bryan@cs.utexas.edu (Bryan Bayerdorffer @ Wit's End) (04/11/90)
There seem to be a number of strange phenomena being experienced by users of the A2630 and older versions of the Microbotics HardFrame. I've just become one of them. With a new 2630 in my rev4.3 A2000, I get random gurus (usually #3 or #4) while booting from the HF/Q80S. These gurus occur at random points in the startup-sequence, so the problem doesn't appear software-related. After several tries, the machine eventually makes it all the way through the s-s, after which everything is fine. The 68030 data cache is always off. Does anyone have a definitive fix for these problems? Various people have claimed that the U9 PAL on the HF needs to be replaced, but at least one person had no success with this. Does the U5 bus control PAL on the rev4.3 motherboard need to be replaced as well? Another question: does anyone have a HardFrame entry for the CardROM file used by setCPU? I'd appreciate mail on this. About wIconify becoming befuddled by the Fastbench screen label: Instead of using setcpu NOPATCH, or zapping the Fastbench string in the setcpu binary back to 'Workbench,' it's only necessary to run setcpu after wiconify has been installed. Once wiconify has a pointer to the WB screen, it never looks at the screen title string again, so it can safely be changed.
daveh@cbmvax.commodore.com (Dave Haynie) (04/11/90)
In article <480@mohawk.cs.utexas.edu> bryan@cs.utexas.edu writes: > > There seem to be a number of strange phenomena being experienced by >users of the A2630 and older versions of the Microbotics HardFrame. I've just >become one of them. With a new 2630 in my rev4.3 A2000, I get random gurus >(usually #3 or #4) while booting from the HF/Q80S. These gurus occur at random >points in the startup-sequence, so the problem doesn't appear software-related. >After several tries, the machine eventually makes it all the way through the >s-s, after which everything is fine. The 68030 data cache is always off. Older Hardframes have a definite problem with bus arbitration that will only show up with A2620, A2630, GVP, etc. CPU cards with on-board DMA-able memory. They have a fix for this, which is apparently a PAL change. Some DMA controllers on some older A2000s run into a problem with a Signetics 74ALS245 buffer chip at position U605 on the motherboard. The suggested fix for this problem is the addition of a 1K resistor between pins 11 and 20 of U605. This fix is already incorporated into A2000s from Rev 4.5 on up. I don't know how badly the Microbotics board is affected by this problem; it almost certainly is, and almost certainly sees much less of a problem than the A2090. >Does the U5 bus control PAL on the rev4.3 motherboard need to be replaced >as well? There is no U5 bus control PAL. That's only on 4-layer A2000s, the German designed motherboard. All US and most other 2000s are the 2-layer "B"2000 designed in West Chester, the one with Fat Agnus and a megabyte of RAM on the motherboard. > Another question: does anyone have a HardFrame entry for the CardROM >file used by setCPU? I'd appreciate mail on this. You don't need a CardROMFile entry for HardFrame. It uses a byte-wide boot ROM which gets copied into RAM automatically on bootup. -- Dave Haynie Commodore-Amiga (Systems Engineering) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy Too much of everything is just enough