Asteroids Deluxe - odd problem...(fixed)

Well, this one had me stumped.

Asteroids D with a RAM note self test - replaced suspect RAM, and then
there were no more tones, and the test screen was normal. However when
you went to attract mode it would flash the HSTD a couple of times, and
then show the screen without HSTD fo ra few moments, that would then
flicker and then the watchdog would cut in and the screen would be blank.

So I figured it was a heat problem - started using a bit of spray on the
chips and found that the alternate IC for the 74LS245 (80something) if
cooled would allow the screen to come back. Great, replaced it with a
245 and the screen was then blank totally - self test and attract...

Now I had put in sockets for both buffers 245/80xx(what is the number?)
by taking a pair of regular 20 pin and splitting them in half and
trimming off the extra bits - so I could plug either chip in its correct
socket.

Now I should mention that I had hooked up the Fluke to this board and
run all tests. Eprom sigs looked great, RAM short or long both passed
for as long as I cared to run the tests - both scratch and Vector RAM.
This was the point where I asked for the AD signature info (thanks!)
figuring I had to go into Signature analysis to find the problem.

After another day of letting it sit on my bench waiting I thought - why
not try the original buffer (assumed defective, but held onto to return
in bag of parts to customer) to see if it seemed to work still using the
cold spray (Note that the LS245 was useless for any screen display - but
the Fluke thought it was just fine as a buffer - RAM/ROM tests passed
just fine as did the Signature Analysis for AD!)?

Yup - it would work as long as I cooled it down. Reset the HSTD 2055
EERAM and that got rid of the intial HSTD display issues but the board
would still not go into attract (with original buffer) unless cooled down.

Then I started to look at the RAM - noticed that none of the RAM matched
at all and though it couldn't hurt to put a matched manufacturers pair
of RAM for the Scratchpad...board now started working just fine in attract.

So, it turns out one of the scratchpad RAM was loading down the Data bus
just enough that the 245 couldn't drive it, but the 80XX could IF it was
cooled down!

I do not want to say how many hours I worked on this board...not that
the customer is going to pay for the time beyond what it should have
taken me...just interesting that loading can pass the Fluke RAM test as
well as the game's self test, but be enough to stop it worm working
properly.

Weird?

I would like to find something to test 2114s and 2101/5101s - none of my
RAM test devices do these...something where one could adjust the speed
and possibly the loading factor...

John :-#)#

-- 
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