Re: A83: Two graphical problems to solve.
[Prev][Next][Index][Thread]
Re: A83: Two graphical problems to solve.
Hi!
> The delay necessary depends on the calculator. You may be interested
> in looking at the Ion library source (ionf.z80). I use a minimum of 65
> clocks (total, not the delay routine), but that seems to still be too fast
> for some calculators.
Thanks for the tip. I wanted to anyway (and now there's no excuse not
to. (-: )
Is it really calculator dependent or does that depend on the batteries?
IIRC, in devices like Palms, calculator speed is related to the voltage
of the batteries (and the oscillator, of course). Now assuming that the
LCD driver's delay rate is constant, new batteries require a longer
delay in the code, while those already in use may work just fine with a
smaller delay rate.
So 65 (perhaps even 70) clocks is a safe guess. Hmmm ...
In any case, what happens exactly if the delay is too fast for the
calculator?
> And again, see ionf.z80. It certainly wouldn't be hard to port a TI-86
> routine to the 83...
Not this one. The Eble/Yopp routine benefits from from the 86's screen
length of 16 bytes, and a neat trick with the RLD-Routine. There's no
way this can be duplicated, as the 83/83+'s screen length is 12 bytes
and therefore doesn't conform to nibble-boundaries. To be honest, I've
already written a routine, but the ~160 clock cycles can't compare to
the 127 of Ti-86's. I'd like to compare it to other routines in order to
find anything that I may have overseen. (And I still have to test it for
*possible* bugs...)
Thanks again,
Martin
Follow-Ups:
References: