MiSTer FPGA - 7800 Core

[Rodney Hester]

By the way, I did find this:

 

https://www.reddit.com/r/MiSTerFPGA/comments/x6k19h/current_atari_2600_paddles_solution/

 

from which I cribbed their input config file, and it seems to work...mostly.  I've made the following observations with the currently-released 7800 core and the MiSTer D9 test firmware from January 12:

 

- Only one of the two paddle buttons work, even with "Allow Multi Paddles" enabled

- There appears to be range limiting - for example, in Warlords, I can't get the Yellow player far enough to the left edge of the screen to hit the Green player's leftmost bricks

- There is some pretty intense jitter going on.  I've tried two sets of legit Atari paddles with new(ish) G5 pots from Best Electronics and one set of Flashback paddles and all three exhibited the same high level of jitter, so it's not the controllers

[BrianC]

Should I keep the January 28th firmware or replace it with the Mr firmware from the 26th? I looked for more information on the Mr discord, but didn't find anything.

Edited February 1, 2023 by BrianC

[Rodney Hester]

13 hours ago, Rodney Hester said:

- There is some pretty intense jitter going on.  I've tried two sets of legit Atari paddles with new(ish) G5 pots from Best Electronics and one set of Flashback paddles and all three exhibited the same high level of jitter, so it's not the controllers

I can't seem to edit my post...but I'm wondering whether this could be related to having fast USB polling enabled.  The earlier video of someone playing Kaboom doesn't seem to exhibit this problem - mine is shaking like a Chihuahua in a foot of snow.

[BrianC]

Are you using the latest firmware from this link or the Mr. Firmware from a few posts before in this topic? I'm not getting jitter with the latest firmware. They actually jitter less on MiSTer than on actual hardware. This is with fast USB polling on. The reddit link you posted is a few months old and before the firmware that was posted in this topic. 

 

2600-daptor D9 - 9-pin controller to USB interface

Edited February 1, 2023 by BrianC

[dualcam]

21 hours ago, BrianC said:

I noticed a new January 28 firmware on the 2600-daptor site. Does this have the Mr stuff applied?  

Just posted a 2600-D9 special firmware version for MiSTer -

http://2600-daptor.com/2600-firmware D9.htm

 

This is just for the paddle scaling in the 7800/2600 core.  Using the Atari 2600 dedicated mode (all switches down) will give the paddle scaling for MiSTer.  Selecting paddles with AutoSense (all switches up) with give the old scaling for Stella.

 

Am holding off on anything the driving controller - have learned there are pending changes to the core for this.

 

Tom

http://2600-daptor.com/

[BrianC]

Thank you!

[Rodney Hester]

I've successfully tested the new firmware in Windows, but since it shows up as a different device, I don't know the proper way to do the initial MiSTer "Define joystick buttons" setup.  Can anyone describe what actions to take for each of those input questions?  (Obviously the cribbed config I borrowed from Reddit isn't going to work anymore LOL)

[Rodney Hester]

Progress continues...

 

I've more or less figured out how to map the buttons to B and A and the axes of the paddles within MiSTer Main (I'm not going to explain how yet because it's a PITA and somewhat broken).

 

That being said:

 

- I'm *still* getting wonky jitter even on yesterday's D9 firmware.  Maybe I really do have 3 sets of marginal paddles, but that seems really hard to believe.  LOL

- The button mapping (in the core!) is still an unholy mess.  Why does there have to be a separate paddle button defined (seriously - I see *no point* in this existing at all!), versus just using Fire1?  If Fire1 is mapped to paddle 1's button (in multi-paddle mode) and Fire2 is mapped to paddle 2's button, both joystick and paddles (single or multi) will work perfectly with the D9 without any need of further (re)configuration.  As-is, every time you swap controllers you have to redo the 7800 button definitions, which is a totally unnecessary PITA.

 

It's that second paddle button that makes the MiSTer Main mapping a royal PITA as well, but until point 2 above is resolved, it won't make much difference.

 

Not in ANY way trying to sound down on the core author or the D9 maker...you are both brilliant and awesome people, I'm just expressing a slight bit of frustration at being sooooooooooo close to Atari nirvana 40 years on.  LOL  Hopefully the feedback helps in some way.

 

SO...CLOSE!  😃

[dualcam]

21 hours ago, Rodney Hester said:

Progress continues...

 

I've more or less figured out how to map the buttons to B and A and the axes of the paddles within MiSTer Main (I'm not going to explain how yet because it's a PITA and somewhat broken).

 

That being said:

 

- I'm *still* getting wonky jitter even on yesterday's D9 firmware.  Maybe I really do have 3 sets of marginal paddles, but that seems really hard to believe.  LOL

- The button mapping (in the core!) is still an unholy mess.  Why does there have to be a separate paddle button defined (seriously - I see *no point* in this existing at all!), versus just using Fire1?  If Fire1 is mapped to paddle 1's button (in multi-paddle mode) and Fire2 is mapped to paddle 2's button, both joystick and paddles (single or multi) will work perfectly with the D9 without any need of further (re)configuration.  As-is, every time you swap controllers you have to redo the 7800 button definitions, which is a totally unnecessary PITA.

 

It's that second paddle button that makes the MiSTer Main mapping a royal PITA as well, but until point 2 above is resolved, it won't make much difference.

 

Not in ANY way trying to sound down on the core author or the D9 maker...you are both brilliant and awesome people, I'm just expressing a slight bit of frustration at being sooooooooooo close to Atari nirvana 40 years on.  LOL  Hopefully the feedback helps in some way.

 

SO...CLOSE!  😃

Never had any negative feedback about jitter in MiSTer.  If you run into a new problem or something not being addressed, best course is to contact me directly (email or msg here). 

 

Note I don't actually don't have MiSTer, but in watching the raw paddle output, I can see a touch of jitter, I think a result of the new scaling.

 

Here is an update -

http://2600-daptor.com/2600-daptor D9 Mr 230203.hex

 

Let me know the results.

 

Tom

http://2600-daptor.com/

 

[dualcam]

Posted a 230204 firmware update -

http://2600-daptor.com/2600-firmware D9.htm

 

This is slightly different than the 230203 in my prior post.

 

Not seeing any jitter - if you do see some, please post a video.  I wonder if people may have forgotten that in Kaboom!, the buckets will start rhythmically swaying back and forth a bit after you press the paddle button and are playing the game.  This happens on the real console - this is the way the game behaves.

 

EDIT: And looks like this rhythmic swaying is just when the player on the left 1/4 or 1/3 of the screen.

 

Tom

http://2600-daptor.com/

Edited February 4, 2023 by dualcam

[Jstick]

I might be able to help with the jitter; I ran into this issue myself with a pair of upgraded Best paddles that work flawlessly on original hardware.

 

In my case, I found that by touching the metal D9 port on the adaptor, the jitter disappeared entirely.  
 

I can only assume this is a grounding issue of some sort, but haven’t diagnosed this any further yet.

[xeex]

3 hours ago, Jstick said:

I might be able to help with the jitter; I ran into this issue myself with a pair of upgraded Best paddles that work flawlessly on original hardware.

 

In my case, I found that by touching the metal D9 port on the adaptor, the jitter disappeared entirely.  
 

I can only assume this is a grounding issue of some sort, but haven’t diagnosed this any further yet.

Wow, you're right. I've been struggling with the same problem. I tore apart the pots in my paddles and cleaned everything thoroughly - still had jitter with every firmware, including the one released today. I've been considering new pots from Best. Saw your post, touched the shield around the D9's port, and the jitter disappeared. Curious.

[Jstick]

On 2/3/2023 at 4:46 AM, Rodney Hester said:

- The button mapping (in the core!) is still an unholy mess.  Why does there have to be a separate paddle button defined (seriously - I see *no point* in this existing at all!), versus just using Fire1?  If Fire1 is mapped to paddle 1's button (in multi-paddle mode) and Fire2 is mapped to paddle 2's button, both joystick and paddles (single or multi) will work perfectly with the D9 without any need of further (re)configuration.  As-is, every time you swap controllers you have to redo the 7800 button definitions, which is a totally unnecessary PITA.

I can definitely help with this one:

 

For the paddles to show as 2 controllers, a workaround was implemented in the MiSTer core (the 2600daptor d9 presents paddles as multiple axes on a single controller, MiSTer requires it to be 2 separate controllers).  The paddle button mapping is part of that workaround.

There is a way to have separate mappings for Paddles and non-paddle games, by using an .mra file.  MRA files allow you to have different configurations for the same core.  So I have both "Atari 2600 (Paddle)" and "Atari 2600" menu entries, and choose one or the other depending on which type of game I'm playing.  Using this method you will only have to do the button mappings once, after which they will never interfere with each other.

[dualcam]

13 hours ago, Jstick said:

I might be able to help with the jitter; I ran into this issue myself with a pair of upgraded Best paddles that work flawlessly on original hardware.

 

In my case, I found that by touching the metal D9 port on the adaptor, the jitter disappeared entirely.  
 

I can only assume this is a grounding issue of some sort, but haven’t diagnosed this any further yet.

Thanks, that is puzzling.  The metal ring around the D9 is supposed to be connected the USB ground.  You could open the adapter and double check that I didn't forget to solder it (see pic).

 

Have you tried playing on a regular PC/laptop?  Do you see this jitter there?  If not, then this grounding problem might be an issue with the MiSTer hardware.

[xeex]

On 2/5/2023 at 10:15 AM, dualcam said:

Have you tried playing on a regular PC/laptop?  Do you see this jitter there?  If not, then this grounding problem might be an issue with the MiSTer hardware.

I just tried in Stella and see the same behavior. I had been passing this off as a problem with my paddles, but now I think something else is wrong. The soldering on my adapter looks OK, though.

[Jstick]

On 2/6/2023 at 1:15 AM, dualcam said:

Thanks, that is puzzling.  The metal ring around the D9 is supposed to be connected the USB ground.  You could open the adapter and double check that I didn't forget to solder it (see pic).

I opened my D9 adaptor up as well, and the 2 posts indicated with arrows in the picture are soldered.

[xeex]

I haven't tried my paddles on original hardware yet, so I can't rule out that they're still defective in some way. That's next on my to-do list. I do appreciate the time you're taking to make MiSTer users happy.

[dualcam]

18 hours ago, xeex said:

I just tried in Stella and see the same behavior. I had been passing this off as a problem with my paddles, but now I think something else is wrong. The soldering on my adapter looks OK, though.

Is this with the same USB cable? 

 

Tom

http://2600-daptor.com/

[xeex]

12 hours ago, dualcam said:

Is this with the same USB cable?

Yes, it was the same cable. I tried two different cables, actually, on MiSTer and in Stella, but didn't notice a difference between them. In all cases, touching the metal shield around the DE-9 connector eliminated jitter. I'll try to cobble together a video demonstrating what happens in Stella, but that might take me a while. And I still have to test these paddles on original equipment.

[Jstick]

12 hours ago, dualcam said:

Is this with the same USB cable? 

 

Tom

http://2600-daptor.com/

Ok, I had a chance to test this more thoroughly (2 different D9 adaptors on latest MiSTer firmware, 2 USB cables, MiSTer 7800 core and Stella on PC).  Here is what I've found:

 

If the dips are set to Autosense, paddles work properly in Stella (tiny bit of infrequent jitter).  In MiSTer, there is some jitter at the low end (paddle left) which reduces about half way across the screen and disappears when reaching the right of the screen.

 

If the dips are set to MiSTer mode, there is a large amount of consistent jitter in both Stella and MiSTer.  It appears that the scaling may be amplifying the amount of inherent jitter.

 

In all cases, touching the metal shielding immediately and completely eliminates any jitter and paddle movement is flawless.

 

Unfortunately I won't have access to my real 2600 setup for some time to test there, but the paddles worked perfectly on it previously.

[Rodney Hester]

On 2/5/2023 at 8:53 AM, Jstick said:

For the paddles to show as 2 controllers, a workaround was implemented in the MiSTer core (the 2600daptor d9 presents paddles as multiple axes on a single controller, MiSTer requires it to be 2 separate controllers).  The paddle button mapping is part of that workaround.

That was sort of the thing I was driving at - no workaround should have been needed.

 

If the input device is a D9, in ALL cases the input can be treated as a 2-button 7800 controller (with B and A buttons), where the analog axes map to the joystick directions.  This works for the CX40 and clones, the 7800 controller (stick and pad), and paddles, all when connected to the D9.  Not only is there no need for a separate single paddle button (which makes the second paddle's button useless in multi mode anyway), there's no need for complication of any kind.  The analog axes of the joystick conveniently already map to the proper directions for the paddles in both single and multi mode.

[Rodney Hester]

9 hours ago, Jstick said:

Ok, I had a chance to test this more thoroughly (2 different D9 adaptors on latest MiSTer firmware, 2 USB cables, MiSTer 7800 core and Stella on PC).  Here is what I've found:

Two quick questions:

 

- What MiSTer USB hub are you using exactly (i.e. where did you get it from)?  Different vendors handle supplemental power differently, and I suspect some may indeed have grounding issues.  For reference, see this thread: https://misterfpga.org/viewtopic.php?t=840 and this product: https://ultimatemister.com/product/usb-hub-2-1-with-cables-and-bridgeboard/.  Note that USB hub is unique in that it taps 5V DC power for USB directly from the bottom of the DE-10 Nano via the two pins in the upper-left corner of the photo and therefore does not require external 5V power via barrel jack, however, that also means that grounding for the entire board is achieved solely through the USB bridge and may be insufficient.  (I am NOT an electrician and would welcome further input here!)

 

I also found this commentary specific to poor grounding as part of the design: https://misterfpga.org/viewtopic.php?p=21713#p21713

 

Schematic for the 'generic' USB hub v2.1 (with barrel jack) here: https://github.com/MiSTer-devel/Hardware_MiSTer/blob/master/releases/USBHub_2.1.pdf

 

- How are you powering your MiSTer (is it via the I/O board or directly into the DE-10 Nano?), and is the power source 2-prong or 3-prong (i.e. does the power source have explicit grounding)?

Edited February 8, 2023 by Rodney Hester

[Jstick]

47 minutes ago, Rodney Hester said:

That was sort of the thing I was driving at - no workaround should have been needed.

 

If the input device is a D9, in ALL cases the input can be treated as a 2-button 7800 controller (with B and A buttons), where the analog axes map to the joystick directions.  This works for the CX40 and clones, the 7800 controller (stick and pad), and paddles, all when connected to the D9.  Not only is there no need for a separate single paddle button (which makes the second paddle's button useless in multi mode anyway), there's no need for complication of any kind.  The analog axes of the joystick conveniently already map to the proper directions for the paddles in both single and multi mode.

That makes sense, unfortunately the MiSTer is simply not set up to accept 2 separate controllers (i.e. P1, P2) from a controller presenting itself as a single device.

 

44 minutes ago, Rodney Hester said:

Two quick questions:

 

- What MiSTer USB hub are you using exactly (i.e. where did you get it from)?  Different vendors handle supplemental power differently, and I suspect some may indeed have grounding issues.  For reference, see this thread: https://misterfpga.org/viewtopic.php?t=840 and this product: https://ultimatemister.com/product/usb-hub-2-1-with-cables-and-bridgeboard/.  Note that USB hub is unique in that it taps 5V DC power for USB directly from the bottom of the DE-10 Nano via the two pins in the upper-left corner of the photo and therefore does not require external 5V power via barrel jack, however, that also means that grounding for the entire board is achieved solely through the USB bridge and may be insufficient.  (I am NOT an electrician and would welcome further input here!)

- How are you powering your MiSTer (is it via the I/O board or directly into the DE-10 Nano?), and is the power source 2-prong or 3-prong (i.e. does the power source have explicit grounding)?

I'm not using any of the MiSTer hubs, but rather a standard external USB 2.0 hub.  I have tried multiple hubs, powered and passive, with the same results (I've never had any issue with these hubs, which I've been using with the MiSTer for years now).  In any case, the fact that I get the same results in Stella when plugged directly into my PC without a hub pretty much rules out the hub being the issue.

 

I'm powering MiSTer via the 2-prong power adapter included with the DE10 Nano, connected directly (No I/O board).  I have nothing else connected during testing that draws any load (besides the D9), so the PSU should be sufficient.

Edited February 8, 2023 by Jstick

[dualcam]

11 hours ago, xeex said:

Yes, it was the same cable. I tried two different cables, actually, on MiSTer and in Stella, but didn't notice a difference between them. In all cases, touching the metal shield around the DE-9 connector eliminated jitter. I'll try to cobble together a video demonstrating what happens in Stella, but that might take me a while. And I still have to test these paddles on original equipment.

 

11 hours ago, Jstick said:

Ok, I had a chance to test this more thoroughly (2 different D9 adaptors on latest MiSTer firmware, 2 USB cables, MiSTer 7800 core and Stella on PC).  Here is what I've found:

 

If the dips are set to Autosense, paddles work properly in Stella (tiny bit of infrequent jitter).  In MiSTer, there is some jitter at the low end (paddle left) which reduces about half way across the screen and disappears when reaching the right of the screen.

 

If the dips are set to MiSTer mode, there is a large amount of consistent jitter in both Stella and MiSTer.  It appears that the scaling may be amplifying the amount of inherent jitter.

 

In all cases, touching the metal shielding immediately and completely eliminates any jitter and paddle movement is flawless.

 

Unfortunately I won't have access to my real 2600 setup for some time to test there, but the paddles worked perfectly on it previously.

 

OK, I will message you guys about returning the adapters so I can check them here.

 

Tom

http://2600-daptor.com/

[Rodney Hester]

6 hours ago, Jstick said:

That makes sense, unfortunately the MiSTer is simply not set up to accept 2 separate controllers (i.e. P1, P2) from a controller presenting itself as a single device.

Ordinarily, yes, but that's precisely what multi mode (in the 7800 core) is for.  If it can separate axes into separate controllers, why not buttons?