[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Zip GSX problems, newbie-ish questions
- Subject: Re: Zip GSX problems, newbie-ish questions
- From: Rubywand <rubywand@swbell.net>
- Date: Tue, 15 Aug 2000 00:24:36 -0500
- Newsgroups: comp.sys.apple2
- Organization: SBC Internet Services
- References: <8n9qr3$1ca$1@eskinews.eskimo.com>
- Xref: supernews.google.com comp.sys.apple2:448
John King writes ...
>
> First ...... I am experiencing random
> crashes on a 8 MHz 16 K Zip GSX card.
> The CPU socket cable had a broken pin,
> and I "fixed" it with a wire, but the
> system still crashes down to the monitor. I ordered a 40 pin IDC socket
> header to make the male end of the cable,
> I thought I would use an IDE cable
> cut at about 6" inches for the female end. Zip users, what are your
> thoughts on this plan? Any serious problems with this?
>
You have your original cable to use as a check (using an Ohm meter) for
which header pins go to which 40-pin plug pins; so, the new cable should work
okay. 6 inches may be a bit long. Usually, the Zip-GSx goes in Slot 3. I'm
pretty sure this requires something less than 3 inches of cable. As a rule
you do not want the cable to be longer than necessary, especially if you plan
on increasing the speed of the accelerator.
By the way (just in case), Slot 3 with the accelerator card inserted is
left at the usual Control Panel default setting. (It is _not_ set for "Your
card".) Also, see the suggested DIP switch settings info at
http://home.swbell.net/rubywand/Csa2ACCEL.html#001 .
> Second .... I read that some or all Zip cards do not sit down in the edge
> connector completely (compared to other cards), and that this my result in
> poor connections. The fingers on my card are scratched up to 1/2 way. If
> this really does cause problems (in your
> opinion ), is it feasible to grind
> off some board material (assuming no traces
> are in danger) to make the card
> sink further into the edge socket?
>
When we had contact problems, the result was a white screen on
power-up-- i.e., we never got to a genuine system crash. Our Zip-GSx card
extends maybe 1/16 inch beyond the contact tabs. Our Slot connector was
barely wiping the lower edges of the card contacts. The fix was to replace
the Slot connector with one that contacts the card higher up (not a very easy
swap).
Shaving the bottom of the card to bring the card contacts lower into the
socket is an interesting idea. Probably, rubbing the card edge on a piece of
sand paper would be safer than using any kind of grinder. Not so sure about
trying to shave anything from the circuit board (to allow deeper seating).
Perhaps cutting grooves in the edges of the Slot connector would achieve the
same effect.
However, if your card contacts seem to be meeting Slot wipers 1/2 the
way up, and if the accelerator is operating at all, then, it seems unlikely
that you need to worry about any bad Slot contacts. (You could check you Slot
contacts with the card removed and look for any wipers that are notably
farther from the center than others. These may be making poor contact. You
can 're-sprong' a wiper using a large needle to gently pry the wiper toward
the center of the Slot. Naturally, though, you do not want wipers on opposite
sides to contact when no card is plugged in.)
Anyway, since there are some other possible reasons for crashes, nothing
may be wrong with your card, the cable, or the Slot connections. One possible
error is installing the Zip Init. The Init is intended for installation on
particualr diskettes. It's function is to over-ride the usual (DIP switch)
settings in order to have things the way you want when some particualr
diskette is booted. If you have installed the Zip Init on your usual boot
disk without paying attention to the settings it establishes, that could be a
source of problems. (Deleting the Init should cure such a problem.)
Another common problem is running a system which, in effect, overloads
the power supply. A Zip-GSx with most memory expansion cards should not be
enough of a load to produce noise and glitching which will crash a system.
If, however, you add a SCSI interface card, Focus hard drive, PC Transporter,
Second Sight board, or similar power sucker, then, you can expect crashes on
most IIgs's using the standard, unmodified, power supply.
One test for overloading is (with power OFF) to pull all of your Slot
cards except the accelerator and memory expansion card. If overloading is the
source of crashes, the system should not crash when the load is reasonably
small. (For more about overloading and a few fixes, see
http://home.swbell.net/rubywand/Csa2POWER.html .)
> Last .... if I manage to get a stable system, would going from 16k to 64k
> result in about a 20% boost in performance?
>
....
The boost from a larger cache RAM depends upon the application you are
running. If 16k is large enough to run an application with minimal cache
loads/writes, more cache will make no difference. In general, the larger the
application, the greater the improvement.
When a Transwarp accelerator's cache is increased from 8k to 32k, the
result for moderate-size applications is a 33% improvement in speed. (See
http://home.swbell.net/rubywand/Csa2ACCEL.html#021 .) Something like that is
probably around the max boost you can expect in going from 16k to 64k on a
Zip-GSx. 20% sounds about right as an average improvement.
Rubywand