3.1 ROM problem - strange 3.0 ROM wiring

[Safalra]

Although I've been using PCs for most of my computer work for several years, I've kept my A1200 around, and recently I decided that it was time I upgraded to OS3.1.

I bought some 3.1 ROMs (unused), read through the installation guide that came with them (and the one in Amiga Format 134), and opened up my A1200, only to find something that didn't look like the pictures. Each ROM is securely housed in the chip-holders, with the two spare holes at the notch end, but there was something else: four green wires.

Each chip has its 3rd and 9th legs missing, and instead have green wires soldered on to them. The two green wires cross over and are then soldered straight onto the chip-holders (so that the 3rd 'leg' is connected to the 9th hole and the 9th 'leg' to the 3rd hole). Clearly this makes getting them out impossible without a soldering iron, and even if I did I presume I'd need to I need to perform similar surgery on the 3.1 ROMs.

So, is there anything I can do? Note that any advanced rearrangement, soldering, or other changes are probably not something I should attempt, as I seem to be cursed to break computers (I've been in my current job as a web programmer for two months and I've already completely destroyed a PC, a ball mouse, and an optical mouse, just by using them normally...).

[bloodline]

This sounds weird... though I don't know the ROM pinouts so I have no idea what is going on... can you post a photo?

[Safalra]

bloodline wrote:
This sounds weird... though I don't know the ROM pinouts so I have no idea what is going on... can you post a photo?

Even better - I can post two photos (although neither is very good):

[alexh]

http://www.amiga-stuff.com/hardware/2mbit-maskrom.html

The modification for some reason swaps pin 38 (A10) and pin 32 (A16).

The only thing I can think of is that you had at one time a "piggy back" kickstart switcher?

If I were you, cut the wires with a wire cutter somewhere in the middle. Making sure the wires dont touch anything (perhaps put tape on the end) just fit your new chip as normal.

If it doesnt work... then maybe you have some sort of developer A1200 and need to bend the NEW ROM legs and re-apply the patch wires.

I feel confident that if you cut the wires the new ROMS will just work if installed correctly.

If you leave enough wire on each end you can always re-connect the old roms later.

[Castellen]

I've never seen such a modification before either.  In this case it's swapping two address lines.
No idea why you'd need to do this.  Assuming they're correct ROMs, this would make data appear at incorrect addresses.

If this modification has been like this from new, then at a guess it's to correct a problem if those data lines in the ROM had been accidentally swapped during a prototype manufacturing run from the ROM manufacturer??

If this is the case and the rest of the computer is standard, then the 3.1 ROMs should work.
Hard to tell from the photos, but it looks like the wire is only soldered to the socket's outer contact, meaning that if you cut off the wire flush at the soldered joint, the 3.1 ROM should still plug in and work normally.

[Safalra]

alexh wrote:
[...]
The only thing I can think of is that you had at one time a "piggy back" kickstart switcher?
[...]
If it doesnt work... then maybe you have some sort of developer A1200 and need to bend the NEW ROM legs and re-apply the patch wires.

I wasn't aware of it being anything other than a normal A1200 until I looked at the ROMs. It came in a new Desktop Dynamite pack back in 1993. The only modification we've done is installing a 400MB internal IDE harddrive and an 8 MB memory card in the trapdoor slot.

I'll try what the two of you have suggested and assume it's the 3.0 ROMs' wiring that's weird and not the motherboard's. I let you know if it works.

[Doobrey]

Castellen wrote:
If this modification has been like this from new, then at a guess it's to correct a problem if those data lines in the ROM had been accidentally swapped during a prototype manufacturing run from the ROM manufacturer??

Is the '9311' the datestamp of the ROMs? ie week 11 of 1993
I just looked at amiga-hardware.com's   A1200 Rev1D3 piccy and it's ROMs are labelled '9307' ..and no reworking on them either.
 I guess the '1043' and '1044' refer to the data that Macronix programmed them with, and these are the same too.

It'd be interesting to find out what rev motherboard these are installed on.

[Safalra]

Doobrey wrote:

It'd be interesting to find out what rev motherboard these are installed on.

The inside of the top of the case says 'Rev B'.

Update on my attempt to install the ROMs:

As I lifted the old 3.0 ROMs out I found that the chip-holders they were in were themselves inside another set of chip-holders on the motherboard, so I was able to lift the 3.0 ROMs out, wires and all, and plug the 3.1 ROMs in to the 'true' chip-holders. Unfortunately, when I powered up the Amiga it doesn't work. It makes the usual clicking noise and the power light goes dim, but it doesn't then brighten up again, and rather than booting it shows either a purple screen, a red screen, a black screen, or a black screen with thin green lines moving upwards (each time I cycle the power it seems to pick one of these four options at random).

[vic20owner]

Uhoh! If I remember correctly, you align the roms with the RIGHT side of the socket, not the left side.  So, pin 1 of the chip actually goes into the pin 2 hole.

Is that how you installed them?

If so, then it could actually be that you have a motherboard which had a production bug which was "fixed" at the factory with those additional sockets.  You may need to swap those addresses after all.

The fact that BOTH roms required the switch makes me think it's the motherboard not the roms.

[lurkist]

I posted about this a while back.

http://www.amiga.org/forums/showthread.php?t=15391

I have removed the pics from my webspace, but they looked pretty much the same.  I`ve had no problems with the machine.

Cheers

[vic20owner]

lurkist, what rev motherboard was that you saw the same rom mod on?

This may be a clue.  Also, were they 3.0 roms, and did they have the same chip numbers?

This could be the answer we are looking for.

[Safalra]

vic20owner wrote:

Uhoh! If I remember correctly, you align the roms with the RIGHT side of the socket, not the left side.  So, pin 1 of the chip actually goes into the pin 2 hole.

Is that how you installed them?

Yep.

If so, then it could actually be that you have a motherboard which had a production bug which was "fixed" at the factory with those additional sockets.  You may need to swap those addresses after all.

The fact that BOTH roms required the switch makes me think it's the motherboard not the roms.

Given my record with computers (since my lament in the first post in this thread my PC monitor has also started working only intermittently) I'm not too keen on the idea of performing surgery on the ROMs. I think I might just give up and plug the 3.0 ROMs back in and then put the 3.1 disks and manual up on eBay (I don't think I can sell the 3.1 ROMs as I can't confirm that they work).

[Doobrey]

Safalra wrote:
 I'm not too keen on the idea of performing surgery on the ROMs.

You could always get 2 pairs of 40 pin sockets and perform the hack on them (ie treat the second pair as if they were the ROMs).
That way you can test the 3.1 ROMs without damaging them.

[Safalra]

Doobrey wrote:

You could always get 2 pairs of 40 pin sockets and perform the hack on them (ie treat the second pair as if they were the ROMs).
That way you can test the 3.1 ROMs without damaging them.

Ah, cunning. I might try that.