Re: Stability of 15 MHz ZipGS

[rubywand@aol.com (RUBYWAND)]

In article <19JUN199615221718@vax2.concordia.ca>, spec@vax2.concordia.ca
(Mitchell Spector) writes:

>....
>Below is what I have "feeding" off my 60 watt supply:
>
>Slot 1: Quickie Interface
>Slot 2: MDIdeas SuperSonic stereo w/digitizer daughterboard
>Slot 3: ZipGS 15/64
>Slot 4: ComputerEyes GS
>Slot 5:
>Slot 6: Apple II 3.5 Controller Card
>Slot 7: RamFAST SCSI (Rev D) 1MB cache
>MemExp: RAM GS 4MB board
>
>- Apple SuperDrive
>- Apple 3.5
>- Apple 5.25
>- Mach III Joystick
>- Quickie scanner
>
>Obviously all that was just too much for a 60 watt supply, with the
>ZipGS being the final straw. Hopefully replacing it with a 150-200
>watt PC supply will solve all this ....

     In an article for AppleWorks Forum in 1992, John Links talks about
trying to push the Zip-GS to 15MHz. He indicates that it _might_ be doable
using 35ns RAM and running the board at 5.5V. He was able to get a Zip-GS
to work reliably at 12.5MHz with 45ns RAM running at 5V.

     Based upon the above, if your 15MHz mod includes the faster RAM and
gets 5.5V via a regulator on the +12V line, then there is good reason to
suppose that it is a viable configuration. (Maybe with the faster RAM,
going to 5.5V is not necessary-- I don't know.)  However, experience with
a stock Zip-GS 10/64 on a ROM-01 machine makes it pretty clear that a
standard IIgs with a Slot load like yours is very likely to experience
glitching with a Zip-GS running anywhere near 15MHz.

     The glitches come mainly from two places:  1. The leads coming from
the Power Supply are too skimpy. At loads much beyond 2.5A they act like
resistors producing a noticeable voltage drop which is propagated as
noise.  2. The Slot power line traces are too thin. They are another
source of voltage drops and noise. The drop in supply voltage (+5 and +12)
to the motherboard and the Slots also has the effect of reducing noise
margins.

     The cure is to reduce the resistances between the PS and the
motherboard and between the PS and power-sucking boards in Slots. Going to
#12 gauge leads for +5, +12, and GND will eliminate significant drops from
PS to motherboard. (The Buggie PS uses fairly hefty leads; so, this could
be a good alternative.)

     Heavier PS leads may get rid of your glitching. If not, the next step
is to run #16 gauge or better leads from the motherboard's  PS connector
pads to +5 and +12 points on Slot 3 and Slot 7.

     There is one more factor to consider. Most internal GS Power Supplies
do not deliver +12V at much beyond 1.25A. If your Zip-GS is powered from
the +12 line, then a combination of other +12 loads could lead to
interrupted supply to the Zip-GS. For example, you would probably want to
get a Buggie PS before adding a Second Sight card.       


Rubywand