Welcome, Guest. Please login or register.

Author Topic: Picked up a 1000 today...  (Read 13865 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline DonutKingTopic starter

  • Jr. Member
  • **
  • Join Date: Aug 2011
  • Posts: 61
    • Show only replies by DonutKing
Re: Picked up a 1000 today...
« Reply #44 from previous page: November 22, 2011, 09:27:31 AM »
So the mailman dropped off some GAL chips today :) I've managed to get everything programmed without too much fuss, unfortunately installing and testing them may have to wait for another day...

I've wasted some time fiddling with that XEL sidecar I posted earlier in the thread. I bought some 1Mbx4 RAM chips to upgrade it to 2MB total. It came with 1MB of RAM, in 8 256x4 chips. It can be jumpered for 2MB.
So I tried to install 4 1Mbx4 chips in alternating sockets, leaving every second socket empty. The amiga will boot in this configuration, but it gives a guru error straight away on every boot - once you're past that it boots fine.

Sysinfo detects the 2MB of RAM. On the Phoenix utilities disk there are a couple of RAM testers. Both of them fail with 2MB of RAM, but they pass with 1MB.

One doesn't give an address but one does, it seems everything after address 280000 fails. On Wikipedia there is a Zorro memory map, which says that 200000 is the start of an 8MB of address space for Zorro expansion RAM- 280000 is half a megabyte into this range, If this is not correct please enlighten me :)
Since the capacity of my RAM chips is half a megabyte (4 megabits) it sounds like the second chip is dodgy... except swapping chips around has no effect, it always falls over at that address. I've even tried 4 different chips and it has no effect.


No other configuration of RAM chips will boot. Swap empty/populated sockets, all four chips next to each other on the left or the right side... If I fill all eight sockets with the 1Mbx4 chips it will boot, but will fall over at the same address. Using eight 256x4 chips and jumpering for 1MB works perfectly without errors....


So have I overlooked something, or am I just being greedy for wanting a 2MB fastram sidecar... :/
 

Offline DonutKingTopic starter

  • Jr. Member
  • **
  • Join Date: Aug 2011
  • Posts: 61
    • Show only replies by DonutKing
Re: Picked up a 1000 today...
« Reply #45 on: November 26, 2011, 01:51:30 AM »
Well I've given up on making that sidecar work with 2MB RAM. I've done everything I can think of and it still doesn't work and nobody else seems to have any ideas either :confused:

Still works perfectly with 1MB though, although I'd have loved to get 2MB working, 1MB is better than a boot up the arse :)


Another disappointment was the arrival of an 8735 Agnus which I installed into the board, and configured for 2MB Chipram, but then it wouldn't boot :(

Turns out that I require a new version of the U60 PAL chip to make the board work with an 8735 - otherwise the board will work without that U60 chip, but only with an 8732B or 8732AB if you want 2MB chipram.

That version of the U60 chip was never developed, the most recent version does not support the 8735 - so I've had to give up on that one for now. If you only have a 1MB Agnus you are stuck with 1MB chipram without that U60 chip. The U60 seems to do some address line hacking so you can get 1MB chip + 1MB slow on a 1MB agnus.

The other thing with the U60 is that it apparently doesn't work in a GAL, even on the 2005 boards all the other chips were GALs but the U60 remained a write-once PAL. Andrew Wilson (the board designer) said he never got it to work in a GAL.

So I didn't think this would work but I figured, what the hell. nothing to lose....








IT WORKS! :banana: no idea how or why... but it does :confused:

I ran the PHNXRAME.JED file through paltogal.exe, then burnt it onto a Lattice GAL22V10D, populated all the RAM sockets on the board with 256x4 chips, jumpered for 1MB chipram + 1MB slowram. It's not possible to configure for 2MB chipram with this configuration.

The board passes all the memory tests I tried, even with the sidecar installed with 1MB :)




Strangely enough, using the U60E chip made the phenomena-spectre 'dots' demo work perfectly- I tried it on the phoenix before and the graphics were corrupted, same as on an A500, only the original A1000 ran it properly.

There's some real wizardry going on in that U60 chip...

I have also fitted the kickstart switcher GAL, plus a 2.04 kickstart ROM...





So now I've got 1.3 and 2.04 installed :) The interesting thing about this board is that its got 4 sockets that you can use to install a single kickstart image across 1 megabit EPROMs. Since my programmer is not compatible with the normal kickstart roms or anything similar, I can split an image odd/even then high and low and write it to 4 EPROMs. I've ordered some ROM chips and I'll have a go at getting 3.9 plus updates installed, so I'll be able to have a 3 way kickstart switcher :)

I also fitted the FPU GAL, a 20MHz crystal and a 68882... unfortunately the system doesn't boot with the 68882 installed, screen goes black-grey-white over and over again. If the GAL is left in but the 68882 is taken out it boots normally.... if I leave the GAL and the 68882 in but take out the crystal it does the reboot loop thing again... so it makes me think the GAL I programmed isn't at fault, but maybe the 68882 is faulty... would it stop the system booting like that though? unfortunately I don't have another FPU to test, or a different card that I can try the 68882 in. :(

Under KS1.3 it will boot with the FPU installed but there is graphics corruption and missing text in the AmigaDOS window. something screwy is going on there.


I also tried fitting the chip for the 2000 CPU expansion slot but the system doesn't boot at all then. I don't have a card to use anyway but it should still boot with the chip installed - will have to look into it further.
Still waiting on some parts for the SCSI controller to arrive, but I did track down a Yamaha scsi-ide adapter so I'll try and get workbench installed on a CF card- that'll be the next part of the project :)
 

Offline save2600

  • Hero Member
  • *****
  • Join Date: Jul 2006
  • Posts: 3261
  • Country: us
    • Show only replies by save2600
Re: Picked up a 1000 today...
« Reply #46 on: November 26, 2011, 05:25:43 AM »
Wow, this is an incredible blow by blow tutorial in a lot of ways - thanks DonutKing!

The more I learn about this board though, seems more and more features were never finished or implemented. The direct use of the 8375 is a let down for sure, but I bet the DKB MegaChip and equivs would work. :)

Haven't read every single detail, but have you managed to get the SCSI portion working yet?

Oh, and I've read were, if you fill all of those Kickstart sockets in, certain fat floppy drives no longer fit? You have to mount a slimmer floppy drive instead?
 

Offline mechy

Re: Picked up a 1000 today...
« Reply #47 on: November 26, 2011, 05:33:08 AM »
Excellent work there and thats one sweet a1000 phoenix setup.

its sounding like you have a faulty fpu,assuming its installed the correct way.
hopfully its not a china bootleg remarked part ;))

it will be really cool to see it boot off the cf.

the phoenix was one board i always wanted and couldnt get. i had a chance to get one back in the day for $150,and i should of.

mech
 

Offline amigadave

  • Lifetime Member
  • Hero Member
  • *****
  • Join Date: Jul 2004
  • Posts: 3836
    • Show only replies by amigadave
    • http://www.EfficientByDesign.org
Re: Picked up a 1000 today...
« Reply #48 on: November 26, 2011, 05:53:00 AM »
Great thread DonutKing!

A real service to the community with all the pictures and explanations you are providing.  It will come in very handy, should someone else find a Phoenix board to restore.
How are you helping the Amiga community? :)
 

Offline DonutKingTopic starter

  • Jr. Member
  • **
  • Join Date: Aug 2011
  • Posts: 61
    • Show only replies by DonutKing
Re: Picked up a 1000 today...
« Reply #49 on: November 26, 2011, 06:39:05 AM »
Quote from: save2600;669067


Haven't read every single detail, but have you managed to get the SCSI portion working yet?

Oh, and I've read were, if you fill all of those Kickstart sockets in, certain fat floppy drives no longer fit? You have to mount a slimmer floppy drive instead?


Basically just waiting for various chips to arrive before I can get SCSI working. International postage seems to vary a LOT and coming into Christmas surely doesn't help....

Not sure about the kickstart interfering with the floppy.... It looks like it should be fine. The FPU and crystal fit under the floppy.

I'm hoping it's just the FPU that's faulty  cos thats easiest to fix... Just replace it :)


Glad you guys like it, happy to help out anyone else trying to get one of these boards going :)
 

Offline DonutKingTopic starter

  • Jr. Member
  • **
  • Join Date: Aug 2011
  • Posts: 61
    • Show only replies by DonutKing
Re: Picked up a 1000 today...
« Reply #50 on: November 29, 2011, 12:25:36 AM »
Some further research...

On the Phoenix utilities disk there is a list of jumper settings which isn't in any of the manuals you can download off the web. I've managed to extract it using WinUAE.

http://dl.dropbox.com/u/9151127/amiga1000/jumpers.doc

I think I know why the system doesn't boot when the U70 chip for the B2000 CPU slot is installed - there is a row of 3 jumpers, L118, which need to be on if the chip is installed without a B2000 CPU card.
These jumpers are actually missing from my board, I've got some spare IDC headers lying around so I might solder them in just so I can leave the U70 chip in the board (I'll probably lose it otherwise which will be the cause of much angst if I ever do track down a B2000 CPU card...)

There are also some jumpers for the FPU:
Code: [Select]
Link       Position                                Ref.    Norm.
------------------------------------------------------------------------------
L26        L of U61                                COPRO

This 2-pin link forms the left end of a 6-pin link block. It feeds into pins
18 and 19 of unused the 20L8 PAL U61 that controls the maths coprocessor. It
is not used at the moment.


L27        L of U61                                COPRO

This 2-pin link forms the right end of a 6-pin link block. It links pin 14 of
the 20L8 PAL U61 to earth.
 

L28        L of U61                                COPRO   IN

This 2-pin link forms the middle of a 6-pin link block. It feeds the DTACK
signal from Gary to the 68000, and should be in if there is no maths
coprocessor present. If it is out, DTACK is not passed through, and the 20L8
PAL controlling the maths coprocessor has to mix the DTACK signal from Gary
with the DSACK from the maths coprocessor, to produce a DTACK signal from
the 68000.

Although it says PAL U61 that socket is actually for a kickstart EPROM and is empty on my board. I'm pretty sure the chip it is supposed to be referring to is U21, which controls the FPU.

This schematic would confirm it...
http://phoenix.a1k.org/pix/Sheet4-CoProcessor.jpg

On my board only L28 is on, the others are off. Seems that L26 is unused, while L27 I'm not sure about?
It appears to me that L28 should be off if an FPU is installed - however, the board will not boot AT ALL if this jumper is out, whether an FPU is installed or not. Just sits at a black screen. I also tried different combinations of L27 and L28 but I couldn't get it to work.

Of course I can't discount that the FPU is not faulty yet...

any advice/suggestions? :)



I've also uploaded the other files I extracted. The catalogue might be of interest to some people. These boards originally retailed for $945! a 1MB RAM expansion was $399 and an 8MB expanion was $999.

http://dl.dropbox.com/u/9151127/amiga1000/catalogue.doc
http://dl.dropbox.com/u/9151127/amiga1000/68881.doc
http://dl.dropbox.com/u/9151127/amiga1000/scsi.doc
« Last Edit: November 29, 2011, 12:27:57 AM by DonutKing »