Mark

I had a similar problem with the ISD25xx speech chips - very
occasionally a message would be overwritten for no apparent reason.
While I could never isolate the conditions under which it occurred,
it seemed to never happen if a power off  / on had not occurred since
recording so I presumed that the ISD chip was potentially badly
behaved during power up. The solution was a nasty but effective one.
A pnp transistor from Vcc was turned on by the system reset line and
clamped the Isd's record line high whenever RESET is low. This
ensures that the record line is never lower than Vcc (or barely so)
at any stage when the Isd has not got a stable supply. (Attempts to
use more sophisticated solutions failed). I use a Motorola MC33164 as
a reset generator but any of the many supply level monitoring ICs
would do. A cheaper (more parts) solution could be made with 2
transistors and a few resistors.

The 33xxx does not have a R/W line but clamping the /SS line high (as
above) when the power supply is unstable or outside specifications
may well be equally successful.

regards

    Russell McMahon


From: Mark A Winters <markwinters@WORLDNET.ATT.NET>

>I've built a prototype circuit to drive an ISD33000 speech
record/playback
>chip using a PIC 16F84. It seems to work OK *most* of the time, but
>occasionally large sections of the recorded speech are getting
erased. When
>the problem does occur, it seems to be after powering the circuit
down and
>back up again, so I'm guessing it has something to do with a sort of
>brown-out condition that may exist during this time.
>
>Other points that may or may not be relevant to the problem are:
>
>1. I'm running the PIC at 5v and the ISD device at 3.3v,
interconnecting the
>control lines as shown in one of the ISD app notes.
>
>2. My app makes heavy use of the ISD "message queueing" feature to
locate
>the beginning of the various speech segments for recording/playback.