Amiga 2000 - Zorro Troubles

BigRed

Member
Joined
Jan 15, 2019
Posts
38
Country
Germany
Region
Hesse
Hello dear Community of Amiga Lovers :)

I am currently trying to expand an Amiga 2000 I bought sometime last year.
In the meantime I had some trouble with moving, that's why the long delay.
This is also my first venture into big box Amigas, until then I only owned A500 and A1200.

What I have is a quite nice stock A2000 Rev 6 board with no (visible) damage and battery removed
long time ago. It was equipped with a GVP SCSI Card.
I also installed a Picasso IV which worked as a scandoubler, but I could not use the RTG because this needs at least a 020.
I toyed with it a little while and everything seemed to be working.

This week I started installing some parts I ordered and that is where the trouble started.
First on the list was an ATX Adapter. I pulled the Picasso and the GVP, installed the ATX Adapter from Ian Stedman and
it worked. Then I reinstalled the GVP which also worked.
Next on the list was the Vampire, which I installed using Acill's CPU Slot Adapter.
I removed the stock CPU & the Kick Rom.
Booted and everything worked fine.

Next I installed the Picasso, which did not work.....In fact, neither the Picasso nor the GVP are detected in the Early Startup Menus Diagnostics :blink:
I tried the following:
1. Reinstalled the Kick Rom
2. Removed the Vampire and reinstalled the stock CPU (on the CPU Slot Adapter)
3. Tried every available Zorro Slot
4. Inspected the Picasso & cleaned the contacts.

All to no success. When the Picasso is installed, neither Picasso nor GVP are detected.
If I remove the Picasso, then GVP is detected and everything works fine.
Since the Picasso is not a cheap piece of equipment I am a bit worried.....

Does anyone has any ideas?

Cheers!
BigRed
 
I know from my efforts on the Beta OS 3.1.4 team that we fixed a number of race conditions that were present in OS 3.1's expansion.library and earlier. The conditions showed up with very fast CPUs and various card combinations. I don't want to mislead you that every bug is gone, but if you are still on an OS 3.1 ROM, there may be some hope with a 3.1.4 ROM.

For example, I got the devs to build in CPU hardware-access delays between consecutive card configs, and certain card register read/write processes that a fast CPU like a 68040 or 68060 (or Vampire) could out-pace some expansion board's logic. Even the A2091 benefited (one of the test examples) as did a few other Zorro II cards in both ZIII expansion backplanes and the A2000's ZII. The short version of it is that AutoConfig writes a destination address for the board to move to, and when that happens, the board goes there, and releases the Config_Out signal so the next slot's board (if present) can show up at the $E80000 config address. If the board doesn't show up fast enough, and the CPU has already read $E80000 and found nothing, the board and everything else in the remaining slots never gets configured.

That the GVP shows up at all is telling me you have a later Vampire that resolved the main issues with AutoConfig in general - that's good. It didn't work at all in the beginning, as it was initially geared toward A500's that were likely to be stock systems.

I don't own a Vampire, so I don't have any direct experience with expansion testing against it.
 
Last edited:
Thanks for your answer, and an interesting one at that :)
The problem is, that the Picasso is not recognized even if I remove the Vampire card. (Sorry for the long text btw)

The only difference to the last known working configuration at the moment is:
- Stock 68k CPU is on Slot Adapter in the CPU Slot
- PSU is now an ATX one with Ian Stedmans Adapter (http://www.ianstedman.co.uk/Amiga/designs/Amiga_ATX_Adaptor/amiga_atx_adaptor.html)

I will later today try to remount the CPU into the original socket and try again.

Cheers!
 
Good news everyone! (Or at least, for me ;-) )

Installing the stock 68k in the CPU socket instead of the CPU slot works. Both the GVP and Picasso are detected.
So the culprit seems to be either the slot itself or the adapter board.

Any ideas on that?
 
You should be able to continuity check the 86-pin CPU slot card using the AmigaWiki Rev 6.2 schematics back to the 68000 CPU socket both on the motherboard and the slotcard. There are some signals that the slot has that do not apply to the actual 68000 chip (like CFG_IN and CFG_OUT - AutoConfig signals), but most should match.

/BOSS is the signal that allows a properly designed accelerator slotcard to silence the socketed 68000 chip in the B2000 motherboard. It's a simple logic gate on the _BGACK signal on the 68000 chip (on the motherboard). The schematics will show it on page 2 of the PDF. The 86-pin adapter card solutions typically require the CPU be removed from the motherboard as it doesn't have the proper 'take over' logic for this signal.

On the topic of CFG_In and CFG_out, make sure there are no electrical connections to these lines. The design concept (of the CPU slot) is to have CFG_In grounded as it enters the slot, with some logic bridging it over to the CFG_Out signal on the next slot (if no AutoConfig) card is in a slot. A card in a slot, if it supports AutoConfig, pulls up the CFG_Out line (preventing the downstream card from showing up) and itself appears at $E80000 when CFG_In is low (which is what starts the process). When it configs (gets programmed), it drops CFG_Out and the next card appears.
 
Are you using the latch mode or momentary mode with the ATX adapter? I have one of Ian's ATX adpaters in one of my A3000's and once I put all the cards back in, it wouldn't start up. With no cards or only a couple, it was fine, but fully loaded, it was a no go. I traced the problem back to the the ATX adapter being in momentary switch mode. Once I put it in latch mode, with a latching switch, the problems went away and all was fine.
 
Are you using the latch mode or momentary mode with the ATX adapter? I have one of Ian's ATX adpaters in one of my A3000's and once I put all the cards back in, it wouldn't start up. With no cards or only a couple, it was fine, but fully loaded, it was a no go. I traced the problem back to the the ATX adapter being in momentary switch mode. Once I put it in latch mode, with a latching switch, the problems went away and all was fine.

"Wouldn't start up" as in "got no power?" or in "refused to boot?"

I just put a jumper on the switch header, so the Amiga powers up once I use the switch at the back of the ATX PSU.
I will try reconnecting a switch to the Stedman Board and going to latched mode. Thanks for you input!

Cheers!
 
Are you using the latch mode or momentary mode with the ATX adapter? I have one of Ian's ATX adpaters in one of my A3000's and once I put all the cards back in, it wouldn't start up. With no cards or only a couple, it was fine, but fully loaded, it was a no go. I traced the problem back to the the ATX adapter being in momentary switch mode. Once I put it in latch mode, with a latching switch, the problems went away and all was fine.

"Wouldn't start up" as in "got no power?" or in "refused to boot?"

I just put a jumper on the switch header, so the Amiga powers up once I use the switch at the back of the ATX PSU.
I will try reconnecting a switch to the Stedman Board and going to latched mode. Thanks for you input!

Cheers!

Hopefully all the boards will be detected for you with latch mode!

For me, with the adapter in momentary mode it would power on for a moment then shut down. If I held the momentary switch down it would power up, I can't remember for sure if all my cards were detected or not when I did that.
 
Sadly, that didn't do the trick.

Another user (on youtube actually) told me that the Picasso's Firmware may be too old.
I'm going to check those versions, and find a way to upgrade those.

Thanks so far & Cheers!
 
Back
Top Bottom