Close, but no cigar..

[Pr0f]

So at the end of a long haul and not expecting a victory on the first power up I got a somewhat garbled screen

I suspect it could be a number of things - as I have the following build:

Aurora - not new and repaired with replacement of the largest logic device, the 'MAC' CPLD.

ROM is a Minerva 1.98 in a 64K flash chip (jumpers configured for 48K, chip in socket, no carrier)

Tetroid Super Gold Card

Qubide card running Qubata 3.10 version, and additional GAL to decode 4 Master devices on the bus.

SuperHermes

The fact I am seeing some sort of message on the screen and a stable video (even if it has garbage on it) suggests the main video circuit is good and it's booted as far as the SGC ROM to get that message - which I presume is "Rom version not found contact miracle systems"

Case is a Pandora black rectangular one with AT style power supply


I saw only 1 post on the forum - but I wonder if the video artefacts and missing characters might point to some issue on the Aurora and decoding addresses properly?

I plan over the next day or so to remove Qubide card first, and the floppy drive cable on the SGC and see if a more minimal system might boot up, after that it's pulling the Flash chip to check it's contents...

[BSJR]

...
The fact I am seeing some sort of message on the screen and a stable video (even if it has garbage on it) suggests the main video circuit is good and it's booted as far as the SGC ROM to get that message - which I presume is "Rom version not found contact miracle systems"

Case is a Pandora black rectangular one with AT style power supply

I saw only 1 post on the forum - but I wonder if the video artefacts and missing characters might point to some issue on the Aurora and decoding addresses properly?

I plan over the next day or so to remove Qubide card first, and the floppy drive cable on the SGC and see if a more minimal system might boot up, after that it's pulling the Flash chip to check it's contents...

The screen problem could be the Aurora chip. I reseated my chip so often that I broke one of its legs and needed a new chip. Also its socket was considered a weak point in the Aurora production and Tony Firshman kindly resoldered mine at an Eindhoven meeting.
Still every now and then this chip needs cleaning the legs to get an undistorted display after long periods if inactivity.
The "contact Miracle Systems" message can only come from the SGC and may also come from a bad chip contact.

BSJR

[tofro]

I would really recommend you diminish the amount of potential combination failures and limit yourselves to test things separately:

Test the SGC and QubIDE separately in a known working QL first.
Once that works, test the SGC with Aurora and a known working QL IPC, fix whatever you come along on your way
Next, add the QubIDE, then the Hermes.

There are so many potential reasons for problems that you will have a hard time finding them with an "all-at-once" approach.

[Pr0f]

Yes - my plan was to dial it back in stages - to bring the machine back to the most basic setup.

Question - will the QL boot without an IPC installed? This is custom case - so without the Superhermes - I have no keyboard at all.

It's also clear that it must be booting as far as the patching process - which suggests the basic boot sequence is working to that point.

Good call on testing the other cards with a standard QL though - as I have 2 QL's I can test the SGC and Qubide card in.

[Ruptor]

There are so many potential reasons for problems that you will have a hard time finding them with an "all-at-once" approach.

It is exciting to see what goes bang.

[Pr0f]

So last nights testing -

Unplugged the QUBIDE card - and the display text is now readable. However, I still have some 'snow' on the screen which looks regular and indicates a probable address line issue.

After some further research and looking at Marcels notes on the Gold Card boot - it seems the SGC will boot from it's own ROM - so that error message - now readable is coming from SGC boot sequence. This suggests it can't properly identify the Flash ROM on the aurora, which again could be related to address lines.

I removed the flash and read it's contents in my chip programmer - chip is ok, and it matches one of the ROM images I cut for my ROM switcher, so I can plug this SGC into the black box QL, select that same ROM image on the dip switches and test the SGC can work with it correctly - it should be ok - ROM is basically Minerva 1.98 and TK2 2.23

The next steps:

Test SGC in black box QL
Test QUBIDE in black box QL (I will need to add the regulator chip and put a jumper on the board for bypass when needed)

I suspect both of these will be ok, and this leaves the Aurora board itself. Fun weekend ahead !

[Derek_Stewart]

Hi,

Did you program a new MAC PLD?

[Pr0f]

No - not yet - I have the circuit and parts to do this with the hi-lo programmer - but I haven't built that yet.

I have the spare MAC that I got from you Derek, which is in the Aurora now. It looks like the timing is working ok - as I have stable video and it can take text from the SGC. It's quite an odd problem.

Hoping to get some time this weekend to check the other parts out in isolation - so I can pinpoint where I am looking for the issue.

[Derek_Stewart]

Hi,

I have compiled in a Virtual Machine, all the Aurora PLD code into JEDEC files to be programmed int eh Aurora PLD chips.

Whether new MACH chips can be purchased is for purchase of repairs to existing Auroras, another question.

A long time ago, I was given the PCB file for the Aurora, produced in Protel PCB software, which I imported into Altium Designer and was able to edit the PCB file.

The main problem with the PCB design, was the inner power and ground layers, which at the time of producing the the Aurora, there was no PCB ploygon fill functionality.

I must admit to not understanding this, but after a long read of many PCB documentation I have understood what the Power and Ground Planes are and the use of filled planes.

I have reconstructed the inner planes so that they are electrically correct and should work,

I also started adding names the circuit network, using the PCB schematic. This to check everything is correct.

In the process of this work, Kicad 7 came out, with gives me nearly the same functionality as Altium Designer. I do not have to use Pirated software and not use Windows.

I have started to enter the schematic into Kicad Esschema, but have found some differences between the PCB and the Schematic. Which I am investigating with the 2 Aurora boards I have.

So still work inprogress.

I do not think the PCB files are open source as well.

[Peter]

A long time ago, I was given the PCB file for the Aurora, produced in Protel PCB software, which I imported into Altium Designer and was able to edit the PCB file. [...]

Do you actually want to change the Aurora design, or is this just because you don't have the Gerbers?