Poorly Atari 1040STE

[BrettRogersUK]

I've got a diagnostic cartridge but with nothing showing on the screen except a load of whiteness I'm not sure it'll be much help.

Have you got any diag cart output over serial?

I don't know, haven't tried that yet. I will try it over the next few days and report back. Not very well at the moment.

Cheers,

Brett

[rubber_jonnie]

Well......... as expected, when removing the switcher using my new desoldering gun, more tracks lifted (I don't know why, I was being very careful) when removing it. It was mainly the links between the address lines on the HI/LO rom sockets that lifted. So I traced these and put them right.

Hey Brett.

Usually the reason pads lift is because they've had too much heat and the bond between the copper and fibreglass breaks down. Where the boards are multilayered, 2 in the case of the STE, and the vias are through hole plated, it does tend to make them a bit more resilient, but still not indestructible.

Consider this scenario, much like yours.

You remove the original ROMs, using a soldering iron and de-solder braid. You linger a while because the solder isn't behaving and wicking up. You stay much longer than you would to solder a component in place (Maybe 10s max) and keep trying to desolder. Eventually it wicks up, but you've been there for 20-30 seconds, taking it beyond the specs it can be heated for, and causing the initial damage.

You remove the components, and solder in something new, so another maybe 10s of heat is pumped into each pad. This adds to the damage already done when removing the component.

What happens is that with each cycle the copper expands and shrinks as it heats and cools, and that can lead to cracking in the via and the pad lifting. Additionally, if you use a very hot iron, it amplifies the problem as it makes the copper hotter, causing more expansion and therefore more contraction. Consider too if you spray IPA on to clean off flux residue whilst the copper is still hot, this can cause sudden cooling and thermal shock to the copper.

Now iron temperatures are something that you need to figure out depending on each scenario, because it depends on the type of solder and flux, the thermal mass of your soldering iron, the thermal mass of the item being soldered, ambient temperature etc etc. It's not a simple thing.

For my own setup, and taking into account the above factors, I run my iron at 370C. It's more than enough for the particular items I tend to be soldering (30+ year old computers!) and allows me to solder most things in 5-10 seconds, even where there is a large thermal mass like a ground plane.

I usually set my desolder gun to the same 370C, and will only increase the temp in the case of particularly stubborn ground planes. I can normally desolder a pad in no more than 10s at this temp.

These temps may not work with your particular setup, so don't take that as read, practise on dead PCBs to get a feel for what works for you.

The big problem with a desolder gun is technique. I've made many mistakes myself since I got one, but the biggest is applying too much pressure to the PCB. This causes pressure damage to the pads/vias when you add in the heat factor.

Just use enough pressure to hold the gun to the solder whilst it melts, almost as if you were soldering, and once you feel it's ready, suck the solder out. If you have tried to desolder and some solder is left, move on to another pin rather than heating the same on again and again, and of you can, move to another pin as far away from the troublesome one as possible. Add more solder to it if you need, and use flux, it will help clear the solder out.

Having the right tools is always a good place to start, but they can cause as much damage in inexperienced hands. Trust me, I have made a lot of mistakes over the years, even with the right kit.

Patience is the only way forward here, but I'd suggest the additional pads that lifted were already damaged after the first time you desoldered the ROMs and the additional soldering/desoldering just added to the damage to cement its failure later.

Getting back to your specific problem, I'd still say that you have a pad/via/track problem. Since you only worked on the ROMs, it has to be that area. The diag cart almost certainly won't work on screen, but you may see something if you use a serial cable as derkom suggested.

There is a 'However' though. In some cases, like my STE with the single broken pin on the TOS switcher, the diag cart did nothing at all, so be prepared to have no answers from it. There is more detail on the diag cart here: https://www.exxosforum.co.uk/forum/viewt ... =91&t=3143

My best suggestion at this point is stop. Your machine is in need of a serious repair now by the sounds of things, and continuing further may cause irreparable damage, and given this machine has significant sentimental value for you, I'd suggest that probably don't want to be in that situation.

I know how much you want to repair it, but there is a fine line between fixing it and doing more harm.

[exxos]

I agree with what @rubber_jonnie said above.

Though one slight " issue" I just un-forgot , is the diagnostic cart actually needs the first few lines of ROM code to actually boot the cartridge in the first place. So even with a serial link you may still likely have a dead machine. But at this point I cannot really say either way.

I have half a feeling that EMUTOS may have a option to boot from the cartridge port ? So maybe worth looking into that and see if your machine will actually boot that way. But I have no idea if that actually requires working motherboard ROMs either. Might be worth pinging the EMUTOS team and asking them about it.

[sporniket]

Though one slight " issue" I just un-forgot , is the diagnostic cart actually needs the first few lines of ROM code to actually boot the cartridge in the first place.

It seems very likely to me, it jumps into the ROM first, then the boot sequence checks for the clue that an executable rom is present on the cartridge port, then jump to said ROM.

This page http://atari.8bitchip.info/cartST.html contains a description of the magic values/bitfields used to tell the TOS at which point in the HW initialisation it can jump to the cartridge ROM.

[exxos]

It seems very likely to me, it jumps into the ROM first, then the boot sequence checks for the clue that an executable rom is present on the cartridge port, then jump to said ROM.

This was posted by @troed a while ago

Code: Select all

CODE: SELECT ALL


_main:
/* disable the interrupts */
/* this is useless on cold boot, */
/* but it matters for software reset */
        move    #$2700,sr

        reset                  /* reset all hardware */

/*
 * Check for diagnostic cartridge
 */
        cmp.l   #$fa52235f,cart_base    /* magic - is cartridge present? */
        bne.s   nodiag                  /* no -> go on */
        lea     nodiag(pc),a6           /* save return address */
        jmp     cart_base+4             /* execute diagnostig cartridge */
nodiag:

        lea       memcret(pc),a6
        bra       memchk
memcret:
        bne.s      noreset
        move.b    (memctrl).w,(memconf).w  /* initialize memory controller */

resetvec:
/*
 * Check, if we should jump to a reset vector after warmboot
 * The high byte of resvector must be zero,
 * it must be even, and cannot be entirely zero.
 */

[rubber_jonnie]

I agree with @exxos and @sporniket above.

If it can't access the ROMs, it won't start the diag cart. This ties in with my experiences of ROM problems and not being able to use the diag cart, so whilst you should try, be prepared for it to be a bust.

[BrettRogersUK]

Yeah I think I'm gonna wait for @Steve to become free and send the board to him to look at it.

Thank you to everyone who has tried to help me. It was my own stupid fault in the first place I should have left well alone.

[BrettRogersUK]

I agree with @exxos and @sporniket above.

If it can't access the ROMs, it won't start the diag cart. This ties in with my experiences of ROM problems and not being able to use the diag cart, so whilst you should try, be prepared for it to be a bust.

Yeah as expected, no output on the serial. I don't know why it can't access the ROMS. All the data and address lines are all present with no breaks in them. But something seems to be happening to stop the machine from reading the roms.

Time for someone with a bit more experience than myself to have a look at this I think.

Thanks,

Brett