controller calibration with Pete's Test Cartridge

[bradhig1]

I just got Pete's Test Cartridge today and I use a masterplay clone with a genesis controller on 5200.   On the joystick calibration screen the dot is not in the center of the blue square. It is a little to the left  Should it be in the center?

[+-^CrossBow^-]

No...it won't usually be in the center.

 

I don't usually use Pete's cart for calibration and use my flash cart with the original 1.1 diagnostic rom on it. That is because the values in the service manual for where everything should be adjusted is based on that diagnostic rom. But generally the Horz should be at 112 +/- 1 and the vert should read about 113 - 120 in most cases when everything is perfectly centered. But that is why you should use a loop back test board because the values of the controller inputs on those test boards are fixed to make the adjustment on the 5200 side more precise. Once that is done, then you also have to manually move the arms on the trimmers of the controllers themselves to calibrate them. 

 

But this is when using stock controllers with the 5200. Good games to test are Gyruss to make sure your ship will rotate all the way around. That game is very picky about the calibration being good on your system. Also Popeye will have issues moving up and down the stairs if the calibration isn't good.

 

The worst game about calibration has to be Bounty Bob Strikes Back. That game is crazy sensitive to the calibration being spot on. If it isn't you will have issues moving left and right. So I've adjusted plenty of systems where all games work great. Then put in bounty bob and bob will just walk to the right and won't respond to the left. That is when I know I have to do some more tweaking to get is working just right.

 

it can be a tedious process but so satisfying once everything is set right. But..keep in mind that doing all of this is essentially calibrating your system to your controllers. So other controllers you purchase might have to be calibrated separately afterwards to work properly. Although as long as you have the console set correctly per the loop back test boards, then you know all other adjustements just need to be done on the controller arms themselves and not the pokey adjust on the console.

 

[+DrVenkman]

2 hours ago, -^CrossBow^- said:

I don't usually use Pete's cart for calibration and use my flash cart with the original 1.1 diagnostic rom on it. That is because the values in the service manual for where everything should be adjusted is based on that diagnostic rom. 

Pete’s Test Cart shows the same values, not just the graphical display. It also lets you verify button functionality for both fire buttons, all twelve keypad buttons plus start, pause and reset. If you do a *lot* of controller calibrations like @MakerMatrix and I were doing last fall/winter, and what EdLaddin will likely be doing once Ed is ready to put his arcade units on sale, it’s definitely quicker and more direct than using the diagnostic rom. 

[bradhig1]

countermessures tends to be picky with calibration because sometimes the tank just moves down or right until I adjust the settings.

[+-^CrossBow^-]

1 hour ago, DrVenkman said:

Pete’s Test Cart shows the same values, not just the graphical display. It also lets you verify button functionality for both fire buttons, all twelve keypad buttons plus start, pause and reset. If you do a *lot* of controller calibrations like @MakerMatrix and I were doing last fall/winter, and what EdLaddin will likely be doing once Ed is ready to put his arcade units on sale, it’s definitely quicker and more direct than using the diagnostic rom. 

Not debating, but on the 5200s I've calibrated Pete's gives me different values from what the diagnostic 1.1 shows. As an example lets say I adjust a 5200 using the 1.1 diags to 112 - 116. On petes it will show different values with the loopback test board installed. In fact it is hard to get values because the loopback board will cause Pete's cart to keep changing screens since that is part of the function of the loopback controller board. The 1.1 diags will stay put on this screen and only does the auto mode if I select that.

 

If I were to use Pete's cart I would have to use a 5200 controller with the arms in the dead 9pm and 6pm locations to read what values I get then. While I can certaintly adjust the pokey trimmer to match it up with the controller, that doesn't mean the next controller will give me the same values and hence..it might not work. I know that you and Herb experienced alot of differences in that regard from system to system and controller to controller. Or at least that is what Herb has stated on FB more than once.

 

[+DrVenkman]

From what I gather, Pete's was put together by the folks doing repairs and adjustments in-house at Atari who wanted the fastest tool available to calibrate controller pots in the factory or repair center without the need for test harnesses. It does one thing and one thing only - calibrates controllers. 

 

What our experience was, building the gamepad controllers, if the POKEY adjust resistor was set to the scale recommended in the Field Service Manual and gave full-range motion on the graphical box display, it would play all 4-way games perfectly, and acceptably on some (but not all) analog control games due to the nature of the short-throw thumbstick and self-centering springs. This was repeatable across literally 100+ controllers. The same properly-adjusted POKEY would give similar values with modern thumbsticks or properly-maintained vintage Atari sticks. 

The issues we ran into building thumbstick controllers were three-fold: the current design that Scott put out requires a LOT of very close-pitch soldering for every single unit. That takes an awful lot of time and effort for each one. Second, the MCP42100 digital potentiometer chips have channel-to-channel variability and chip-to-chip variability. Fully 20% of the ones we purchased would give substantially high or (usually) low value outputs in one or more axes (usually the X-axis for reasons never determined). Pull the chip and replace it with one known to be good and the same controller would work great. Put that same out-of-spec chip in an otherwise good controller and it would be bad. Definitely issues with the chips there.  And the final issue we ran into was even if we had known-good controllers working and tested on one of our known-good working and tested 5200's, sometimes customers had wildy out of-spec POKEY adjustments and didn't want to disassemble their consoles to adjust them or in one case, the adjustment resistor didn't actually work - probably a dirty or worn out component on the console. The customer sent the controller back at our expense and it works fine for us today. The joys of vintage systems.  

 

So in reality, the kicker for dealing with controller calibrations is that you need at least one working controller to first verify that your POKEY adjust is correct because that's the system you're using to then calibrate additional controllers.  So if you have both an uncalibrated POKEY to begin with, PLUS a misbehaving vintage stick, you don't have a known-good reference to work from to make the initial POKEY adjust, and then determining if the vintage stick merely needs pots adjusted or if they're worn out. 

 

Tl;dr: Pete's is designed to work stand-alone without need for a test harness and it works great *IF* you have a known-good POKEY setting in the system to begin with. 

[edladdin]

Here's a discussion that lands right in the middle of my current efforts!

 

I can certainly verify that every 5200 will yield different results!  I have three consoles to use for testing and my prototype "Super 52" controller will deliver three different results for the "at rest" center position on the Hall-effect self-centering analog joystick we're using.  In the end, we're either going to add a pair of on-board trimmer pots, or an auto-calibrate function in which you will hold the cursor at the center of the screen while doing a long press on a function key on our keypad.  The problem with the last idea is that most people are not going to have a copy of Pete's to get the centering right, and not that many games lend themselves to that either.  Centipede and Missile Command?

 

Very grateful to @DrVenkman and @MakerMatrix for clarifying for me that the values served up by Pete's cart are on some arbitrary scale, rather than anything concrete like kHz. 

 

Trivia:  Pete's Test Cart came up in one of the FB forums recently and Dan Kramer talked about the actual guy, who apparently was the sole developer of this bundle of tests.

[edladdin]

On 4/15/2020 at 12:05 PM, bradhig1 said:

I just got Pete's Test Cartridge today and I use a masterplay clone with a genesis controller on 5200.   On the joystick calibration screen the dot is not in the center of the blue square. It is a little to the left  Should it be in the center?

Referring to the original question... 

 

Doesn't the MP Clone have a pair of trimmer pots (knobs) so you can make those adjustments back to center? 

I have an original MP and it does not have any adjustment feature, which always has led me to wonder how they got away with leaving that off given the console variability.

[+DrVenkman]

18 minutes ago, edladdin said:

Doesn't the MP Clone have a pair of trimmer pots (knobs) so you can make those adjustments back to center? 

I have an original MP and it does not have any adjustment feature, which always has led me to wonder how they got away with leaving that off given the console variability.

There are different versions of the concept that are all basically clones of the Masterplay idea, though different designers over the years have come up with different approaches.  @smbaker’s “Yet Another Masterplay Clone” design is the one I have built and have the most experience with. His design indeed has two trimpots, one for each axis. 

 

Apologies for not having time right now to find the exact thread and quote it, but my own experience shows that sometimes even a pretty wide-range trimpot is *not* enough to get adequate response if your system has a badly out-of-spec POKEY Adjust resistor. My two Masterplay clones would simply not work whatsoever with my then-daily driver 4-port until I finally opened up the console and adjusted the POKEY. Since that tweak, both my examples have worked perfectly. 

[bradhig1]

mine has trimmer pots and in countermessaures I have to adjust them several times during gameplay.

[Danjovic]

The Time Precise method should render immunity against internal potentiometer setup.

[bradhig1]

I wouldn't know how to adjust the pokey. I don't know what cart to use or what values to look for.

[+DrVenkman]

1 hour ago, bradhig1 said:

I wouldn't know how to adjust the pokey. I don't know what cart to use or what values to look for.

The Field Service Manual describes it. There's a small variable resistor, accessible inside the case but outside of the RF shielding. You load up the 5200 PAM SALT diagnostic cart or, more conveniently, Pete's Test Cart, and adjust it such that - if your controller is top-dead-center, the X-axis and Y-axis values fall into the range described above. If you can only get one of the two axes into the proper range, either something more serious is wrong with your 5200 or more likely, the controller you're trying to use is itself out of calibration or has a worn-out potentiometer inside.

[Danjovic]

Pokey measures time which in turn is proportional to R, C and V being

• R the resistance of the potentiometer
• C the capacitance of an internal capacitor
• V two parameters here: Vac voltage, adjustable internally, and the Vih the positive threshold of Pokey chip pot input.

Time is measured by counts in the range of 0 to 227, and a known good controller with shaft at center position shall provide a reading at half of the maximum which means a value about 113.

 

Again with a know good joystick, if the values are below the expected it is possible to bring them up - non-intrusively - by either

• Adding a series resistor with the potentiometer axis
• Adding a capacitor from the axis to ground
• Adding diodes in series with Vac to lower the value - it will affect both axes

 

On the other hand, if the values (for a known good joystick) are above expected the non-intrusive solution space is limited to bring up Vac by either:

• Use Vcc (pin12) instead of Vac (pin 9) - switched by transistors or analog switch so games won't take a joystick by a trackball
• Use a DC-DC converter on Vac pin - Take care to not exceed the 6.4Volts that is equivalent to the maximum value provided internally

 

For Masterplay Clones the solution is simpler, as the resistor values can be reduced or increased to to bring down or push up the counting values.

 

 

 

 

 

 

[+DrVenkman]

32 minutes ago, Danjovic said:

Again with a know good joystick, if the values are below the expected it is possible to bring them up - non-intrusively - by either

This is exactly what the POKEY Resistor inside the console does ... it adds or removes resistance. The only “intrusive” thing about adjusting it is removing a handful of screws to take off the top case, then tweaking the knob.  

[Danjovic]

14 minutes ago, DrVenkman said:

....The only “intrusive” thing....

Sorry, I  forgot to put quotes in "intrusive", but that was exactly what I meant: "without opening the console".

But frankly speaking, if someone is able enough to implement my suggestions either modifying the controller or building an external adapter, that someone will surely be able to open the console to adjust the pot

[SpotAnime]

I just wanted to bump this because ITC recently posted a video on controller calibration using a loopback board and an alternative method for those who do not have a loopback board.

 

 

QQ - are these loopback boards available today? Or are they old Atari service equipment that have since gone scarce?

[+-^CrossBow^-]

12 minutes ago, SpotAnime said:

I just wanted to bump this because ITC recently posted a video on controller calibration using a loopback board and an alternative method for those who do not have a loopback board.

 

 

QQ - are these loopback boards available today? Or are they old Atari service equipment that have since gone scarce?

I purchased the one you see in that video from Best several years ago. I also tried to get a 2 port version but he was sold out of them. I believe I heard from someone over the summer that he was sold out of all of them now? I say this because I actually scanned my 4 port loopback board on both sides and posted the pics along with the resistor values used so that they could in theory be replicated. The hardest part would be finding panel mount female 15pin connectors I would think. but you could just as easily solder on controller cables in their place as long as you kept them fairly short to prevent adding in any additional resistance to the readings.

 

Also, I didn't state this in the video, but technically a loopback for making the pokey adjustment is only needed to get the readings from port 1. Because all the ports will be affected equally by the adjustment. 

 

 

[nick3092]

On 3/16/2021 at 5:06 PM, -^CrossBow^- said:

The hardest part would be finding panel mount female 15pin connectors I would think. but you could just as easily solder on controller cables in their place as long as you kept them fairly short to prevent adding in any additional resistance to the readings.

 

 

Wouldn't these work?

 

https://www.digikey.com/en/products/detail/norcomp-inc/182-015-213R531/858357

 

If someone were willing to do a run of 4 and 2 port boards, I'd certainly be in for one of each. 

[edladdin]

1 hour ago, nick3092 said:

Wouldn't these work?

 

https://www.digikey.com/en/products/detail/norcomp-inc/182-015-213R531/858357

 

If someone were willing to do a run of 4 and 2 port boards, I'd certainly be in for one of each. 

 

PC cables do not have quite as long a throat on the molding, so you can have issues with the plug not reaching deeply enough into the socket on the console.

 

Easiest source for "game style" molded DB15 cords is to purchase Neo Geo extension cords and lop off whatever you don't need.  Here's one example, though they are widely available from $10-15 each.

 

https://www.amazon.com/Tomee-Controller-Extension-not-machine-specific/dp/B07L6FC5R3