Yarrb - redesigned

[DaveLecky]

Hi again, it’s started to error on other games again.

Chuckle Egg seems ok for some reason..

Dave

[cjm]

Hi Chris,

Yes, the pin was out ok at installation, I did reinsert it during your tests, it is back in now.

Dave

That might be your problem - pin 10 on IC44 should not be connected when the YARRB is installed.

[DaveLecky]

Hi cjm, that was a misprint, it was out after your tests.

Dave

[cjm]

Hi cjm, that was a misprint, it was out after your tests.

Dave

OK - it needs to be out for those mods to work. I'll update my earlier post to make that clear.

[hoglet]

Started ok, but when killed, the screen started to scroll up endlessly as before,until reset.
I have played this game and passed this point a few times, before the error, a couple of days ago.

Givin the diversity of symptoms, I think there multiple issues here.

The rolling screen symptom is quite distinct. It can only happen if the HDMI adapter stops receiving vsync (but continues to receive hsync).

I can only think of two ways that can happen:
1. A bad connection on the vsync signal (e.g. a dodgy socket for the 6847)
2. A stray write to the 8255 control port (#B003) that sets Port C upper to be an output (causing a conflict).

As the rolling is fixed by pressing BREAK, I think (2) is much more likely.

During the reset code, the Atom initializes the 8255 by writing #8A to #B003. This sets Port C upper as inputs, so the vsync signal be be read by the 6502.

I can reproduce the rolling on my Atom by doing ?#B003=#82. This clears bit 3 which sets Port C upper as outputs. It seems when there is a conflict between the 6847 and the 8255, with both driving vsync, then the 8255 wins.

This could be just another manifestation of the known YARRB clock issue that Roland first encountered here:
viewtopic.php?p=250933#p250933

Until this is addressed, I would leave the noise killer out. That way, we are just dealing with one issue at a time!

Dave

[DaveLecky]

Hi Dave,

Thanks for your input, I suppose I can reflow the connections on the 6847, but that’s about it.

I’ll leave out the noise killer as you suggest, I certainly notice the difference in screen clarity with it out.

In your posts regarding the hdmi cable being cut at core 4 ( I think) to stop crosstalk.....should I do this or has the problem been programmed out.?

Dave

[DaveLecky]

Hi Folks,

Would changing the two crystals be worthwhile...?

Dave

[hoglet]

Thanks for your input, I suppose I can reflow the connections on the 6847, but that’s about it.

As I said, I don't think that is the problem here.

I’ll leave out the noise killer as you suggest, I certainly notice the difference in screen clarity with it out.

I just suggesting to leave it out temporarily, while we debug the YARRB issue.

Once you have a stable YARRB, then try adding the noise killer back in.

In your posts regarding the hdmi cable being cut at core 4 ( I think) to stop crosstalk.....should I do this or has the problem been programmed out.?

The Pi RGBtoHDMI software has been made more robust to this issue. If you aren't seeing any issues, then I would leave as is.

Dave

[cjm]

Hi Folks,

Would changing the two crystals be worthwhile...?

Dave

If there is a clock issue then it's worth trying the mod again but this time with pin 10 of IC44 bent out.

viewtopic.php?f=44&t=19947&p=278886#p278886

Chris

[DaveLecky]

I’ll give it a bash later....

Dave

[DaveLecky]

Hi Chris,

I remade your mod again, this time I left out pin 10 on ic44 as directed.

The problem is still there when I killed in Asteroids, but other newer games, like chuckle egg, run ok .
I wonder if everyone has the same problem with this game on hoglets menu A.

Could someone check this game for me to see if it is an existing problem
Symptoms....after playing a couple of screens, when I get killed the screen starts to scroll up...ta.
I’ve been using this game as a test, so if it’s faulty, I’ve been getting led up the garden path....

I re-ran Rolands wee program, viewtopic.php?f=44&t=18020&start=240#p296160

Is has run for 20 minutes now with no deterioration of the screen like before, so that is promising.

Thanks

Dave

I have reinstated my original IC’s too, so I’ll see what happens tomorrow when I look at it again.

[chrisn]

The problem is still there when I killed in Asteroids, but other newer games, like chuckle egg, run ok .
I wonder if everyone has the same problem with this game on hoglets menu A.

Could someone check this game for me to see if it is an existing problem

Asteroids is working OK on my Atom + Yarrb2.

[hoglet]

Could someone check this game for me to see if it is an existing problem

It seems alright for me as well - I even managed to get through the asteroid storm once!

Does the start of the rolling always co-incide with the Beep at the end when you hit an asteroid?

Does this program also trigger it?

Code: Select all

DO P.$7;U.0

This should just beep continuously.

Dave

[DaveLecky]

Thanks for checking, mine doesn’t seem to like the old commercial games.....weird.

Dave

[DaveLecky]

Hi Hoglet,

It does start with the beep at the end, and your program triggers it too.

It sounds like a machine gun, continuous single bleeps.

Dave

[hoglet]

It does start with the beep at the end, and your program triggers it too.

OK, that's interesting.

Code: Select all

  Handle <BEL> subroutine
  -----------------------

- Returns with X=0, Y=128, and the sign flag set.

FD1A  A9 05     LDA @#5         Get control code to set 8255 PIA port bits
                                C0-C3 to input; A, B, and C4-C7 to output
FD1C  A8        TAY             Set up outer loop counter
FD1D  8D 03 B0  STA #B003       Set port C0-C3 to input, so speaker O/P=1
FD20  CA        DEX             )
FD21  D0 FD     BNE #FD20       ) ..a short delay
FD23  49 01     EOR @#1         Toggle C0-C3 between input and output
FD25  C8        INY             Increment outer loop counter
FD26  10 F5     BPL #FD1D       ..continue for 122 outer loop cycles
FD28  60        RTS

The "Bell" code in the OS alternately writes 5 and 4 to #B003. I think on your Atom, occasionally one of those writes is also setting bit 7, which changes the direction of port C and triggers the screen rolling.

Another data point, in the screen shot you posted earlier:

Most of the corruptions correspond to a setting of bit 7 and/or bit 5, i.e. they follow a pattern, rather than just being random.

Other than pointing this out, I don't really have any further concrete suggestions.

Roland, can you check the photo here:
download/file.php?id=63422&mode=view

Is JP2 in the right position for a SY6502A?

(For an NMOS 6502, don't you have to link pins 1 and 2, which I think are the left and centre pins?)

Dave

[roland]

Roland, can you check the photo here:
download/file.php?id=63422&mode=view

Is JP2 in the right position for a SY6502A?

(For an NMOS 6502, don't you have to link pins 1 and 2, which I think are the left and centre pins?)

JP2 is in the wrong position. Like Hoglet says, for a NMOS 6502 (and the older 65C02 CPU's like the Rockwell R65C02) the pins 1 and 2 are linked. The other position is for the "modern" W65C02 from Western Design Center. Having this jumper in this position(2-3) will connect the incoming CPU clock to the PHI2 line that goes back to the Atom main board. This can cause unpredictable results due to timing issues.

[DaveLecky]

Hi Dave and Roland.

I changed over the position of the jumper to the left, and it booted up fine and I tested the Asteroid program and others.

The YARRB was installed as a new install, with my own IC’s inserted, and the noise killer.

The Asteroids game worked ok, but did crash a couple of times without finishing the game properly, leaving part of the game on the screen and not going to the, “ you are dead screen “.

I had to go out for a while, and when I returned and switched on, the Atom MMC menu slipped/crashed again....SH!^^....

I replaced the 47 ohm resistor again, and it has continued to work ok, running Asteroids 20 plus times.

I will insert Rolands chips without the 47 ohm resistor, later tonite, to see what happens.

Thanks for your help during this journey guys.

Roland, It may be a good idea for me to purchase the IC’s you sent, just in case of future problems, if that is ok with you.
If you need them back, it’s no problem, I’ll return them in the box you sent.....Let me know your thoughts by PM.
I still owe you for the postage and the AtoMMC rom.

Dave

[roland]

I replaced the 47 ohm resistor again, and it has continued to work ok, running Asteroids 20 plus times.
I will insert Rolands chips without the 47 ohm resistor, later tonite, to see what happens.

O NO!!!! If it works ... don't touch it , leave it as is

Roland, It may be a good idea for me to purchase the IC’s you sent, just in case of future problems, if that is ok with you.
If you need them back, it’s no problem, I’ll return them in the box you sent.....Let me know your thoughts by PM.
I still owe you for the postage and the AtoMMC rom.

I'll pm you about that.