Keyboard stops responding after pressing reset with TK-II Stereo board in SIDE2 Loader

[SpicyChronos]

Hi All,

 

A little background.

 

I have a U1M, SIDE2 combo going in an 800XL.

 

I recently assembled a TK-II Stereo board without the PS/2 keyboard interface.

 

I don't know if I have found a bug in FJCs firmware for SIDE2 (ver3.02), or U1M (ver3.02), or both or not at all. with the TK-II stereo board. Or if it is an issue with the TK-II board, or my 800XL.

 

Here is what I can duplicate.

 

With the A4 line connected to the TK-II board, the computer will boot, other times it won't. I'll just get a black screen.

I have the Stereo option enabled in U1M.

If I disconnect the A4 line from the board and turn it on, it'll boot constantly.

When I boot it up and then connect the A4 line while it's on (probably not safe) and select a game, like Yoomp!, it'll say stereo not detected. I press reset, and it'll then detect stereo.

If I power cycle the computer, SIDE2 Loader comes back up and the keyboard will not respond.

Pressing Help and Reset, and go into the U1M setup, keyboard will respond, and I disable stereo mode and save, the keyboard will work again.

 

 

 

 

 

[Rybags]

Stereo detection methods vary - by introducing the second Pokey after reset it'll present itself in an uninitialised state so possibly not be picked up.

Keyboard failing and general crashes - maybe there's some sort of chip select problem where the second Pokey is also getting the select signal for the first.

Note that Pokey has 2 CS lines which is for the purpose of easily supporting multiples of them, so you'd want to check both those - CS0 is active low and activated by any $D0xx page access by the LS138 chip, CS1 active high - normally I believe is just held permanently high by external pullup but in a stereo Pokey config would be connected to the inverse of address line A4 only instead.

Edited February 22, 2020 by Rybags

[SpicyChronos]

Hi Rybags,

 

I thought it might be shoddy cheap Chinese resistors I used, as two of them are for Pokey select (A4) and Stereo mode (MO). Replacing those, did not help. If I leave everything connected, and switch the Stereo mode in U1M, this is when the keyboard will not respond. So enabled, no keyboard, disabled, keyboard works.

 

According to the schematic for the TK-II (attached), it shows that both Pokey chips are connected at pin 30 (CS0), and CS1 are connected high to 5 volts.

 

I checked all my soldering, no shorts, swapped the POKEY chips, and the same result. I can replace the resistor network (RN3) as this could be the wrong part. and see if it will fix anything. as I might have installed the wrong part.

 

tk-ii-stereo_main.pdf

[flashjazzcat]

8 hours ago, SpicyChronos said:

If I power cycle the computer, SIDE2 Loader comes back up and the keyboard will not respond.

This usually happens stuck or undriven keyboard register, since the internal code for 'L' is permanently present.

8 hours ago, SpicyChronos said:

Pressing Help and Reset, and go into the U1M setup, keyboard will respond, and I disable stereo mode and save, the keyboard will work again.

If stereo was not detected in the first place, there would be no stereo setting to change (the option would be greyed out).

6 hours ago, Rybags said:

Stereo detection methods vary

The method used in the currently available firmware is the one you described on the forums some years ago; specifically, it sets up a timer interrupt on the primary chip and checks to see if the IRQ appears on the secondary POKEY's address space (which it will if there is no stereo board present). In the soon to be released update, that method is dropped in favour of KMK's method which puts the potential second POKEY into a reset state and checks to see if the primary chip is halted. Despite the fact your method appears to work with 100 per cent of existing boards, the detection method was changed to accommodate Candle's latest Simple Stereo board, which apparently does not work with your detection method, for reasons which were never clearly explained.

4 hours ago, SpicyChronos said:

If I leave everything connected, and switch the Stereo mode in U1M, this is when the keyboard will not respond. So enabled, no keyboard, disabled, keyboard works.

The only thing that changes when you turn stereo on or off is a single bit which allows the secondary POKEY to appear at $D21x.

 

One scenario in which the keyboard can become unresponsive is if you run the 1088XEL/XLD U1MB firmware on an emulated machine (in Altirra) without stereo POKEY emulated. This is an impossible condition on real hardware (even if the second POKEY chip is removed), however. The test used in the standard XL/XE U1MB firmware is not known to have any adverse side-effects regarding the keyboard.

[SpicyChronos]

Hello FlashJazzCat,

 

Thank you for the explanation of what is possibly happening with the TK-II stereo board and U1M.

 

I will still look at that resistor network, and verify I put the right part in. I'll update shortly.

 

 

[SpicyChronos]

All,

 

It was the wrong part. RN3 wasn't the resistor network or array.

 

I have it working now. 

 

I didn't mean to step on any toes over this, I appreciate all the help and hard work that goes into designing these great modifications for our little 8-bits.

 

I look forward to seeing the newer firmware updates that come down the pipe, giving us more options to keep these going.

 

Now to wire in the stereo RCA jacks.

 

 

 

[+mytek]

I'm glad you figured it out (without any help from me). I saw your earlier post, and started thinking about it, and was going to verify that you had high speed version of the 4066 (74HCT4066), instead of the much slower CD4066. But then the weekend with the wife kinda put a damper on that. And if it hadn't been that, the resistor networks were my next suspect.

 

[SpicyChronos]

Helo Mytek,

 

I actually had two of the resistor networks wrong. The one for the POKEY select and for the Left Channel. I wasn't getting any audio out. Started troubleshooting at the POKEY chip itself, and had audio at the 74HCT4066, and the one side of the C4. the other side was very quiet, and with the machined socket I installed, I couldn't tell what part I installed. Desoldered it, and it was the wrong part too. Now I have a working stereo board. I plan on fixing my first attempt, as I used the wrong pin headers to plug into the socket of the POKEY, and build one more as I ordered 3 boards from the PCB manufacturer. I have two more Atari's I want to upgrade, another 800XL and a 130XE. Plan on doing other mods to these too.

 

I'm going to try and do Hias's 512K memory upgrade in the 800XL, and for the 130XE, not sure what I'll do, besides the Stereo board, maybe another U1M.

 

Thanks all!

[+mytek]

Was there a problem with the parts list on my website? Seems like you had a lot of wrong parts. Please let me know and I'll be sure to make corrections .

 

[SpicyChronos]

Hi Mytek,

 

It wasn't anything wrong on your website or parts list, I got the parts swapped. I think I was in a rush to build another board, and didn't watch the parts I used. All the bags looked the same from Mouser.

 

It's all good!