exxos's Atari, Amiga & retro forum

Hiya Atari Gurus.

This project started as a use for my spare PAL STFM that was only blessed with TOS1.02 and 512k of ram so was basically not good for much. I already own a 1mb PAL STFM with a GOTEK drive and then only this year was lucky enough to finally score an STE which I upgraded to 4mb and I use with my Ultrasatan.

I had the idea that I could upgrade this SACRIFICIAL 3rd STFM to 4MB of ram with Exxos new ram expansion board for my C070859-001 Rev2.0. I got further excited when I realised that this board would be a candidate for his 16mhz Accelerator board and then of course his TOS board that would let me run TOS 2.06. I also saw an empty Blitter socket so thought.... why not solder in one of his crazy blitter sockets too. Ha Ha I was going CRAZY in this attempt to create a pretty special STFM.

But OH DEAR i pushed for simply TOO MUCH......

I own a PC shop in Australia and have desoldering equipment so first job was to remove the 68000 cpu and fit one of EXXOS sockets in preparation for the accelerator to come later. Logically I remounted the original 68000 to confirm the board successfully started and operated okay with the new socket.

I then decided to desolder using my desoldering station the original 512K of ram chips. This was fairly massive but I enjoyed it and used it as a learning experience. I literally removed all 16 chips and was lucky to only break 2 legs. I then soldered in the new L shaped 2019 4mb board and tested the board with just this mod and up came 1024K. I had teething problems when I tried to jumper 4mb of ram but eventually solved that by shortening the MMU wire to pin66 and then fitted a switch to move between 1mb and 4mb. I ran quite a few memtests successfully in both configurations b4 I thought..... okay time to solder in the crazy blitter socket. I chose NOT to fit the blitter chip and desolder W3 and W4 as that may just muddy the waters. Again I tested the board and everything seemed to operate okay as far as I could test on my monochrome SM124 monitor.

All my problems seemed to really start when I fitted the v2.3 accelerator.

The board posted straight away with the matching booster TOS 1.04/2.06 board in BOOSTER mode with the patch wire soldered into the accelerator. I excitedly ran my sysinfo and noticed it was only running at 7mhz or so it said. Then 30 secs later the machine crashed with multiple bombs and simply would not return to GEM desktop ever again. I tried the tos 2.06 mode and yes it attempted to POST with the black atari logo in top left corner but 4or 8 bombs would always appear b4 any GEM desktop. I tried soldering the 3rd wire from the GAL chip to earth to ensure that it would just boot at 8mhz with the same results.

Naturally I realised at that stage something horrible was going on so I removed all power and thought I must return the original 68000 to its socket but of course I hadnt put the multi TOS board into GLUE mode. I soldered the jumper on it to glue mode and made sure it was in 1.04 mode only and I was returned to the GEM desktop again and I began to think........ I might quit while i am ahead until even this combination crashed with bombs... YIKES.

I didnt know what was going on so like a lot of you would suggest, I took my spare psu out of my GOTEK STFM just in case it was power related. No change just as unstable. Bombs bombs crashes,

The final straw came when I thought......Okay lets go right back to basics and put my TOS 1.02 dual chips which I marked using stickers properly so I couldnt get them confused and placed them in their correct locations. And then YES my GEM desktop returned and I tried to run something on my floppy but then bombs again. I removed power again and tried to push down harder on the original 68000 cpu in its socket assuming I had poor continuity with the CPU and then my odyssey turned into a NIGHTMARE. When I reapplied power I now get simply a BLACKSCREEN on my SM124 and no attempt by the floppy drive to boot. Just the green power lite on the keyboard.

I have subsequently tried to swap psus, put back the accelerator just in case my original 68000 got zapped, swap the TOS board in glue mode instead of the 2 xTOS 1.02 chips. Everything I try produces the black screen and no floppy boot. Thinking this MUST be related to force being applied to the 64pin CPU socket I have resoldered and fluxed all 64 pins and cleaned them up and they look great. My experience on my Amiga suggests that it could be a TOS firmware not starting as a bad kickstart produces the same result. A black screen with no color flashes or diagnostic beeps.

The only other thing I can think of is that I didnt test the 4mb ram board mod for long enough and as soon as I disturbed the board by fitting the accelerator and simultaneously the TOS 1.04/2.06 board I may have somehow destabilized that ram board causing all the instability that followed that change.

But why wont the board POST now???

Can you guys help with all of your experience on Atari ST boards ?

krisfeigl wrote: 24 Jun 2019 14:45 I chose NOT to fit the blitter chip and desolder W3 and W4 as that may just muddy the waters.

Do you mean you DID desolder the jumpers after fitting the socket ?

krisfeigl wrote: 24 Jun 2019 14:45 All my problems seemed to really start when I fitted the v2.3 accelerator.

The board posted straight away with the matching booster TOS 1.04/2.06 board in BOOSTER mode with the patch wire soldered into the accelerator. I excitedly ran my sysinfo and noticed it was only running at 7mhz or so it said.

sysinfo will not report the speed of the booster.. its just not designed to recognise it. You need to run GB6 to see if machine is running boosted or not.

krisfeigl wrote: 24 Jun 2019 14:45 I tried the tos 2.06 mode and yes it attempted to POST with the black atari logo in top left corner but 4or 8 bombs would always appear b4 any GEM desktop.

Bombs after the logo is generally floppy drive not connected.

krisfeigl wrote: 24 Jun 2019 14:45 I removed power again and tried to push down harder on the original 68000 cpu in its socket assuming I had poor continuity with the CPU and then my odyssey turned into a NIGHTMARE. When I reapplied power I now get simply a BLACKSCREEN on my SM124 and no attempt by the floppy drive to boot. Just the green power lite on the keyboard.

Did you do that with the power on ? you may have killed your CPU.. Or need to resolder the CPU socket and even resolder ROM sockets.

Meanwhile, you need to get your MB working with STOCK CPU & ROM's before anything else.. If pushing on the CPU casued the black screen.. then likely you have a bad solder joint or broken something on the board or socket depending how hard you pushed on the CPU.

I've encountered problems of late when testing the beta IDE board. I pulled the board, complete with CPU, from my other test machine, plugged it into another STFM board. Nothing, black screen with a white border. It did the same with the CPU removed.

It's back up and running now, but with the CPU from the other machine, which it wouldn't work with originally. Makes no sense at all

My point? There are differences between all the STFM boards, even those of the same model, and the tolerances, already poor in the late 80s and early 90s are now even poorer, so even the addition of a simple socket is enough to throw things out of whack.

Take the STE. A simple change of chip to a 68HC000 is enough to resolve DMA issues.

My advice is to go back to as close as original as possible, which I think you've tried to do, and then stop, take stock and review all the work you've done (RAM upgrade included) under a large magnifying glass and bright light to make sure there are no solder whiskers, bridges or cold/broken joints, and that you haven't damaged any of the tracks or vias by applying pressure to the CPU.

Once you've done that, have a look at the MMU and GLUE chips, they are often problematic since the tension in the contact springs can weaken over time. I did a post about this a little while back: https://www.exxosforum.co.uk/forum/viewt ... hilit=plcc

Not only that, but the contacts on the chips and the PLCC sockets are not gold plated (Cheapskates!) and so are susceptible to oxidisation, and you can use something like DeOxit, or other similar products to clean and deoxidise the contacts on both.

My advice would be that if you're using spring based sockets, then you can sometimes get a little extra movement in, but once a chip has stopped moving in a turned pin socket, it won't go any further, so you really must stop pushing!!

As for the PSU, I have had PSU problems before, and a re-cap has done wonders, so even though you do have a working spare one, it may not work properly with all boards. If neither is re-capped, I'd suggest you do so.

Good luck!

Thanks for your responses guys.

I never apply much force ever while any power is applied to any motherboard. If I do its only to see if a socket may be flakey or continuity is flakey.

The blitter socket was only soldered in place. The W bridges were left as they were and the socket left unpopulated.

My motherboard features surface mount glue and mmu chips which are unlikely to suffer from all of the PLCC issues that effect the clipped in jobbies unfortunately. I almost wish it was as simple as clipping in and out some PLCC chips to check them for oxidation.

I AM concerned about the PSUs and may plan an order to EXXOS for one of his new RED board super psus in my next order.

I do suspect the 4mb board soldering may still be dodgy in some way and so need to run across all of the joints yet again. I have 2 other 68000 cpus from Amiga 500s I havent tried yet

I have a CD32 fitted with a Terrible Fire 330 68030 board that is also blackscreen ATM but believe it is due to a bad PSU and possibly the 1000uF caps on its main board so I am not having much luck really at the moment.

Pffft Retro Computing. 30 year old equipment. You could just scream sometimes ;-) although I really pushed the envelope with all that I did to this board.

If it wasnt so cost prohibitive I would be tempted to just ship the board to EXXOS form here in OZ to find out what has gone wrong......simply just to know!!!

I will let you know if I get any life out of this thing again. :-)

Just tried 3 different cpus. No change however if I damaged the new cpu socket I would be very surprised.

I also waited 30 seconds on my mono monitor too see if there were any diagnostic beeps or border flashes but still nothing. Tried swapping floppies as I have a spare from the GOTEK STFM again no change. Tried with no keyboard connected. Made sure that the reset switch wasnt stuck or hadnt failed as I had to reglue it back together as the head broke off when i orignally dissassembled the motherboard from the case. It seems to be okay I THINK!! In my life fixing 10,000 PCs over 23 yrs I have had up to 10 pcs which wont turn on as the reset button on the case has failed stuck in and engaged. The board will never start as a consequence so something to look out for!

Gone over the board as much as I could with a magnifying glass and couldnt see anything obvious but I know that it is very hard to see a bad joint. There were certainly no bridges or contamination that was obvious.

I returned the original Tos 1.02 chips and will put the original 68000 back in it and will resolder the ram board again tomorrow and perhaps the TOS rom sockets as there MIGHT be a crack in a joint due to the first force on them in 30 years.

Beyond that I will be really quite stumped. What multimeter checks might be recommended? I dont have an oscilloscope but might consider getting one to try and find out why i am not even getting to the white screen b4 the GEM desktop. Exxos what is the order of the Power On Self Test that occurs on an Atari ST? If the ram board I have added doesnt initialise or present a valid ram bank at startup will this possibly produce a black screen symptom with no audio beep failure or diagnostic?

Oh b4 the board died when trying TOS 2.06 with the booster in place I got a BAD ROM CRC IN CHIP E message with 3 bombs underneath. Was this error an indication of what was to come?

Thanks for any help offered.

TOS 2.06 is meant for STEs, so that's likely why you got that error. It'll likely only work wit the accelerator installed and working. Use the multimeter to buzz out the lines you soldered from the MB to the 4mb adapter from a point on the mainboard away from the riser pin. If that's OK go from the riser pin to a point on the 4mb board away from the riser pin. By doing it in stages you can eliminate the mainboard, the riser pins and the 4mb upgrade.

As a word of advice, only ever upgrade one thing at a time and make sure that everything is working before moving on to the next thing, as it means you only have one thing to troubleshoot at a time.

I would check the RESET pin on the CPU and make sure, when the reset button is pressed, it goes to 0V on the CPU. Then when switch is released, it should go to 5V.

BAD ROM CHIP is likely bad connection on ROM socket somewhere.

When the ST starts up one of the first things it does is set the monitor to 50Hz ( assuming UK TOS) By default its 60Hz. So if the monitor isn't switching Hz then your likely not running ROM at all.

I would suggest changing the ROM sockets for some new ones. They cause all sorts of issues in the STE when people change the ROMs etc. Of course getting STOCK working first is the best idea.

DIAGROM may help diagnose faults. It should allow you to test memory etc even on a apparent dead machine.

Just a brief update.

I resoldered all of the 4mb board joints. No change.

I then as Exxos suggested looked at the TOS Rom sockets. I had nothing to lose so I reheated and applied solder to the 2 x 28pins of both of the ROM sockets. Interestingly No change initially but during ONE of my reset button resets up came a white screen with 2 bombs and then a few seconds later gone again. It was long enough for the floppy to start booting so I DO believe I am a victim of 30 Yr old TOS Sockets causing these problems.

Perhaps the best option is to directly solder in EXXOS multitos board and do away with any sockets. Its still jumpered in GLUE mode and of course it has a socketed modern dual bank rom itself. When I think about it all my problems started with removing those TOS1.02 chips really coinciding with the BOOSTER board of course.

Judging by Exxos comments heaps of people have had bombs, black screens and all sorts of problems with bad TOS ROM sockets in STE and STFMs.

Will update if it comes back to life soon.

Have you metered the ROM lines back to the CPU ?