Test Cart or the PAM SALT cart, you should use them when making any adjustment, otherwise you're flying blind.

 

3) The digital pot chip might be out of spec - we ran into a surprisingly high percentage of these that would read low on one or both axis measurements and had to be replaced. 

 

 

 

 

I think I might I have issue #3. My standard 5200 controllers and my Wico Command joysticks work just fine. The solder points on the homemade 5200 controller PCB are solid and not loose. I just ordered another chip. They are actually very inexpensive so it should not be an issue. I'll reply with the results as soon as I receive and install the other chip.

[CZroe]

I've built over 100 of these and never, ever run into a bad stick. Do you have a copy of Pete's Test Cartridge or the PAM SALT (diagnostic) cartridge? There are three possible problems that cause what you're seeing.
 
1) Marginal or bad solder joints on the (very!) small pads for the 0.1 uF capacitors that serve to effectively "amplify" the apparent range of the modern stick to meet what the 5200 POKEY circuit expects. If those caps aren't properly soldered, the stick won't register properly.
 
2) The POKEY Adjust resistor on your 5200 has drifted out of spec over time - this is more common than most people realize with machines pushing 38 years old. You do have to open the console to adjust it, but it's a very minor thing, much like adjusting the color potentiometer. You really shouldn't adjust it though without a way to measure the effects of your change, so if you have Pete's Test Cart or the PAM SALT cart, you should use them when making any adjustment, otherwise you're flying blind.
 
3) The digital pot chip might be out of spec - we ran into a surprisingly high percentage of these that would read low on one or both axis measurements and had to be replaced. 
 
 
 
 

Regarding #2, I've found that anyone following Best Electonics' guide for calibration with a controller that has US-made CTS pots will have a miscalibrated console since CTS pots are only around 350-390k ohms these days (instead of the specified 500k ohms). Because there seems to be little acknowledgement of this before I analyzed several NOS CX-52 sticks in 2017, many were blindly following those guides and miscalibrating their consoles if they had one of those sticks. Because I have not done a good job spreading this information, many may have continued following older guides and miscalibrating their consoles.

It's fine if all your controllers will have CTS pots but not fine when you need the console to support others.

Panasonic and Alps pots are fine.

[+-^CrossBow^-]

On 10/29/2020 at 6:36 PM, CZroe said:

Regarding #2, I've found that anyone following Best Electonics' guide for calibration with a controller that has US-made CTS pots will have a miscalibrated console since CTS pots are only around 350-390k ohms these days (instead of the specified 500k ohms). Because there seems to be little acknowledgement of this before I analyzed several NOS CX-52 sticks in 2017, many were blindly following those guides and miscalibrating their consoles if they had one of those sticks. Because I have not done a good job spreading this information, many may have continued following older guides and miscalibrating their consoles.

It's fine if all your controllers will have CTS pots but not fine when you need the console to support others.

Panasonic and Alps pots are fine.

I'e always followed the calibration routine and numbers from the actual service manual. Although I have one of the loopback boards that provides static and consistent readings for making the adjustment in addition to the diagnostic rom. That seems to work best. I then have 3 games I use for testing. Bounty Bob Strikes back, Gyruss, and The Last Starfighter. All three of those games are especially sensitive to the adjustment being proper in addition to your controllers.

 

[CZroe]

I'e always followed the calibration routine and numbers from the actual service manual. Although I have one of the loopback boards that provides static and consistent readings for making the adjustment in addition to the diagnostic rom. That seems to work best. I then have 3 games I use for testing. Bounty Bob Strikes back, Gyruss, and The Last Starfighter. All three of those games are especially sensitive to the adjustment being proper in addition to your controllers.
 

Exactly, but when the controller you are using has bad pots (low resistance) then you will be miscalibrating the console regardless.

In order to act like their mylar PCB and button membrane products solve everything short of physical damage, Best Electronics was basically telling everyone with the problem (bad "Made in USA" CTS pots with low resistance) that the issue was their console and then walked them through adjusting the console. If the console is correct (it usually is), the real fix is to replace the pots or use a capacitor in-line with them to increase the RC delay. If the console has been misadjusted to suit one of these faulty controllers then you need to readjust the console with a known-good controller and then fix the bad controller with a cap for each pot.

The issue causes you to be unable to reach the bottom or right sides of the screen in analog games like Missle Command which usually prevents you from going down or right in digital games.

You can't just put a resistor in-line with the pots or else you will be raising the minimum resistance and causing the opposite issue, limiting to top and left-most ranges. Adjusting the console seems to work until you go back to an in-spec controller with Made in Japan Alps or Panasonic pots.

[Lan Di]

OK, I've replaced the digital pot chip and the problem was not resolved.  Like I stated earlier, all my other controllers work fine, but not the 5200 home made. So I don't think the internal 5200 pot is out of spec. Unless I need to recalibrate the internal POT for each controller. 

[Lan Di]

I've added photos. 

[+DrVenkman]

On 11/5/2020 at 6:50 PM, Lan Di said:

Unless I need to recalibrate the internal POT for each controller.

No, if the POKEY Resistor is in-spec and the other controllers work properly, then the issue is going to be with the build. It's possible you also might have a bad ATtiny chip, but you should see an error when you program it, if that's the case. You could always simply pull the chip and try reprogramming it.

 

If that doesn't work, you're going to be down to tracing continuity between the various component pins on the board and trying to track down any bad solder joints. 

[Lan Di]

On 11/12/2020 at 4:13 PM, DrVenkman said:

No, if the POKEY Resistor is in-spec and the other controllers work properly, then the issue is going to be with the build. It's possible you also might have a bad ATtiny chip, but you should see an error when you program it, if that's the case. You could always simply pull the chip and try reprogramming it.

 

If that doesn't work, you're going to be down to tracing continuity between the various component pins on the board and trying to track down any bad solder joints. 

After a ton of troubleshooting and a small amount of money spent, my DIY 5200 controller is working great! I ensured that the solder joints were not loose. I used my multimeter to test for continuity and capacitor voltage. I even opened my 5200 and adjusted the pot, and I even reprogrammed the controller's ATTINY 85 chip. And still nothing changed. So I decided to use six wires and wire the analog stick to the right side of the PCB. I soldered the left side to the six solder pads used for the right side analog and BAM,  now works flawlessly. 

[8bitAndy]

I have a question about the pots which is related to this controller design. Forgive my ignorance of electronics and I would appreciate any corrections where I misunderstand something.

 

According to Scott Baker's blog, the digital pot maxes out at 100K resistance instead of the required 500K resistance the Atari expects. This is compensated by adding a 220nF capacitor in the controller.

 

Another way to look at it is that the digital pot has 1/5th the resistance, so will charge the internal capacitor in the 5200 five times faster. An additional 220nF capacitance must make the total capacitance five times bigger. If the internal capacitance is 55nF, then an additional 220nF will be 275nF, or five times the amount. 

 

One of my controllers has an out-of-spec pot that has 381K resistance. This will charge the capacitor 1.3 times faster than an in-spec pot. To compensate, I want 55nF * 1.3 total capacitance, or about 16nF additional capacitance.

 

Should it be possible for me to fix my out-of-spec pot by adding a 16nF capacitor, or am I misunderstanding something?

 

Edited November 27, 2020 by 8bitAndy

[CZroe]

I have a question about the pots which is related to this controller design. Forgive my ignorance of electronics and I would appreciate any corrections where I misunderstand something.   According to Scott Baker's blog, the digital pot maxes out at 100K resistance instead of the required 500K resistance the Atari expects. This is compensated by adding a 220nF capacitor in the controller.  

Another way to look at it is that the digital pot has 1/5th the resistance, so will charge the internal capacitor in the 5200 five times faster. An additional 220nF capacitance must make the total capacitance five times bigger. If the internal capacitance is 55nF, then an additional 220nF will be 275nF, or five times the amount. 

 

One of my controllers has an out-of-spec pot that has 381K resistance. This will charge the capacitor 1.3 times faster than an in-spec pot. To compensate, I want 55nF * 1.3 total capacitance, or about 16nF additional capacitance.

 

Should it be possible for me to fix my out-of-spec pot by adding a 16nF capacitor, or am I misunderstanding something?

 

I have already done exactly this except I recall the nF required being around 20nF. It worked like a charm! I have a NOS crate of four controllers and every one measured 380k-390k-ohms, which made them unusable for most games. For example, most would not register down or right. Games like Missile Command where you move a cursor across a range cannot reach the far right or bottom edge of the screen.

 

I bought this assortment:

https://www.amazon.com/dp/B06XSGNTMM/ref=cm_sw_r_cp_apa_fabc_y6uWFbXNZQ1TN?_encoding=UTF8&psc=1

...though it doesn't seem to be available anymore.

 

They were all "Made in USA" pots marked/branded "CTS." Japanese pots from MAtsushita/Panasonic and Alps all seem fine where these CTS pots seem to have degraded universally.

 

Here is a comparison of Matsushita/Panasonic, CTS, and Alps potentiometers (variable resistors/pots) that I made in 2017 with examples from different revision CX52 Atari 5200 controllers:

https://imgur.com/a/gDb9j

[BitMerge6502]

On 11/3/2018 at 1:15 PM, amiman99 said:

Here is a preview of the 3d Case:

What's the licensing on the STLs? Would I be allowed to post a "remix" on Thingiverse (with all due credit & license given to original author)? 

[amiman99]

13 hours ago, BitMerge6502 said:

What's the licensing on the STLs? Would I be allowed to post a "remix" on Thingiverse (with all due credit & license given to original author)? 

I don't have this 3D model on Thingiverse. Unless someone posted it there.

When you create "remix" of this, you can post it here.

 

Personally I would add a clip for the board, so it's easier to assemble, and additional screws for the handle. Also an edge lip around the shell.