[Video] Mandatory Fixes: bus resistors on an STFM

[Badwolf]

Morning,

Some Friday STFM for you.

I tackle one of the most obvious stability-enhancing modifications you can make to your ST today: lowering the resistance of the bus pull-ups.

I adopt a addition-only approach. It's perhaps not the best way to tackle this, but maybe the least destructive for those who, like me, suffer from a deficit of desoldering talent.



The mandatory fixes thread can, as ever, be found here: viewtopic.php?f=17&t=1585

Cheers!

BW.

[exxos]

:bravo:

Yeah I tagged them on the back a few boards as well. The 70789 series of boards is the most stable. But some of the other boards even changing to the HC CPU will cause it not to boot at all. Even changing the brand of ROM can cause it not to boot. Even plugging in and the diagnostic cartridge can make or break the machine booting. It really is a total nightmare.

[Badwolf]

Yeah I tagged them on the back a few boards as well. The 70789 series of boards is the most stable. But some of the other boards even changing to the HC CPU will cause it not to boot at all.

Ah shite! You've just reminded me I failed to add an HC PLCC 68k on shop order. Bah. I knew there was something else!

That's for another Sooper Sekrit project that may go nowhere, BTW :ugeek:

Even changing the brand of ROM can cause it not to boot. Even plugging in and the diagnostic cartridge can make or break the machine booting. It really is a total nightmare.

This crusty old board has been pretty solid, yeah, but the last thing I want is to be looking for problems with my stuff that aren't there. I'm going on the basis you've called them mandatory for a reason, so let's do it!

I'm planning to record a video for each change I make that is on the mandatory list and have them all on a play-list. Can be used as refrence for people in the future then.

Cheers,

BW.

[exxos]

This crusty old board has been pretty solid, yeah, but the last thing I want is to be looking for problems with my stuff that aren't there. I'm going on the basis you've called them mandatory for a reason, so let's do it!

Indeed. After having nothing but problems with these boards for 25+ years, after doing the fixes they have pretty much been reduced to nothing. The amount of time I spent back in the day doing ROM upgrades , be it chip or PCB types, and failing just because the brand of the ROM would break a machine. Same happens on the STE as well. You end up chasing faults which are not actually there. You end up concluding the ROM chips are faulty even though they pass verification. And changing them "proves" they are faulty because a different brand works. But in reality it is just pure dumb luck at the end of the day what works and what doesn't.

Years ago I was saying not to use Texas Instruments branded EPROMs because I had nothing but problems with them. The SGS brand always worked. of course back then I did not know what I know now. Ironically my old job we also found Texas Instruments EPROMs would tend not to work on some particular manufacturer of boards. We also concluded to only use SGS brand. Though I wonder if those boards were suffering from similar issues as the STFM do.

I'm planning to record a video for each change I make that is on the mandatory list and have them all on a play-list. Can be used as reference for people in the future then.

Great stuff!

I guess if you could prove something is failing and the fix fixes them it would also be an advantage. But sometimes it is difficult to do "on demand". sometimes it can take several hours to get the "one off crash".

Coincidentally I did ping @agranlund A couple weeks ago ( I don't think he has been online for a few weeks?) about doing a more advanced system "stress test" program for diagnosing pull-up and DMA related issues ( mainly in mind for the STE though).

[Badwolf]

Years ago I was saying not to use Texas Instruments branded EPROMs because I had nothing but problems with them. The SGS brand always worked. of course back then I did not know what I know now. Ironically my old job we also found Texas Instruments EPROMs would tend not to work on some particular manufacturer of boards. We also concluded to only use SGS brand. Though I wonder if those boards were suffering from similar issues as the STFM do.

I reckon I'm seeing something similar with FPUs on DFB1.

I've had reports that a pair of chips, which both run at 32MHz on the TT will fail at 25 and 40MHz respectively on DFB1. I think there may be a control line termination issue too, but more agressive bus pull-ups is possibly an answer.

I guess if you could prove something is failing and the fix fixes them it would also be an advantage. But sometimes it is difficult to do "on demand". sometimes it can take several hours to get the "one off crash".

Yeah, and proving that's the problem is never easy.

DMA might be an easier one to prove, but haven't encountered anything on that front yet.

Cheers,

BW.

[exxos]

I've had reports that a pair of chips, which both run at 32MHz on the TT will fail at 25 and 40MHz respectively on DFB1. I think there may be a control line termination issue too, but more aggressive bus pull-ups is possibly an answer.

Possibly. Such things become difficult to prove. I have had no end of problems with clock lines even if they look good. Slapping something like 33pF on the clock can be work-around. Not sure where the FPU clock is being driven from offhand.. PLDs at least the ones I have used a not very good at pulling up signals by themselves and often need a pull-up resistor as well. Also schmitt buffers I have to use on the STE running 32MHz. directly driven from a OSC itself I found on the H5 prevents the machine from booting without a 470R load resistor on it. The list is basically endless.

[MacTech05]

I have a 1040 STFM Motherboard C103225 REV.2, which is not pictured in the mandatory fixes Bus Resistor Upgrade page.
I am trying to identify the Bus Resistors to upgrade on this mode. I am including photos of my motherboard, if that will help. Of course, as is the case with other 1040 STFM motherboard revisions, there are both the red and black resistor packs installed.

I would appreciate it very much if someone, might help me deterimine which of the pull up resistor packs should be replace with the 2.2K pull up resistor packs.

Thank you.
Randy

C103225_Rev2 MB 1 .jpg

C103225_Rev2 MB 2.jpg

C103225_Rev2 MB 3.jpg

C103225_Rev2 MB 4.jpg

C103225_Rev2 MB 5.jpg

C103225_Rev2 MB 7.jpg

C103225_Rev2 MB 8.jpg

[exxos]

Just meter the address and databus pins from the CPU to random resistor packs until you locate the relivent ones.

[MacTech05]

exxos,

Thank you. I guess I need to get schematic for the Motorola 68000 processor to determine which pins are databus and address bus. I only need to replace the databus resistors and I need to use the 2.2K pullup resistor packs, which I already have. I just ordered more of them so I can replace the databus resistors on my 1040 STE as well.

Randy

[MacTech05]

Does someone have a schematic for the Motorola 68000 processor Address and Data bus pins? I thought I had it figured out, but in toning out via continuity test with my meter, from what I believed to be the databus pin, only 1 resistor pack toned out (continuity), RP2 (black), checked out. No other resistor packs toned out from the pin I used on the Motorola 68000.

I also tested from the pin I believed to be the address bus. Only 1 resistor pack toned out, RP8 (black).

This is the reason I believe I am not performing the test or I am testing from the wrong pins on the 68000. If I am using the correct pins on the 68000, the test is straight forward. I should have been able to successfully tone out (continuity test) at least 4 or 5 relevant resistor packs, possibly 6,

To test resistor packs from the 68000 databus, I used the first pin, the pin at the dot at the notch end of the 68000.

To test resistor packs from the 68000 address bus, I used the last pin on the opposite side and opposite end from the notch on the 68000.

Am I on the correct pins? If I am, then what could I be doing wrong? I am certain that more than 1 resistor pack on the databus and 1 resistor pack on the address bus should be relevant and should be replaced with the 2.2K resistor packs.

If I am on the wrong pins, which pins for the 68000 databus and address bus should I be using to test for the relevant resistor packs to replace.

Thank you.
Randy