Re: Hardware idea

[dempson@actrix.gen.nz (David Empson)]

Dennis Jenkins <djenkins@usb.com> wrote:

>       As for DRAM, I figured that a DRAM device would have no battery.  You
> are correct, it would eat the battery alive.

Even if you did use DRAM, you would also need a refresh controller which
ran while the main power was off.  This isn't an issue with static RAM.

> Would it be possible to use some of the Apple II's build in circuitry to
> do the DRAM refresh?

Not likely.  You could get timing signals to drive a refresh circuit,
but that's about it.
 
> How to the standard IIgs and IIe ram expansion cards handle refresh?

Standard slot RAM cards have their own refresh circuitry.  (The original
Apple language card is a notable exception, but it cheats by running a
cable to one of the RAM sockets on the motherboard to get the refresh
signals.)

A 64K auxiliary memory card in the IIe is refreshed by the motherboard.
I suspect that larger auxiliary slot cards can also use the motherboard
refresh cycles (video accesses) to refresh all their banks.

IIgs memory expansion cards are refreshed by the motherboard.  Some
non-standard cards can't use the motherboard refresh support and have to
implement their own.

>       Will GS/OS be able to use this device if it contains just a standard
> ProDOS block driver?

Yes.  GS/OS will use a generated driver if it find a ProDOS or SmartPort
driver on a card.

> Based upon the recent post about a home brew IDE interface and Dave' comments
> about a smart port driver... How difficult would it be to make the card use a
> smartport interface instead of prodos block driver?

Mainly a lot more code to write.  The ProDOS firmware interface is very
simple, but SmartPort has a lot more calls and more information that
needs to be passed around.  The call interface is also more complex,
resembling that used by the ProDOS-8 MLI.

I've written a ProDOS block driver (for a networked file server).  It
fit into some unused space in the $C800-$CFFF area in the card's
firmware, and made use of existing support for Pascal in the card, so it
was reasonably compact (less than 256 bytes, anyway).