Re: Asteroids Deluxe - odd problem...(fixed)

From: Mark Shostak <shostakmark_at_gmail.com>
Date: Mon Aug 30 2010 - 01:58:34 EDT

> The Fluke test does not occur at system speeds, rather it simply dumps the
code
> onto the bus via a PIA, then toggles the R/W and reads it back through the
same
> PIA...slow...won't catch RAM problems related to speed unless you are
using one
> of the Z80QT pods...

Just to clarify, the Fluke pods _do_ run ram (and other) tests on the bus
with the normal cycle times of the UUT (i.e. at full speed). Accesses to UUT
memory space are performed directly by the uP in the pod (same uP type as
that of the UUT), which is clocked directly from the UUT clock, and don't
have to traverse the PIA, which would be slow. The CAT Box, however, does
use this slower method.

Also, the PIA is actually a RIOT (6532), which provides Ram and Timers for
the uP in the POD, when it is running "locally" (i.e. not running the UUT),
and also provides limited IO (hence the R, IO & T in the name) for
communicating with the 9010, and local GPIO for sensing and controlling bus
access (e.g. UUT power sensing, etc.).

-Mark

On Sun, Aug 29, 2010 at 10:07 PM, John Robertson <pinball@telus.net> wrote:

> Kevin Moore wrote:
>
> A tool like that would be very useful. I've run into this situation a few
> times before. You would think the Fluke long Ram test would catch it. But it
> doesn't sometimes.
>
> Kevin
>
>
> The Fluke test does not occur at system speeds, rather it simply dumps the
> code onto the bus via a PIA, then toggles the R/W and reads it back through
> the same PIA...slow...won't catch RAM problems related to speed unless you
> are using one of the Z80QT pods...
>
> Hmm...I wonder if I could make CPU adapters for the Z80QT pod, and as I
> understand the pod runs RAM/ROM tests at system speed it is a more accurate
> test. The QT pods also can generate a Checksum instead of the Fluke
> Signature for ROM checking. They are a bit rare though...
>
> John :-#)#
>
>
> On Sun, Aug 29, 2010 at 9:35 PM, John Robertson <pinball@telus.net> wrote:
>
>> Matt J. McCullar wrote:
>>
>>> Well done, John! I know what it's like to fix a "tough dog." Just last
>>> week I had a _Ms. Pac_ motherboard that turned out to have a bad 2114 RAM
>>> --
>>> it kept cycling through the initial boot-up sequence. I had an external
>>> RAM
>>> tester that said it was okay (of course), but I've seen this sneaky type
>>> of
>>> thing happen before. Substitution nailed it.
>>>
>>> You're right about mixing up different manufacturers of RAM chips!
>>> Guaranteed to cause all manner of weird things to happen.
>>>
>>> Matt J. McCullar
>>> Fort Worth, TX
>>>
>>>
>>>
>> This sort of thing - RAM that passes tests by test fixtures, but fails in
>> the 'real' world - is a pain.
>>
>> I'm thinking of something that plugs into the CPU socket and after you
>> 'tell' it where the RAM is located allows you to dial up the speed you want
>> to test it at - or use the system clock. A simple Pass/Fail would suffice,
>> you figure out which RAM is defective by the allocated space and if you are
>> testing upper or lower bits (0 - 3, 4 - 7). This could be used with any
>> static RAM...
>>
>> First though I'm going to dig out my old Atari RAM/ROM tester (comparator)
>> and see if it can be used this way - why design something new when I may
>> have a tool on hand that can be easily modified to do the job I want? Now to
>> track down my manual - TM-120 is the one for the Atari Automatic ROM/RAM
>> Tester...probably in the filing cabinets, and I see no scans posted at the
>> usual suspects.
>>
>> John :-#)#
>>
>> (Vancouver, BC, eh?)
>>
>>
>> ---------------------------------------------------------------------------
>> ** Unsubscribe, subscribe, or view the archives at
>> http://www.vectorlist.org
>> ** Please direct other questions, comments, or problems to
>> chris@westnet.com
>>
>
>
>
> --
> John's Jukes Ltd. 2343 Main St., Vancouver, BC, Canada V5T 3C9
> Call (604)872-5757 or Fax 872-2010 (Pinballs, Jukes, VideoGames)
> www.flippers.com
> "Old pinballers never die, they just flip out"
>
>

---------------------------------------------------------------------------
** Unsubscribe, subscribe, or view the archives at http://www.vectorlist.org
** Please direct other questions, comments, or problems to chris@westnet.com
Received on Mon Aug 30 02:00:23 2010

This archive was generated by hypermail 2.1.8 : Mon Aug 30 2010 - 08:50:02 EDT