• Please review our updated Terms and Rules here

Torch Triple X c1985.

Crashedfiesta said:
Let me know what documentation you have as I have the Caretaker Manual, Triple X Handbook, Inside System V, System V Administrator's Manual, Programming System V, System V User Manual Section 1 and System V User Manual Section 2 to 6, as well as some addendums to the Triple X Handbook. I've also got instructions on fitting and using the Ethernet Chip Set, some information on the OMTI jumper settings, a price list for various Torch upgrades (£700 for a Torch engineer to come and fit a 1MB memory upgrade - bargain) and various other bits and bobs. I should probably start scanning some of it.
Click to expand...
Only documentation I have are the two versions of the service manual that are 'out there' https://bitsavers.org/pdf/torch/triple_x/, and your schematic. I think the Caretaker and the Triple X Handbook would be useful to have scanned.

Good to have another Caretaker version, which confirms it was still being developed after release. Maybe we'll find 1.0 and 1.1 sometime.

Emulation will be in MAME when I'm ready to submit it. Regarding the key disc, hopefully the emulation will allow us to see what it's looking for and comparing with so it can be reverse engineered. We already have a hard drive image at https://bitsavers.org/bits/Torch/ to play with.

Attached a couple of videos of current progress ... the palette version should match your blue screen, and the no palette is running the same but without palette so you can see what's being written to the video RAM.
 

Attachments

  • triplex_nopalette.mp4
    113.4 KB
  • triplex_palette.mp4
    49.8 KB
Awesome!

I scanned the Caretaker manual. They're bound using those wire springs (which can be a pain) and are a slightly odd page size - I scanned it A4 anyway as that's what the auto document feeder said it was. :) I'll scan the Handbook when I get another few free minutes. I've also started to upload a couple of bits to Archive.org so they should start appearing for anyone else who has one of these.
 

Attachments

  • TorchTripleX_Caretaker.pdf
    4.4 MB · Views: 5
That. Is. FREAKIN' AMAZING!!

That is what I'm hoping for at some point on the physical hardware i.e. a demand for the key disk.

So when it starts up I don't have the disk controller plugged in, do you think I should I see the 'no SCSI" message even with the pale blue palette (assuming I can get the self tests to all pass)?

Great work!
 
Crashedfiesta said:
So when it starts up I don't have the disk controller plugged in, do you think I should I see the 'no SCSI" message even with the pale blue palette (assuming I can get the self tests to all pass)?
Click to expand...
Without the OMTI board connected I'd expect to see something like this:
0003.png

To get any further with the floppy emulation will require a dump of the OMTI firmware ROM.
 
Here you go. OMTI firmware in .BIN format. Hope that helps! PXL_20231206_204722319.MP.jpgPXL_20231206_204733422.MP.jpg
 

Attachments

  • OMTI5200.zip
    6.9 KB · Views: 1
Been out prodding the motherboard.

On the 68010 the halt signal is high (correct), the reset signal is high (also correct) and bus error is also high. So this CPU should be running code. But the address and data buses are dead with no activity. A few data lines are stuck high but most are at 0v. One address line (A20) is high but the rest are at 0v.

So something is screwing the main data and address buses. Great.

On the plus side I can see activity at startup on the video RAM which seems to confirm it is moving data around, and the beeps indicate that the service processor and caretaker ROM are also working.

Any ideas on what could be screwing up the buses? 🤔
 
Crashedfiesta said:
Scanned the handbook. It was a pain in the rear to do! Stupid small pages kept jamming the document feeder. But 'tis done. :) It's too big to upload here, so here's a Google Drive link:
Click to expand...
Thanks, not much of interest in there though. It does suggest a working machine should go straight to the Torch logo screen, no mention of what the key disc screen indicates. So presumably my emulation has detected a serial mismatch and wants the key disc to reset things. This gives me something to think about as I don't expect to have working floppy anytime soon.

Do any of the System V manuals contain the installation process and key disc usage?

Crashedfiesta said:
On the 68010 the halt signal is high (correct), the reset signal is high (also correct) and bus error is also high. So this CPU should be running code. But the address and data buses are dead with no activity. A few data lines are stuck high but most are at 0v. One address line (A20) is high but the rest are at 0v.
Click to expand...
The 68010 HALT and RESET signals are active low, at startup they are high so the CPU is not running. It only starts running code when the video RAM has the bootstrap code copied to it via the 6303, which then sets 68010 HALT and RESET low to let it run the code.
 
Pernod said:
Thanks, not much of interest in there though. It does suggest a working machine should go straight to the Torch logo screen, no mention of what the key disc screen indicates. So presumably my emulation has detected a serial mismatch and wants the key disc to reset things. This gives me something to think about as I don't expect to have working floppy anytime soon.

Do any of the System V manuals contain the installation process and key disc usage?


The 68010 HALT and RESET signals are active low, at startup they are high so the CPU is not running. It only starts running code when the video RAM has the bootstrap code copied to it via the 6303, which then sets 68010 HALT and RESET low to let it run the code.
Click to expand...
Hmmm. I thought it was the other way around so the HLT and RST lines were active low i.e. if HLT is low then the system is halted. 🤔

For the key disk I do know that if the key value on the hard disk is different to the value held in the battery backed up RAM then it demands the key disk. That's why that damn battery was integrated into the power supply.

There's no information on what method (if any) encodes/decodes the serial number on the key disk or hard disk. I need to get the Greaseweazle image of the disk into a hex editor and see if it's a simple search for a string... I suspect it won't be that simple, but being a machine from 1985 it might be. 🙂

The caretaker manual (chapters 3 and 6) give some details on the key disk and re-installing SysV. The other manuals I have are basically reference books with just an alphabetical list of SysV commands - nothing very useful. There is surprisingly little about setting up a machine from scratch...
 
New version of the schematics attached with a few updates and error corrections. Usual disclaimers apply.

In some good news I've managed to verify that the video RAM is all working correctly. I've also ordered a new DMA controller chip since that seems to be one of the main drivers of the address and data buses, and since it multiplexes them it clearly has a major hand in both. For the relatively small cost it's worth having it even if it becomes a spare. Once it arrives I'll be able to say that the 68010 is working and the 68450 is working (or replaced). But the 68451 MMU is going to be a pain. It looks pretty rare as it was only really used with the 68010 (according to Wikipedia) and there were none shown at all on eBay, whereas the 68450 was fairly common.. Fingers crossed the MMU is OK.
 

Attachments

  • TorchTripleXSchematics_StickleBack_v0_3.pdf
    2 MB · Views: 3
Crashedfiesta said:
New version of the schematics attached with a few updates and error corrections. Usual disclaimers apply.
Click to expand...
Still a few mislabelled pins and tags, I'll put together a list sometime.

How far do you think you are from actually getting through the boot process? When it comes alive I'd appreciate some photos as my palette handling is still not correct and not obvious what colours I should be seeing, especially on the insert key disc screen.
 
Pernod said:
Still a few mislabelled pins and tags, I'll put together a list sometime.

How far do you think you are from actually getting through the boot process? When it comes alive I'd appreciate some photos as my palette handling is still not correct and not obvious what colours I should be seeing, especially on the insert key disc screen.
Click to expand...
Ooh, a list of errors would be great. Having spent the last couple of months looking at them it's difficult to see the wood for the trees. 🙂

I don't really know how close to booting. I've ordered a new DMA controller because a) it seems to be the main chip that touches both the data and address buses i.e. the PBUS and b) they're surprisingly available and not expensive. The PBUS is dead and it's not the CPU dragging it down. I'm really hoping with a new DMA controller it will spring into life, but if it is actually OK then I'm in the realms of removing each component that the PBUS touches to check which is not going to be quick. 🙁
 
OK. So I'm probably going to give up on this machine. A second ATX power supply has gone bang, the common denominator being that it was driving the Torch motherboard. The 5v rail is consistently low at 4.2v and only 3.9v at the 68010 processor. Something on this board is dragging things down and pumping a massive amount of current somewhere - but nothing seems to be getting hot (or even warm). It's beyond my capabilities to fix something like this so unless someone can point me at 'finding shorted components for the complete idiot' it's going on the shelf until I can find time to possibly hand it off to a museum somewhere. Gah.
 
You can always put a high-wattage, low value resistor in series with the +5V supply line to limit the current.

You can then probe various points around the motherboard (I would go for the pins of the decoupling capacitors and tantalums) looking for the lowest DC voltage.

You measure the voltage directly across the capacitors.

This should narrow it down a bit...

Dave
 
Crashedfiesta said:
OK. So I'm probably going to give up on this machine. A second ATX power supply has gone bang, the common denominator being that it was driving the Torch motherboard. The 5v rail is consistently low at 4.2v and only 3.9v at the 68010 processor. Something on this board is dragging things down and pumping a massive amount of current somewhere - but nothing seems to be getting hot (or even warm). It's beyond my capabilities to fix something like this so unless someone can point me at 'finding shorted components for the complete idiot' it's going on the shelf until I can find time to possibly hand it off to a museum somewhere. Gah.
Click to expand...
That's really unfortunate :( I doubt a museum would spend much time with it, apart from adding it to the collection.

There's probably not much more I can do with the emulation either, apart from complete the SCSI part of the OMTI board. The floppy part of the OMTI involves too many unknowns to even attempt at this stage.
 
I'd ditch the ATX power supply and switch to a lab power supply, if you have one. Disconnect all the boards and accessories from power and start with just the main processor board.

If there are any parts which require multiple power rails and might get upset (old DRAMs like 4116es) - remove them.

Bring the 5V rail up from the lab supply with a current limit of a couple of amps max. It'll probably go into current limit.

Then as Dave says, measure the voltage across the various 5V and ground pin pairs and look for the lowest voltage. This works because of Ohm's law and the track resistance - you get a high (relatively) voltage at the power input, then it falls as you approach the short. After the short, it'll more or less stay where it is, unless there's another short somewhere else.

If you have a friend with a thermal camera, they're fantastic for finding short circuits, even when the temperature difference is only a couple of degrees.

Alternatively if you're anywhere near Leeds, feel free to shoot me a PM - I've done this kind of thing before and have some tools which might take the task from "impossible, abandon all hope" to "big headache but fixable".

Also to satisfy my curiosity, what brand of ATX supplies are you using? I'd expect a decent one (Seasonic or similar) to completely shut down if there was a shorted rail. Staying running with all but the 5V rail seems a bit whack, as the 5V is usually the reference rail - the others are regulated based on that.


daver2 said:
It depends which museum of course...

Dave
Click to expand...
I'd say it depends more on whether the museum has a skilled volunteer who cares enough about the machine and has the time to sink into getting it working again.
If the goal is to get it working again, a museum may not be the best place - in that situation I'd normally look for an enthusiast to hand such a machine on to.
 
You must log in or register to reply here.
Back
Top