KTRL CD32 - Amiga CD32 custom gamepad

[MickGyver]

Originally Posted by MarkBogaard

Hi, I hope I haven’t been forgotten. My cd32 is coming back from a recap this week and would love to stop using the crappy PlayStation converter haha

You are not forgotten, you are actually the next one in line! I will contact you in a day or two.

Originally Posted by thairacerjp

I MickGyver ! and congratulation for your nice product success , I have a psx with converter and a crappy Amipad32 , did I will have more satisfaction with your pad ? I think yes but can you tell me the advantage ? (multi buttons, setup), thanks !! , it's for CD32, any compatibility problem ?
jean

Thanks!

The thick circuit board and gold plated contacts for the buttons and DPAD makes it feel quite responsive. The configuration options are a nice bonus, especially the ability to map UP to a button. Compatibility with CD32 games is excellent.

The gamepad does not work on all A1200 machines though. The original CD32 have compatibility issues with some A120 machines also but my gamepad is a little more sensitive I guess. On other Amigas (including the CD32) it seems to work well.

[Daedalus]

Originally Posted by MickGyver

Thank you, much appreciated! Do you have another power supply to try? I had problems when developing the KTRL before I recapped my power supply.

No problem! Yeah, there were two different PSUs on the different machines I tried, both in decent shape with good regulation and low ripple.

[MarkBogaard]

Next in line? Yesss!

[MickGyver]

Originally Posted by Daedalus

No problem! Yeah, there were two different PSUs on the different machines I tried, both in decent shape with good regulation and low ripple.

Ok, thanks!

I have found something suspicious when removing capacitor C9 (and C3). When latch (pin 5) goes low, clock/button 1 (pin 6) will also go low before clock pulses are sent from the Amiga, as expected. However, at this point there is some "ghost" pulses on pin 6 (look at the attached scope image) that disappear when i set C9 to 22pF or higher. I guess that if these pulses are interpreted as proper clock pulses then the data will be shifted with an offset, resulting in erroneous button readings.

These "ghost" pulses seems to only go up to 2V and the Atmega shouldn't interpret that as the signal going HIGH but I guess it does. When I add a capacitor for C9, the "ghost" pulses disappear and it works as expected.

So if you have any spare capacitors, you could try replacing C9 with a 100pF or larger capacitor. C9 up to at least 330pF works on my A1200.

[Daedalus]

Very interesting, and looks like a possibly explanation. I haven't looked into it since but I'll try it either tonight or over the weekend and let you know. 2V is kinda borderline for TTL use, so it's strange that those stray pulses don't also affect the original CD32 pad too, but I guess they're just not wide enough at the 2V point to trigger a clock.

[UKShark]

Expressing interest in pre-built.

[Daedalus]

Originally Posted by MickGyver

So if you have any spare capacitors, you could try replacing C9 with a 100pF or larger capacitor. C9 up to at least 330pF works on my A1200.

Okay, so I didn't have any SMT caps smaller than 100nF, but I did have an old 1nF ceramic disc cap, so I said I'd give it a try. Quick test with JoyPortTest and ControllerTest on the A1200 and both appear to be working fine! It's quite a large value compared to your tests, but doesn't seem to affect performance as far as I can tell. In the plot attached, blue is pin 6 and red is pin 9. Pulses now appear in their correct positions with the clock signal, though it's interesting that pin 6 has a short period at ~2.5V before dropping... Anyway, I'll do some playtesting with it over the weekend and see how it goes, but for now it looks good.

Thanks for the help!

[lmimmfn]

Originally Posted by MickGyver

Yep, you are number 5.


The list is not too long but I can't really estimate how much time I have to assemble controllers. I don't accept any payment before I have a controller ready, I add you to my list and will contact you when I have a controller ready for shipping.


Ok, I added you to my list!


I added you to my list and will be in contact later!


I added you to my list already when I got your first message.


Nice, glad to hear that, thanks!


Thank you, much appreciated! Do you have another power supply to try? I had problems when developing the KTRL before I recapped my power supply.

Thanks very much man

[MickGyver]

Originally Posted by Daedalus

Okay, so I didn't have any SMT caps smaller than 100nF, but I did have an old 1nF ceramic disc cap, so I said I'd give it a try. Quick test with JoyPortTest and ControllerTest on the A1200 and both appear to be working fine! It's quite a large value compared to your tests, but doesn't seem to affect performance as far as I can tell. In the plot attached, blue is pin 6 and red is pin 9. Pulses now appear in their correct positions with the clock signal, though it's interesting that pin 6 has a short period at ~2.5V before dropping... Anyway, I'll do some playtesting with it over the weekend and see how it goes, but for now it looks good.

Thanks for the help!

Glad you managed to get it working and thanks again for troubleshooting! With a C9 of that size you might have issues with Ultimate Body Blows CD32 and maybe Pinball Fantasies CD32. UBB is the most "sensitive" CD32 game when it comes to gamepad compatibility, at least in my experience.

The short period of 2.5V on pin 6 is an interesting one. The behaviour is the same with the original gamepad. The way the CD32 gamepad is designed in combination with the how lowlevel (and other CD32 gamepad reading code) works is a bit weird, maybe a "bug"? Pin 6 is normally configured as an input, and the tri-state buffer on the gamepad is outputting the state of the red button (HIGH when not pressed, LOW when pressed). When the CD32 button reading starts, pin 6 is configured as an output and pulled LOW. Now, the tri-state buffer is still outputting HIGH if the red button is not pressed causing a conflict, resulting in a 2-2.5V level on pin 6 (this is the "bug"). After pin 6 is set as output LOW, the latch (pin 5) is pulled LOW. This will disable the red button output on the gamepad (the tri-state buffer output goes to high-impedance state) and will result in a voltage level of 0V on pin 6. Then the clock pulses are sent and the buttons are shifted out from the shift register (Atmega in the case of KTRL CD32) and read by the Amiga.

Maybe this "floating" 2.5V thing is what's causing problems with some A1200 machines even with the original CD32 gamepad?

[gururise]

Any more gamepad avail?

[traxx]

Interested in one assembled. I have A1200 rev 1D4 may I expect some problems being occured and what kind of?

Edit: found the answer

[Daedalus]

Originally Posted by MickGyver

Glad you managed to get it working and thanks again for troubleshooting! With a C9 of that size you might have issues with Ultimate Body Blows CD32 and maybe Pinball Fantasies CD32. UBB is the most "sensitive" CD32 game when it comes to gamepad compatibility, at least in my experience.

Yeah, I figured that anything with particularly tight timing might be upset by it, but as a proof of concept it works. I'll order some more suitable SMT capacitors the next time I'm ordering parts and do some more thorough testing.

Maybe this "floating" 2.5V thing is what's causing problems with some A1200 machines even with the original CD32 gamepad?

Yeah, that was my thinking as well, from memory there are some minor differences in behaviour with some revisions of the CIA chips - specifically that one revision allows you to emulate fire button presses through poking the register whereas the other doesn't.

[phantomphanatic]

MickGyver thanks for adding me to the list!

[wsoltys]

Just wanted to ask if I'm still on the list
Thanks.

[o-leif]

Hi
Am I still on the list? Ordered two assembled controllers

Terve,

Interested buying one assembled and tested gamepad.

[Glen M]

Quick query if I may....

Is there any maximum length to the cable on the joypad? I would ideally like to make it 5 meters so I can sit on the SOFA comfortable and play CD32 on the big TV across the room.

[MickGyver]

I'm assembling a batch of controllers at the moment. After that I need a break, I wont be able to assemble gamepads for everybody on the list now. I need to find somebody that's willing to buy some kits from me and that would assemble them and sell them on (or I need to find another solution). Doing everything alone takes WAY to much time. I have already sold a batch of kits to somebody here on EAB, I'm hoping he finds the time to complete the kits.

Originally Posted by gururise

Any more gamepad avail?

I have a list where I add people interested in getting a gamepad. I added you to it, let me know if you want to be removed.

Originally Posted by UKShark

Expressing interest in pre-built.

I added you to my list!

Originally Posted by traxx

Interested in one assembled. I have A1200 rev 1D4 may I expect some problems being occured and what kind of?

Edit: found the answer

You are on my list. Just to let others know, I explained that the KTRL CD32 won't work on all Amiga 1200 machines. The original CD32 gamepad doesn't work on all A1200 machines, the KTRL might be somewhat more sensitive, but with the new components I have heard good results, still no guarantee though.

Originally Posted by Daedalus

Yeah, I figured that anything with particularly tight timing might be upset by it, but as a proof of concept it works. I'll order some more suitable SMT capacitors the next time I'm ordering parts and do some more thorough testing.

Yeah, that was my thinking as well, from memory there are some minor differences in behaviour with some revisions of the CIA chips - specifically that one revision allows you to emulate fire button presses through poking the register whereas the other doesn't.

Looking forward to hearing about the results when you have found the time to do some more testing.

Originally Posted by wsoltys

Just wanted to ask if I'm still on the list
Thanks.

You are indeed.

Originally Posted by o-leif

Hi
Am I still on the list? Ordered two assembled controllers

Yep, you are on my list.

Originally Posted by Kettu

Terve,

Interested buying one assembled and tested gamepad.

Ok, I added you to my list!

Originally Posted by Glen M

Quick query if I may....

Is there any maximum length to the cable on the joypad? I would ideally like to make it 5 meters so I can sit on the SOFA comfortable and play CD32 on the big TV across the room.

I'm not sure, I haven't really tested. I have tested a 3m cable and that worked fine. The cable on the KTRL is about 1.8m. I haven't been able to find any longer cables for a reasonable price. A will do some testing with one or two extension cables and will let you know about the results.

payment sent my friend

[MickGyver]

I got a new batch of PCBs (90pcs) that are a little different than the old ones. They don't seem to work with the same components as the old batch. I need to figure out the problem but at the moment I don't have any more time to put into this project. Because of this I won't be selling any assembled gamepads nor kits at least for a while, I will probably not assemble any more gamepads myself whatsoever. I will of course keep my list of unfinished "orders" if anything changes.