Need help with initial faultfinding on 1040STe

Troubles with your machine? Just want to speak about the latest improvements? This is the place!

Moderators: Mug UK, Zorro 2, spiny, Greenious, Moderator Team

siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Need help with initial faultfinding on 1040STe

Post by siriushardware »

This is a machine I've had mothballed in the loft for about 18 years since I acquired it from someone at work then (I actually swapped a modest PC for it). As my orginal STFM workhorse was still alive and well at the time I stored the STe and it remained stored until last year when I finally started to have problems with my STFM and its hard drive (those problems were described in other threads).

When I obtained an UltraSatan recently I found that it didn't work with the old STFM warhorse, but I had half expected that because it had ceased to work with my original HDD as well and I suspected that some element of the STFM's DMA port had failed.

So I got out the 1040 STe which I already knew had a 'good' DMA IC, powered it up and found that it worked nicely with the UltraSatan.

Unfortunately, the STe only lasted half an hour.

I was watching it running one of the demos from the UltraSatan SD card when the screen suddenly froze (with the graphics still present on the screen) and the audio output started to produce a sustained bbbbbbrrrrriiiiiiinnnnnnngggg sound, like the CTRL-G bell being repeated very rapidly - (it had been playing a music demo at the moment of failure).

After a cursory inspection I can say that the +5V and +12V supplies are good (on a scope as well as a meter). I happened to have a spare PLCC 68000 handy so I swapped that in - no change, so I put the original chip back in. The reset chain is OK, to the extent that I see the reset signal to various ICs including the processor changing state when I press in / release the reset button. The CPU also has its 8Mhz clock.

For me one significant clue would seem to be the fact that the content of the screen memory continued to be rendered, uncorrupted, onto the monitor (which is a 1986 vintage RGB CRT monitor - there's no frame store built into it) even after the moment of failure. I would imagine quite a lot of the hardware, memory, clocks, etc would still have to be working in order for that to be happening. At a minimum, I guess the screen RAM area of the RAM and the Shifter are probably OK.

When it happened I thought the STe had just crashed and I was very surprised when it failed to come up again after a reset or power down.

I don't have any spares apart from the CPU, so chip swapping of the MMU, GLUE, Shifter, etc is not going to be my first line of attack - can anyone with good knowledge of these suggest a few things to check to try to narrow things down a bit? I have a scope, logic probe, multimeter, the usual general purpose test gear. But almost nothing in the way of STe spares.
User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: Need help with initial faultfinding on 1040STe

Post by exxos »

Have you run a RAM test program on it and left it for a couple hours to see if the ST is OK before you start changing random chips...

If it will not power up at all now.. Then its a probably a bad PSU.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

exxos wrote:Have you run a RAM test program on it and left it for a couple hours to see if the ST is OK before you start changing random chips.
As explained in the first post, I don't have any chips to randomly change - apart from the CPU which was worth a try since it was socketed and took about 5 seconds.

No, it's not the PSU. Rock steady +5V and +12V even under load conditions.

It doesn't run AT ALL, so I can't run any kind of diagnostic software.

The power LED at keyboard bottom-left lights, but the machine doesn't get to the desktop. The FD drive light on the keyboard top-right does not illuminate, the drive does not activate.

There's something pretty fundamentally wrong.
User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: Need help with initial faultfinding on 1040STe

Post by exxos »

Leave it over night, trying again in 12-24 hours time, I bet it will boot again. Try it.. let me know.. Don't turn it on again else you will have to start your 12 hour wait again...
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

exxos wrote:Leave it over night, trying again in 12-24 hours time, I bet it will boot again. Try it.. let me know.. Don't turn it on again else you will have to start your 12 hour wait again...
Not a bad thought, but already done because it took me a couple of days to lay my hands on the other 68000.

So two days after it originally failed I powered it up again as-was and it still wasn't working, so I tried the other CPU, made no difference, put the old one back in and spent a couple of hours looking around with a scope - but I couldn't make much headway, mainly because I have never scoped the various signals on a working STe so it was difficult for me to know whether what I was seeing was normal / correct or not.

Then I left it alone for a while while I tried to fix the colour video sync problem on the STFM, as we discussed in another thread.

With that now fixed, I've come back to the STe.
User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: Need help with initial faultfinding on 1040STe

Post by exxos »

How much RAM do you have in there ?
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

exxos wrote:How much RAM do you have in there ?
It's a 1040STe so... 1MB as far as I ever knew.

When I look at it, however, there are two pairs of two different looking memory modules in the slots. (All four slots are populated). What's the memory configuration in an out-of-the-box 1040STe? Two 512K sticks + two empty slots? Or something else?

Just taking a look at the two different memory module types - both have what look like Atari part numbers

First type is C301962-001 - Two Goldstar GM71256 Ics, two Siemens HYB513256-AJ-10B and some blobbable links of which BR1, BR4, BR5, BR7 BR10, BR11 are made.

Second type is C398125-001 - Eight Siemens HYB41256-10.
User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: Need help with initial faultfinding on 1040STe

Post by exxos »

Try taking 2 out and see if that helps.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

exxos wrote:Try taking 2 out and see if that helps.
Just did that, no difference. I tried both pairs individually in both pairs of slots. No change.

With the STFM connected to the same RGB monitor, what happens after power-on or reset is that a brilliant white, blank screen appears and rolls vertically for about 0.3-0.5 seconds because the video circuit initially runs at some incompatible frame rate (60Hz or 70Hz), neither of which my colour monitor can lock on to. Then, it settles down to a steady solid white screen at 50Hz for a few seconds, then the desktop appears.

With this faulty STe, it never gets beyond the wrong frame rate / picture rolling stage. I turn it on and the (blank white) screen just rolls and rolls and rolls indefinitely because the video hardware is not being initialised. In fact, I don't think any part of the system is being initialised, it's just not getting that far.
User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: Need help with initial faultfinding on 1040STe

Post by exxos »

I've had the exact same issues as you in the past, normally after fitting more RAM to the STE. The STE would work for half hour, then never work again. I put it down to the PSU as I never found anything wrong with anything else. Though at the time I thought the simms were bad and killed the STEs. Until one day the STE would not boot, just white screen. I left it a couple days, then it was working again. I never got to put the idea to the test, though if its not the PSU, then something will have died, but as to what I never found out.

You can't measure the PSU 5V on a scope unless you know exactly what to look for. When the ST turns on, big power surge as ST pulls power during init. Voltages can dip to 3V on the 5V line, then game over. If the ST manages to get past init, the next thing will be motor on, on the floppy drive, surge drops PSU voltage for a few ns and ST crashes anyway. You need fast storage scope and monitor the very first second of power on. If you power on then measure the 5V line on your scope, too late, you missed the voltage drop.

What happens if you unplug the floppy drive ?
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

exxos wrote:I've had the exact same issues as you in the past, normally after fitting more RAM to the STE. The STE would work for half hour, then never work again. I put it down to the PSU as I never found anything wrong with anything else. Though at the time I thought the simms were bad and killed the STEs. Until one day the STE would not boot, just white screen. I left it a couple days, then it was working again. I never got to put the idea to the test, though if its not the PSU, then something will have died, but as to what I never found out.

You can't measure the PSU 5V on a scope unless you know exactly what to look for. When the ST turns on, big power surge as ST pulls power during init. Voltages can dip to 3V on the 5V line, then game over. If the ST manages to get past init, the next thing will be motor on, on the floppy drive, surge drops PSU voltage for a few ns and ST crashes anyway. You need fast storage scope and monitor the very first second of power on. If you power on then measure the 5V line on your scope, too late, you missed the voltage drop.

What happens if you unplug the floppy drive ?
To lessen the PSU load during initial start-up, yes? I have just tried it, it makes no difference.

I've also addressed this differently by trying my spare ST PSU unit, which also makes no difference to this STe fault but is able to run the (now fully working) STFM OK - though I can imagine, without looking it up, that the STe probably has higher power requirements than the STFM (Although, my STFM actually has more than twice the RAM that the STe has).

My scope actually -is- a storage scope, so I can see that there's no significant drop on 5V or 12V during the immediate start-up phase.

As all the memory modules have Atari part numbers on them, I think there's a reasonably good chance that they are the original Atari fitted modules. I don't think anyone has ever upgraded this STe's memory.

It so happens that my 520STFM has a memory add-on which allows up to 4MB of SIMM memory to be put in the machine, although mine only has 2MB in it, and the original onboard 512 remains enabled for a total of 2.5MB. I could always put the two pairs of simms from the STE in that machine one pair at a time to check them, but to be honest I think just trying one pair at a time in the STe is enough to have shown that the memory is not the problem.
User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: Need help with initial faultfinding on 1040STe

Post by exxos »

The only time I saw the STE die was many years ago, that was just upgrading RAM, it killed a lot of machines I had back then. I never found out why, but just assumed it was the PSU. I can't remember if I tried other PSU's or not back then to see if the STE failed to boot or not. I remember the address and databus was still active, so it was like the ST was still operating, but not actually booting. It looks like you have a dead STE there..

Maybe someone else has some suggestions, but I never really spend time with in-depth faults as its normally easier just to get a new motherboard. Though if it comes to the point you giving up on it, you could send it to me to look into it. Most of what I do with the CPU boosters results in really screwy problems like that now so more used to it these days.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

exxos wrote:The only time I saw the STE die was many years ago, that was just upgrading RAM, it killed a lot of machines I had back then. I never found out why, but just assumed it was the PSU. I can't remember if I tried other PSU's or not back then to see if the STE failed to boot or not. I remember the address and databus was still active, so it was like the ST was still operating, but not actually booting. It looks like you have a dead STE there..

Maybe someone else has some suggestions, but I never really spend time with in-depth faults as its normally easier just to get a new motherboard. Though if it comes to the point you giving up on it, you could send it to me to look into it. Most of what I do with the CPU boosters results in really screwy problems like that now so more used to it these days.
Ah, I love difficult faults. Usually. :roll:

I work as an electronics technician, although not specifically on computers, so I have the necessary general electronics knowledge and experience (which is one half of what's needed) but not in-depth knowledge of this particular equipment and a stock of spare parts (which is the other half of what's ideally needed). I basically just can't see something which could potentially be fixed allowed to go unfixed, so I'm going to continue to have a go at it. If nothing else, I'll eventually learn a lot that I didn't know before. Now that the STFM is working with the UltraSatan, the pressure to fix the STe is off (which is probably a good thing). Previously, the STe was the only machine I had which worked with the UltraSatan.

When we were talking about the STFM fault in another recent thread, simbo2 indicated that he might have a few spare ICs which he'd be willing to loan out for diagnostic purposes: Like you, what I've found is normal (?) looking activity where I expect to find it, but the computer just isn't working. I'll go as far as I can with what I've got, but in the end I may just have to chip swap, especially if one of the big custom ICs comes under suspicion (if indeed they have not already). I may eventually take you up on your kind offer to have a look at it (you at least probably have some spare ICs to try) but I'll try to have a more in-depth look at it first.
User avatar
troed
Atari God
Atari God
Posts: 1460
Joined: Mon Apr 30, 2012 6:20 pm
Location: Sweden

Re: Need help with initial faultfinding on 1040STe

Post by troed »

siriushardware wrote:For me one significant clue would seem to be the fact that the content of the screen memory continued to be rendered, uncorrupted, onto the monitor (which is a 1986 vintage RGB CRT monitor - there's no frame store built into it) even after the moment of failure. I would imagine quite a lot of the hardware, memory, clocks, etc would still have to be working in order for that to be happening. At a minimum, I guess the screen RAM area of the RAM and the Shifter are probably OK.

I don't have any spares apart from the CPU, so chip swapping of the MMU, GLUE, Shifter, etc is not going to be my first line of attack - can anyone with good knowledge of these suggest a few things to check to try to narrow things down a bit? I have a scope, logic probe, multimeter, the usual general purpose test gear. But almost nothing in the way of STe spares.
On the STE there's no separate GLUE, both it and the MMU are part of the GST MCU. At least the "GLUE" parts of it should be fully working though since you had a steady image (sync signals ok).

From your later description it sounds like TOS isn't initializing the chips (rolling screen) - so maybe TOS isn't working as it should. (If you named the demo it would help - some demos don't use TOS routines at all but many do).

/Troed
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

Thanks for the info re: GLUE and MMU both in one chip in the STe - I hadn't even noticed that yet.

About the demo - (Edited - it was the '20 Years' demo]. Your theory that the demo might have been using TOS routines and failed because TOS failed is an interesting one.

If there's some doubt over whether the TOS ROMs themselves are still working I have a suitable device programmer to read them with, so I could compare their content with the appropriate ROM image - That would rule out a problem with the actual ROM code. However, my feeling is that there's something fundamentally broken in the machine's subsystem which prevents the CPU from accessing the TOS ROMS in the normal way, and therefore prevents any of the operations which require TOS to be working (pretty much everything, in other words).

The same device programmer has a 74xx / 40xx logic device chip tester feature, so I could laboriously remove all the non-specialised logic ICs and test them independently of the STe - I might get lucky and find that it is one of those which has failed.

Here's one idea - in theory I could write a piece of 68000 code which does something very simple like toggling an I/O port pin hi/lo/hi/lo or continually transmitting a known character out of the RS232 port, and then program that code into two eproms and plug them into the TOS ROM sockets. If just that simple piece of code could be made to run, it would prove that the 68000 was able to fetch and execute instructions and communicate over the system buses, in which case the prospects for diagnosis / a fix would suddenly look a lot better.

There's also the old trick of filling the OS ROMS with NOP instructions which, on a simple 8-bit system, would generate a nice smooth binary count on the address lines and show up any problems on any of them - however, I'm not sure it would be so simple to do something like that on the STe.
Last edited by siriushardware on Wed Feb 04, 2015 10:31 pm, edited 1 time in total.
User avatar
Arne
Atari Super Hero
Atari Super Hero
Posts: 659
Joined: Thu Nov 01, 2007 10:01 am

Re: Need help with initial faultfinding on 1040STe

Post by Arne »

You ever tried the STE diagnosis kit on your STE? I sent a couple to the UK. Maybe someone can lend you his kit. Get the service manual for the STE. At least the MegaST service manual does have a chapter "Troubleshooting a dead unit." That step by step guide may help you if you've got access to a scope (DSO preferred).
There's also the old trick of filling the OS ROMS with NOP instructions which, on a simple 8-bit system, would generate a nice smooth binary count on the address lines (..)
Could work on a 68K also, if SP and PC are initialised. NOP is IIRC $4E71 and as you don't have A0 on a 68K it should count as expected.
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

Arne wrote:You ever tried the STE diagnosis kit on your STE? I sent a couple to the UK. Maybe someone can lend you his kit. Get the service manual for the STE. At least the MegaST service manual does have a chapter "Troubleshooting a dead unit." That step by step guide may help you if you've got access to a scope (DSO preferred).
There's also the old trick of filling the OS ROMS with NOP instructions which, on a simple 8-bit system, would generate a nice smooth binary count on the address lines (..)
Could work on a 68K also, if SP and PC are initialised. NOP is IIRC $4E71 and as you don't have A0 on a 68K it should count as expected.
I would love to have either an original or remanufactured test cartridge / loopback cable kit especially if it worked with the STFM as well since I have one example of each, but it's difficult to justify the cost of such a kit for purely personal use on two machines. You're right, it would be good to be able to borrow or hire such a kit. Hmmm... this is probably a stupid idea, but what if I were to programme the code for the STe test cartridge into a couple of EPROMS and replace the TOS eproms with them temporarily? Would the cartridge code run? I would be surprised if the test cartridge uses any routines in TOS because it might be TOS itself which is broken in a broken machine - so the question is, does a cartridge, when inserted, replace the TOS ROMs and run in the same address space as TOS normally does?

Even if this particular idea didn't work, it would presumably be possible to write a poor man's version of the test cartridge to run from a pair of ROMs placed where the TOS ROMs normally go - that's basically an expansion of the 'simple hardware test code' idea I was talking about previously - although of course doing so would require a very good understanding of the ST / STe hardware. The 'test ROMs' would need to contain everything within themselves (keyboard handling routines, display routines complete with defined character set, text input / output via an RS232 console as well, in case the display is broken).

Anyway, I haven't run into a brick wall yet - there are various things I can still try.

With regard to the NOP trick for testing address lines thanks for the confirmation that it might work. Another benefit from that trick, again with simple systems, is that memory mapped address decoding can also be checked, in so far as the address decoder output associated with a particular device should generate a pulse or a train of pulses as the processor steps through all the addresses in the device's address range - however, I think the MMU (GSTMCU) in the STe might be a little bit too clever to co-operate with that trick. (It probably needs to see an attempt to write to or read from a device before anything happens?)

I've made a start today by trying to draw out the block diagrams of (a) the reset chain and (b) the clock chain in the STe. This is proving to be something of a headache because the STe schematics generally available online are such poor copies that around half of the IC pin numbers are unreadable, so I'm going to have to go checking with a meter to find out which pins really are involved.
SteveBagley
Captain Atari
Captain Atari
Posts: 188
Joined: Mon Jan 21, 2013 9:31 am

Re: Need help with initial faultfinding on 1040STe

Post by SteveBagley »

I used a cheap £10 logic analyser from Amazon on an ST (not E) earlier to find out what was going on as I could see when various things happened and cross-reference it to the TOS source. Just watching DTACK, AS, A23, BERR and a few other bits was enough, see my thread elsewhere, might be harder with an STe to tie onto the processor.

16/32 have ST(e) diagnostic cartridges on sale for £20 btw, I picked one up last week, but if you want to try a the NOP trick, opcode 0 is effectively NOP (its OR #0,d0 or similar) -- but that will change the value put in the initial PC so it starts executing RAM at 0x0, not the ROM (at 0xE00030).

Steve
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

@SteveBagley: That's a lot cheaper than I expected, for a diagnostic cartridge, probably no more than the price of one wrongly guessed 'bad' big chip, and surely less than the price of a whole STe motherboard. I'll think carefully about that. I've got a pretty good understanding of typical 8-bit microprocessor system architecture but I'm still really weak in my understanding of what many of the 68000 control lines (BERR, etc) do, and when. I'm going to have to do some reading up before a logic analyser - one of the few tools I do not have - will be any use to me.

As regards making connections to the processor, luckily, I have an adaptor PCB which was originally used to mount an ATOnce PC emulator on top of an STe processor. The ATOnce itself was designed to plug into a 68000 (DIP version) socket, and then the adaptor board I have converted the DIP pinout of the ATOnce into a plug-in PLCC connector which sits on top of the PLCC socket with the PLCC 68000 still in situ - the upshot of all that is that, If I plug this adaptor onto the top of the STe's processor it makes all the address / data / control bus connections available on an easy-to-access DIP header.

The only thing I can say at the moment is that the physical reset switch is definitely working, and the reset signal is reaching some of the places I checked - for example, the master reset from the 555 is getting through to the two 7406 open-collector buffers which split it into XRESET and XHALT. Both of those signals are arriving on the relevant pins of the 68000. I see them change state when I press and release the reset switch.

From general experience I would have thought that if the physical switch failed it would do so gradually and become progressively more troublesome until it stopped working, or if it did fail suddenly it would do so when it was actually being pressed or released - I wasn't touching the machine at all when it suddenly failed before my very eyes.

As I mentioned earlier, I haven't yet followed the reset signals to every place they should go to because a lot of the chip pin numbers are unreadable on the diagrams which are available, so it's going to be slow work following them all and annotating the diagram with the actual pin numbers - same with the clocks, which I need to follow from the master clock all the way through to its various derivatives and their final destinations.

I've eliminated the power supply by (a) measuring the voltages on-load (b) scoping the supplies, including capturing the first few mS / first few seconds with a storage scope: And (c) by putting in another PSU which runs my STFM OK - so I am reasonably sure it's not the power supply.

Re your question about the drive light: It never comes on.
Last edited by siriushardware on Tue Feb 10, 2015 8:03 pm, edited 1 time in total.
User avatar
Arne
Atari Super Hero
Atari Super Hero
Posts: 659
Joined: Thu Nov 01, 2007 10:01 am

Re: Need help with initial faultfinding on 1040STe

Post by Arne »

Would the cartridge code run?
Of course not.
(..)so the question is, does a cartridge, when inserted, replace the TOS ROMs and run in the same address space as TOS normally does?
You gave the answer yourself.
Both of those signals are arriving on the relevant pins of the 68000. I see them change state when I press and release the reset switch.
What state do they switch to? If /XBERR, /XHALT and /RESET are high then it should continue executing TOS code.
Try getting a PDF copy of "ST Internals" It's got a commented TOS (1.00 though) listing of how TOS boots. It should be pretty similar for all TOS versions on 68000. Then you see that TOS pretty early does a jump into the diagnosis cartridge if present. If Atari had designed the cartridge port for R/W and put some SRAM onto the cartridge then it should be usuable even without any onboard RAM I think. But they missed that chance... :evil:
Just replacing ICs isn't very straight-forward IMO.
User avatar
1st1
Atari Super Hero
Atari Super Hero
Posts: 885
Joined: Mon May 07, 2012 11:48 am

Re: Need help with initial faultfinding on 1040STe

Post by 1st1 »

You also should look with your oscilloscope on the chipselect and clock signals of all chips. If you can't see chipselect for example at the tos roms (CE, OE) then something wrong with the glue logic of the ste mcu. If some of the chips to not get their clock signals, then something wrong with clock generator and clock dividing what is also done by the glue logic and a 74ls74 flipflop. All starts with a 32 Mbz clock which is divided down to 16 Mhz and 8 Mhz and some more. Check also the BERR, HALT, BG, BR, BGACK and so on signals on the CPU bus if they do something after reset. Some are controlled by the 68000, others are controlled by chipset. See service manual of the STE which you can download from devdocs.atariforge.org. For the CPU bus/handshake signals refer to a 68000 datasheet.
Power without the Price. It's not a bug. It's a feature. _/|\_ATARI

1040STFM in PC-Tower (PAK68/2, OvrScn, 4 MB, 1GB SCSI, CD-ROM...) * 3x Falcon 030 * 3x TT030 * many 260 /520/1040ST(F)(M)(+) * 520/1040STE * many Mega ST * 2x Mega STE * Stacy * STBook * 2x SLM605 * 3x SLM804 * SMM804 * SH 204/205 * Megafile 30/44/60 * SF314 * SF354 * 5x Pofo * PC3 * ...
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

Arne wrote:
(..)so the question is, does a cartridge, when inserted, replace the TOS ROMs and run in the same address space as TOS normally does?
You gave the answer yourself.
I didn't actually know the answer, which is why I asked the question.

As far as I knew the diagnostic cartridge code ought to be fully self contained (independent of any TOS routines) and might have been fully relocatable, able to execute from / at any address. But I'm happy to accept what you say without getting into any deep discussion about it. If I'm going to run any test code, then, it will have to be code that I write myself specifically to run at the address of the TOS ROMs. However, I'm nowhere near that stage yet.
Arne wrote: What state do they switch to? If /XBERR, /XHALT and /RESET are high then it should continue executing TOS code.
Try getting a PDF copy of "ST Internals" It's got a commented TOS (1.00 though) listing of how TOS boots.
As I recall, pressing / holding the reset switch was taking the _Reset and _Halt pins on the 68000 low: When the reset switch was released, they went high. I didn't get to checking _BERR. I'm going to have another look at it later tonight.

I've always owned a paper copy of 'Atari ST Internals' (Data Becker) and I still have it now, but it deals (as you say) with the earliest versions of the ST(F)(M), not with the STe which didn't exist when it was printed. At this stage, the differences between it and the STe might confuse me more than the similarities will help. However, it does have useful descriptions of 68000 specific control signals, the functions of the various special ICs, their registers and so on, so it should be helpful even if the functions of those ICs are spread around a different set of ICs in the STe.
Arne wrote: Just replacing ICs isn't very straight-forward IMO.
For those ICs which are in sockets, it would be, if I had any ICs to replace them with :-).

I have already tried swapping the PLCC 68000 for an identical chip that I happened to have - no difference, so I have put the original chip back in. I have access to good tools for desoldering of both conventional and SMD ICs but I really prefer to narrow a fault down to the smallest possible area by 'scientific' means before pulling soldered chips to try to fix a fault.
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

1st1 wrote:You also should look with your oscilloscope on the chipselect and clock signals of all chips. If you can't see chipselect for example at the tos roms (CE, OE) then something wrong with the glue logic of the ste mcu. If some of the chips to not get their clock signals, then something wrong with clock generator and clock dividing what is also done by the glue logic and a 74ls74 flipflop. All starts with a 32 Mbz clock which is divided down to 16 Mhz and 8 Mhz and some more. Check also the BERR, HALT, BG, BR, BGACK and so on signals on the CPU bus if they do something after reset. Some are controlled by the 68000, others are controlled by chipset. See service manual of the STE which you can download from devdocs.atariforge.org. For the CPU bus/handshake signals refer to a 68000 datasheet.
Thanks for the useful pointers - One of the reasons I'm finding this fairly hard going is that all the online versions of the schematics I've found so far are pretty bad copies, with many of the pin numbers not legible. For example, the _CE pins of the TOS ROMs are shown connected together and they head off left to towards a named control line which comes in from another of the 6 sheets which make up the circuit diagram. Unfortunately the name of this control line is utterly unreadable. By a process of elimination I think it may come from ROM2 on the GSTMCU, since ROM3 and ROM4 go to the cartridge port and ROM0, ROM1, ROM5, ROM6 on the GSTMCU do not appear to be connected to anything.

I'll take a look at the service manual you've referenced and see if those diagrams are any more readable.

I'm going to have another look at the STe later on tonight. Maybe It'll be fixed by the time I come back here: More likely, I will just have more questions to ask. :-)
siriushardware
Atari Super Hero
Atari Super Hero
Posts: 551
Joined: Thu Aug 21, 2014 7:55 pm
Location: UK

Re: Need help with initial faultfinding on 1040STe

Post by siriushardware »

Duh... would you believe it, after several sessions spending an hour or two trying to work out why this had died, I settled down for another go at it and... it started normally!

However, I made it die again by flexing / twisting the PCB gently. Looks like it may be an intermittent connection rather than a device fault, which I suppose I should be grateful for. (Yes, there's a small chance that one of the devices may be physically intermittent - no, I don't really want to think about that for the moment!)

The processor socket seems to offer very little resistance to the chip being removed / inserted, so I am going to replace that first and see if that makes the machine stable even when I stress the PCB. At around 25+ years old, it seems the socket contacts have lost their 'spring'. The processor gets quite warm so it will have done a lot of expanding and contracting over the years.

At work, we used 44-pin PLCC devices fitted in surface-mounted PLCC sockets in a particular range of products. They caused so many problems with intermittent / unreliable connections to the chip pins that we now remove the sockets and solder the ICs directly to the PCB whenever those units are returned to us for any reason. You can do that if the original sockets are SMD type because the pins on the bottom of the SMD sockets have the same layout / profile as the pins on the PLCC chips. On the STe, through-hole sockets were used so this isn't an option and I'll have to replace the socket. Fortunately, they still seem widely available.
Dal
Administrator
Administrator
Posts: 4202
Joined: Tue Jan 18, 2011 12:31 am
Location: Cheltenham, UK
Contact:

Re: Need help with initial faultfinding on 1040STe

Post by Dal »

I had exactly this with an STM - which also had a number of other faults before it worked again. It's worth getting retainer clips for the socket to keep the bugger clamped around the chip too.
STE: Desktopper case, IDE interface, UltraSatan (8GB + 512Mb) + HXC floppy emulator. Plus some STE's/STFM's
Post Reply

Return to “Hardware”