With Sound/Speech I exits to monitor after inital detection check

[robjustice]

Using version v0.2, and a Sound Speech I in a IIe, exits to monitor after the initial check. I get this output:

mb-audit v0.2-beta, 2021
                    1  2  3  4  5  6  7
              $00:           ?
              $80:           ?
               SP:
65C02 detected

*

[tomcw]

Thanks Rob - I will take a look this this coming weekend.

[tomcw]

Triage:

• The * prompt means that it's vectored to the Monitor's default IRQ or NMI handler.
• It must be NMI, since I've been testing with my cards so should spot a rogue IRQ.
• The mb-audit code hasn't reached DisplayCardUpdate() yet... so the rogue NMI occurs in Check6522() or DetectSubunits().
• Considering that @robjustice has confirmed that play-sc01-using-ssi263 (v0.7) works on his S/S card, and this mb-audit is initially largely the same, then what is the difference?
• In Check6522(), mb-audit has an extra routine to Check6522Addrlines(), and suspiciously this is setting IER (Interrupt Enable Reg) to $96.
  • NB. this routine runs in "SEI" mode (ie. I=1 / interrupts disabled)... but of course this doesn't stop NMI's!
  • Also at this point, the IRQ/NMI handlers haven't been initialised yet - this happens in the next routine, namely Check6522IRQ()

Anyway, if I change Check6522Addrlines() to initially use "CLI" mode (ie. I=0 / interrupts enabled) and try this on my real h/w, then I too get the same exit to monitor for my slot-4 card:

Slot=4, card = "ReactiveMicro MB-C" (after power-on)

• $C40D,E (IFR, IER): $99, $96 ; <-- here's the problem
• $C48D,E (IFR, IER): $19, $80 ; <-- and here

Slot=1, card = "Phasor" (after power-on)

• $C40D,E (IFR, IER): $00, $80
• $C48D,E (IFR, IER): $00, $80

Then after CTRL+RESET, I still get the same values for both cards.

btw, reading $C400 (ORB), then $C40D becomes $01... then reading $C401 (ORA), then $C40D becomes $00. This is correct as the data-sheet says that reading ORB/ORA will clear these IFR bits.

NB. After a CTRL+OA+RESET, then $C40D reads as $11, and $C48D reads as $19. So perhaps some noise that randomly sets these CB1(b4)/CB2(b3)/CA1(b0) IFR bits. And probably best to avoid enabling these IER bits!

[tomcw]

@robjustice - when you get a chance, can you try v0.3-beta here. Thanks.

[robjustice]

That works, and it now gets past the detection and displays: ... 4 S/S S/S V

However :-( , then, i get failed test: SC01

Test: 31:00:00 Expected:00 Actual:01

[robjustice]

ok, i see this is related to the duration of the speech. I had made it a bit faster previously, but must have bumped the adjustment back. If i set it a bit faster it passes this test ok and reaches the hit ctrl-reset for the reset tests.

[tomcw]

Yes, that's it. I will remove the timed check (or at least make it a very long timeout), so that this passes however you have the card's potentiometer set for its speech rate.

[tomcw]

Hi @robjustice - this new build here (v0.4-beta) should remedy the short timeout on the SC-01 phrase playback.

I've extended the timeout to 42s, but it's conceivable that your manual potentiometer setting could make the speech even slower than this! But I figure that if anyone sets the speech rate this slow, then it's user error.

And when you get passed the reset tests, there's a manual tone test (keys 4,5,6 for your S/S card). Then finally a report of the number of tests passed.

[robjustice]

I tested v0.4-beta and that's working well. If I set the manual pot to the slowest, its passing ok, I think you have a good margin in there! I did try the tone test and that's working ok.

I did notice another small issue with the tone test, i'll raise it separately.

[tomcw]

Thanks @robjustice for all your help in testing the various iterations of this mb-audit code. Without real h/w it's very difficult to get things right. I hope this test suite will be useful in providing confidence in your h/w or diagnosing future h/w issues.