OS Rom to boot to memory test

[xrbrevin]

I just obtained an XEGS which will only boot diagnostic carts. Ive burned a new OS ROM but the fault remains.

Does anyone know of an OS Rom image that boots straight to a memory test? I cannot use my Sys-Check device due to it being an XEGS..!

Cheers!

[+mytek]

I'm pretty sure you tried this, but just in case... did you try holding OPTION on power-up?

[+Nezgar]

Any difference in behaviour with/without keyboard attached?

 

Does the memory test find any faults? Does it test 40 or 48 blocks? (40 means BASIC is enabled)

 

Maybe try programming Shoestring's memory tester replacement OS ROM to thoroughly test the RAM? You'll need to arrange the ROM  image with BASIC in the right place for it to pass the BASIC test too.

 

XEGS 32KB ROM offsets for 27256:

 

$0000-$1FFF - 8 KB Missile Command

$2000-$3FFF - 8 KB BASIC

$4000-$7FFF - 16 KB Operating System

[xrbrevin]

thanks for the replies.

 

1 hour ago, mytek said:

I'm pretty sure you tried this, but just in case... did you try holding OPTION on power-up?

alas, no joy. i dont even get the fart noise and not seen basic thus far. With no (diag-spec) cart i just get a black screen. SIO devices do not respond.

 

30 minutes ago, Nezgar said:

Any difference in behaviour with/without keyboard attached?

yes, this is an odd thing - with keyboard attached it will boot to missile command. if i start a game the cursor just goes to top-left regardless of whether a joystick is connected (pokey issue?). there is a low-tech historic repair on the keyboard cable that i have not tidied up yet tho.

 

The first boot up went straight to memory self test. it got past half way with no errors and then reset itself to a black screen. Ive never managed to get back into memory test since.

The SALT cart shows RAM: 40k, TIA: PAL, ROM: ERR, NTSB: ERR. The 40k tested fine.

 

SuperSALT flashes up a brief orangey-brown screen with SYSTEM ERROR F4. Sometimes it will emit a high pitched tone.

 

I have swapped out thus far: OS ROM, MMU, Freddie, 2x DRAMs. Im wondering if there is an addressing issue with the CPU but its not socketed. (pesky XE desoldering..!)

 

I'll get that Shoestring ROM burnt and see what it does. Many thanks for the link! Just gotta wait for the UV eraser to work its magic.

 

Edited December 20, 2020 by xrbrevin
SuperSALT info

[+Nezgar]

@xrbrevin is your XEGS motherboard missing components from L34-41? If so, take a look at this "Tech Tip" from tf_hh, the symptoms maybe similar... Fix is to solder two 1nF capacitors at least on the keyboard RX lines:

 

 

[xrbrevin]

My L34-41 is populated with beige components but i like the idea of trying the mod. i have ordered some 1nf SMD caps so while i wait i will swap out the CPU

 

[+DrVenkman]

With the additional joystick problems, sounds like a look at PIA is in order to me.

[xrbrevin]

ok, thats another 40pins of xe-spec desoldering.

stronger coffee required..!

Sadly, the Shoestring ROM just made a high pitched tone.

[+Nezgar]

31 minutes ago, xrbrevin said:

Sadly, the Shoestring ROM just made a high pitched tone.

You will only hear that screetching noise with a blank screen for the first 15 seconds or so before anything shows on the screen. Then I believe it needs another 5-10 minutes to complete a full test.

[+mytek]

4 hours ago, xrbrevin said:

odd thing - with keyboard attached it will boot to missile command.

Sounds like SELECT is stuck on, because this would be the typical behaviour of powering up with the keyboard connected and SELECT pressed.

[xrbrevin]

12 hours ago, mytek said:

Sounds like SELECT is stuck on

i think its more a case of garbage signals. i can boot to missile command if i hold down select, it doesnt if i dont.

here is a pic of the shoestring ROM result, it appears this way immediately. i dont think it supports the XEGS directly as the silkscreen 'U' numbers are wrong but it does support the 2x DRAM chip 65XE (4464) which the XEGS also uses.

I get same results with a homebrew GAL MMU and same results with a change of DRAMs.

time to desolder the CPU..

 

 

[+Nezgar]

Swapping SALLY sounds like a good course of action, as I've read there were more problems with later production batches, but another thought (with less pins to desolder) is to try replacing the 74LS08 buffer - with a faster 74HCT08, or even faster yet 74F08, 74ACT08 or 74ABT08.

 

Good read on this thread:

 

[xrbrevin]

On 12/20/2020 at 8:50 PM, DrVenkman said:

sounds like a look at PIA is in order

Thanks to all for the input, DrVenkman called it - the PIA! (Pesky Infantile Article? hehe)

 

Heres how it went:

• Desoldered the CPU, supplanted but no change
• Multimeter'd all caps, resistors and transistors - nothing untoward found
• Found a spare 74LS08 - piggybacked it on but no change
• Desoldered PIA - replaced and it booted to missile command (no keyboard so this is correct), reset - into self-test. rebooted with keyboard - got basic, all carts work, SIO works!

I have had a PIA casualty before, but luckily they are a universal off the shelf part.

 

Great success! The machine is now playing demos from the Side2 cart 😎

Now i have to tackle the keyboard internal dupont connector botches. Hopefully it will bring the faulty keys back.