Shoestring's Atari 8bit RAM tester

[+Nezgar]

I just built v1.5 into a ROM image for my AtariMax 32-in-1 OS board, works great!

 

But I just noticed something... it shows "DRAM: 0x0800-0xBFFF" - which i presume has a typo and should read "DRAM: 0xD800-0xBFFF"

 

Edit: Nevermind... I'm the one not reading it right...

 

Since I'm here, a suggestion for the column headers:

XL, XLF, XE, 600XL/XEGS(4464)

could even stack "600XL/XEGS(4464)", instead of on same line. There's room on the line above.

[shoestring]

Thanks for the feedback. But I think this needs some re-working as I'm not entirely happy using player missile graphics, for reasons mentioned in post #48.

 

I'll think I'll keep the table and indicate which bit is bad via text instead of using the player gfx. What's tricky is that I need to keep the display list in rom ( and i can't make changes to screen memory on the fly ).There are 8 bits, so that's 8^2 = 256 possible combinations and I doubt that I can store that many combos. Maybe just have 8 of them and stop at the first bad bit detected.

[shoestring]

a8diag1-6.rom

 

CRC16 - E00A

 

Some small fixes. 800XL PMG markers were completely off.

[tzok]

Would it be possible to relocate this program to work from a cartridge ROM instead of OS ROM? Even by a cost of limited functionality?

[shoestring]

Thought about it this initially but there's already Sys-Check V2 and SuperSALT which work really well via the PBI.

 

 

 

 

 

I don't believe the cartridge slot is the best place for my diagnostic, simply because the OS ROM checks GINTLK register to determine if a cartridge is present before doing anything useful. So in a completely dead machine with a potentially bad OS ROM, I fear the my diagnostic program on the cart can't run at all. 

Edited February 11, 2020 by shoestring

[+Nezgar]

Random idea for the hardware gurus... A simple PBI or Cart+ECI PCB that's ONLY function is to override the on-board OS. It could be a cheaper alternative option for diagnostic with shoestrings RAM tester, without having to open the machine. Next step feature would be switch(es) to select between 2 or 4 OS's on the external EPROM.

[tzok]

22 hours ago, shoestring said:

I don't believe the cartridge slot is the best place for my diagnostic, simply because the OS ROM checks GINTLK register to determine if a cartridge is present before doing anything useful. So in a completely dead machine with a potentially bad OS ROM, I fear the my diagnostic program on the cart can't run at all. 

If computer is "completely dead", it won't work neither from OS ROM slot, nor from the cartridge. If you set D7 at $BFFD, the on board OS ROM will do pretty much nothing, leaving full control over the hardware to the cartridge ROM. If D7 in $BFFD is set, then the next thing which is done is the unconditional jump under the address held in $BFFE and $BFFF.

 

Such cartridge would be useful for testing if a working machine is fully working (at least its RAM is fully working), not for testing a faulty machine to find out where is the fault.