Re: Speaker Response?

[Eric Jacobs <no@no.no>]

Colin Klipsch wrote:
> 
> Greetings.  I have an obscure question about the Apple II's speaker
> behavior.  (This would be the speaker found on the 8-bit Apple IIs, not
> the //gs.)  I've searched through the on-line technical notes and was
> unable to find the information I'm after.
> 
> I'd like to know the speaker "latency" -- the time it takes for the
> diaphragm to move from one extreme to the other, after a reference to
> a $C03x I/O address.  Let us assume that the diaphragm was already
> resting at one extreme, and that no $C03x address is touched again
> during the ensuing movement.
> 
> Perhaps this information is hidden in the Apple ][ Reference Manual that
> I have, which talks about the speaker's voltages and impedances.  I'm
> afraid I lack the "audiophile" knowledge necessary to comprehend what
> these numbers are telling me.
> 
> Also, I notice that Electric Duet's music routine is touching the
> speaker (via $C030) in intervals as small as 9 CPU cycles.  I know
> Electric Duet accomplishes its magic by twiddling the speaker very
> rapidly to put the speaker into more than its "ordinary" two states.
> This suggests that the latency period is on the order of 9 cycles, or
> perhaps much more?
> 
> Thanks for any help you can offer.

The output of the flipflop is RC coupled to a Darlington transistor
used to drive the speaker. There is a capacitor (0.1 �f) going across
the transistor, and this mainly serves to absorb some of the coil energy
from the speaker, but also provides some integration for the signal.
This may be contribution to this "speaker lag" that you're talking
about. At each transition caused by $C03x the driving circuitry will
also always produce an overshoot of some amount, on the order of about
a volt or so. The effect is much more noticeable at higher frequencies. 
If you're talking about the total amount of time it takes for the
speaker level to reach its normal voltage, I'd say to wait at least 15 
or 20 microseconds.

By throwing the flipflop on and off rapidly, you can create what is in
effect a crude D/A converter. The flipflop can be held for a while
(charging the capacitor) and then off for a shorter time period 
(partially discharging the capacitor). The voltage level out would be
approximately proportional to the duty cycle you toggle $C03x at. In
fact, you can use this technique to play back digitized sound.

-ej