| T.R | Title | User | Personal Name
 | Date | Lines | 
|---|
| 1014.1 | A diagnostic trick I use | PRNSYS::LOMICKAJ | Jeffrey A. Lomicka | Fri Nov 16 1990 10:29 | 25 | 
|  | The 520 to 1040 upgrade you are performing IS one that I would trust. 
There's no weird 'chip socket' connectors and the power lines are
already installed in etch on the board.  All you are doing is
installing parts that Atari would have installed if the machine were a
1040.  It should work great.
It's the number of original 520's with "EZ-RAM" or similar stuff
installed in them that I've seen act "flakey sometimes" that concern
me. I don't really understand it.  Many of them are solid and run for
years with no trouble, but I've seen a few that NaVAUS members and
others have that just won't work quite right with these upgrades.  Note
that they pass all the diagnostics - but they flake out when running
real applications.
Regarding the problem at hand.  One thing you might try is writing a
little program that puts Physbase() in the expanded memory area.  This
works even when the MMU is configured for 512K.  At this point, any bad
RAM chip will be seen as a pattern of vertical lines every 16 pixels. 
If you ground the data out pin of one of the RAM's, you will see the
corresponding vertical line for that RAM.  If you work your way accross
the RAM array, you will eventually hit one that lines up with the
pattern on the screen.  That's the suspect chip.
If you don't get video that looks like uninitialized RAM, there may
still be something left out on the board.
 | 
| 1014.2 | S/W is still a minefield to me!!! | FILTON::SMITH_T |  | Fri Nov 16 1990 11:38 | 20 | 
|  |     Hi Jeff and thanks for the input
    
    	Unfortunatly as a H/W Engineer I have yet to delve into the deeps
    of S/W and would,nt have a clue about the Physbase you were discussing
    
    Any chance of some more info on those lines and the more idiot proof
    you can make it the better. I have a few programs that allow me to see
    how much memory I have after loading but it always calls up only the
    first 512kb and another that allows me to size from a 1040 back to a
    520 and vise versa to enable a user to test if the program works with
    only 512kb etc. They both tend to suggest that they dont see the new
    memory but what you are suggesting sounds as if it may do the trick.
    
    Any more suggestions greatfully appreciated
    
    
    					Yours
    
    					Tim Smith
     
 | 
| 1014.3 |  | ATHINA::EMMANUEL |  | Sun Nov 18 1990 12:52 | 13 | 
|  |     
    To date I have done four ram upgrades on Rev D boards with only one
    problem. There is nothing "strange" about the udgrade you are
    attempting.
    I assume you have used ic sockets for the ram upgrade,this was the
    source of the problem I had once. 
    If you haved checked everything else try ohming out from the chip to
    the MMU and you just might get lucky.
    Quite oftenly some ics don't contact well on cheap ic sockets.
    
    Good Luck!
    
    rgds Memnon.
 | 
| 1014.4 | exit | FILTON::SMITH_T |  | Wed Nov 21 1990 06:40 | 30 | 
|  |     Hi again chaps
    	Since the last inputs and the private mail (thanks Jeff), things
    have progressed a little. I have had a chat with the guy who sold me
    the chips and he has explained about the MEMTEST program he included.
    
    	Unfortunatly since that point I spent some time investigating the
    other two links I found on the PCB marked W3 and W4. Lord only knows
    why I did this but in a fit of "lets see what these do" I desoldered
    them in turn to monitor the result. Thats when my troubles latest began
    as The ST No longer boots. Somehow it affected either the MMU or  
    Memory or possibly the disk controller circuitry as since this point
    whenever you autoboot you get bombs. If you boot without a disk to the
    windows she works ok until you access a disk then its more bombs. I am
    taking home a scope this weekend and intend to crack this and the
    memory problem once and for all.
    
    But use this as a lesson to all potential "Have a GO'ers"
    
    DONT PLAY WITH WHAT YOU DONT HAVE INFO ON (just like we tell our
    customers Eh!)
    
    I will eventually report back the answers to thes problems so we can
    all benifit from it but this may take a while due to impending Tapes
    courses.
    
    Keep your fingers out!!!!
    
    Smithy
    
     
 | 
| 1014.5 | NOW ITS FIXED !!! | FILTON::SMITH_T |  | Tue Dec 31 1991 07:32 | 16 | 
|  |     HI ALL
    
    I have just noticed that I never replied to update on the fix needed
    for the problem of the bombs on boot up.
    
    The problem was eventually tracked down to a faulty MMU chip. This
    caused a corrupt part of memory meaning any autoboot tried resulted in
    a series of bombs some of which would only come out with the top or
    bottom half showing. If you tried to boot from the desktop the result
    was the same although presumably the part of memory alocated to the
    running of the desktop was still OK. Replacing the MMU at a cost of 25
    quid did the trick. 
    
    Cheers
    
    Tim Smith
 |