Re: RAM-4-GS testimonials

["PZ" <skierpaul@yahoo.com>]

> Some of you out here are geniuses and fully qualified to comment on any
> hardware (or software) design.   Some of you appear to have found
> shortcomings in Garber's RAM card.   So, has anyone ever offered
> friendly and helpful feedback to Garber about how to improve his
> product?    Or just these brutal tear-downs?    People usually have
> such a hard time not getting defensive over assaults like these that
> they can't (or don't want to) possibly address the real suggestions of
> value.
>
> ~ J

Sure J, I'll entertain, but I'm essentially redesigning the entire card
here.

A) Add '245's to buffer the data lines like you're supposed to do for
EVERY card that goes into an Apple II.

B) Add series resistor networks to the RAS/CAS strobes and to the
address lines to cut down on the ringing.  Add hex inverters to the
address lines to maintain strong 5v signal integrety.

C) Learn how to do efficient PCB layout - utilize best practices.  If
its ugly, its probably not good, pointy 90 degree angles don't work
well on a racetrack, they certainly don't work well on PCB's, unless
transmitting RF is your thing - build a radio instead.

D) DON'T RAS strobe all 32 ram chips everytime you wish to only address
one bank. This means you need to scrap that '138 demux and use a pair
of '139's.  Quit being sooo a cheap!  This shortcut only works 100% of
the time if you're using the same DRAM in EVERY card you build (and
you're intimate with their specs), but you're not.  SIMMs, by nature
contain chips, produced by different manufacturers from Korea,
Malaysia, Japan, China, Brazil, India.... well you should be getting my
point, you can't count on anything.

- Paul