Individual Computers - New accelerator cards for A600 & A1200

[mfilos]

Originally Posted by DonMac

Really don't know

The original A1200 ROM is 512kb, but when I'm using "maprom *" the machine doesn't start. The machine has a kickrom switch (1.3 and 3.1). Think thats why I must use the 1mb option?

The original kickstart for A1200 is 512KB.
Even if you're using a special and custom 1.3+3.1 Kickstart (meaning 2x256KB for 1.3 + 512KB for 3.1 = 1MB) burned in one Eprom and selectable by a switch, by switching to Kickstart 3.1 you still allocate 512KB for ROM.

That's why you have to use one asterisk on MapROM option (512KB ROM) because the dual asterisk is meant for 1MB custom ROMS.
Maybe this doesn't have to do at all with your problem, but better use the correct options to be sure.

[DonMac]

Originally Posted by mfilos

Maybe this doesn't have to do at all with your problem, but better use the correct options to be sure.

Thats why I tried the "maprom *" function this morning. But now the OS doesn't boot. Have to fix that tonight

After a CWB test re-install (and no ACAtune) a few days earlier the guru's still appear.

Only thing I would like to know is why the maprom ** function works and the maprom * function fails. Will go into that this evening.

[DonMac]

OK.
After removing the ACA1231 and adding the extra * as maprom parameter the CWB boots again.

Don't know why I need the 1MB setting. I do know this system is equipped with a HRTMon Freezer. Could this be an explanation to this?

My A600 with ACA630 uses the single * with maprom.

[Schoenfeld]

Originally Posted by DonMac

Don't know why I need the 1MB setting. I do know this system is equipped with a HRTMon Freezer. Could this be an explanation to this?

ACA accelerators always treat Kickstart as 1MByte in size. The difference of "one star vs. two stars" is that two stars copies the full 1MByte from the standard locations 0x00e0.0000-0x00e7.ffff and 0x00f8.0000-0x00ff.ffff. If you only use one *, then the mirror copy at 0x00e0 will also be filled with the contents of 0x00f8.

If your HRTmon freezer is more than just an IRQ7-button, but a ROM that uses the lower space at 0x00e0, it will of course not work. In any case, it cannot hurt to use **.

I've been working on the card all day yesterday, chasing a ghost: I thought I had found a reliable way to make the card crash, but it turned out to be a stack overflow, and other accelerators also crash in that specific case. Increasing the stack solves the problem for all cards. It might be a good idea to increase stack size in general, as the Amiga OS does not check stack overflows. The ACA1231 has so much memory that you can surely spare a few kilobytes.

As a result of yesterday's work, the memory controller got a full overhaul, but the performance of the card is reduced now. I'll continue on that path today and try to find a way to make the card crash. One thing I have done is to extend the timeout for the watchdog, although I'm almost convinced that DonMac's card has a defect if it keeps on crashing with 8000 0002 gurus.

Whatever the reason is, all cards will be fixed under warranty.

Jens

[DonMac]

Just my luck

At the moment I'm in the process of buying a 2nd A1200. Picking it up this Tuesday. Very curious how this machine is going to react.

I already send a support email to Vesalia a couple of days ago. They forwarded it directly to Individual Computers.

[Schoenfeld]

...and I'm at the point where I found that Burst is causing trouble. For those who have problems, please disable burst. New cards will most probably be shipped with burst disabled in hardware. Performance goes down by about 3.5%.

Jens

[Mr B]

This is only relevant to ACA1231 owners, yes? Is this something that will be a warranty issue? Will handling of such an issue further delay the Indivision MKII? Do you have any indication of how many are affected by the issue?
How does the fix (burst of) affect the performance comparison between a ACA 1230 vs a ACA 1231 @ say 42MHz? (42MHz wasn't picked completely at random. Both cards were made with that CPU.)

Sorry to see the extra work having to be made, especially after we talked about more exotic designs requiring more, and more importantly, more elaborate error proofing. Guess it just proves the point. Adding variables the manufacturer cant control isn't beneficial for someone trying to provide a service to a small group of insane buyers aiming to further there already completely crazy hobbies. Your doing great work. Thanks. Keep it up, please.
B!

[phl22]

For me the only time aca about yellow guru

[Schoenfeld]

This only relates to ACA1231, but not to ACA1230 or ACA630. Since I cannot reproduce the errors here, I am sending two test-cards to customers in Germany tomorrow. The following steps will be decided depending on their results.

Jens

[melbourneben!]

recieved my aca 1231, working great so far

[mfilos]

Since I reported an error with Transcend 133x CFs in the past I can say the following...
Upon power on, you get the classic Kickstart screen but with a single reset (Ctrl-Amiga-Amiga) the issue is solved. Any reboot from that point doesn't have any problem. Only power on this issue occurs.
AFAIK that is a really common issue that exists to some users and depending on the CF or HD they use.
With Sandisk's simple CF you don't get this issue and no reset is needed.

ACA1231 is working great and the new heatsink gets only slightly warm. Nice job on that one Jens!

[qrulf]

Hi, just received my ACA-1231 a few days ago and have had no luck with it at all. I've mailed Vesalia, but until they reply, maybe someone here could help or had an idea about what could be wrong, what to test, or if this is just normal ;-)

My A1200 gets unstable and freezes/ reboots from time to time.

It can boot into workbench but sometimes it just freezes. ACATune recognizes the board and I'm able to maprom and turn off burst & cache (burst seems to be turned off as default).

I tried to disconnect the harddrive (CF & adapter). It can boot up to the insert disk welcome screen, but it then freezes and reboots with a Software Failure afterwards (Error 8000 0004 Task: 08003290):


Configuration:

• I have no other expansions or hardware mods or fixes.
• Motherboard revision 1D1.
• Tried both the original 3.0 ROM and my new 3.1 ROM.
• Using an ATX PSU (380W, 5V@20A, 12V@17A, -12@0,8A). Also tried the original power brick, but that did not change anything.

ACATune -status before any settings set:

[Schoenfeld]

It turned out that even with the increased immunity of the ACA1231 against false assembly of the A1200, we still need board modifications if the timing fixes haven't been applied yet. The info is hidden in this thread:

http://eab.abime.net/showthread.php?p=726765#post726765
http://eab.abime.net/showthread.php?p=726776#post726776

...and a more complicated one:

http://eab.abime.net/showthread.php?p=753433#post753433

Should you continue to have trouble with the card, please contact Vesalia.

Jens

[Retrofan]

qrulf what Hard Drive are you using? Anyway my only problem is with ACA 1230/56 and Fast Ata. It seems an "slower" Cf Hd is the best, as it solves it (at least for a time), I have to try with the slowest one.

[1time]

I have made all the timing fixes and that resolved many of my problems.

[DonMac]

Bought a second A1200 and installed the ACA1231. Worked like a charm. No problems at all (even burst enabled as a test).

The first A1200 now houses a Blizzard 1230 so now they are both perfect machines.

[jimbob]

. .must . . . . resist . . ..

I don't even have a 1200, but I hope I'll pick one up some day and when I do, all the IC stuff will be sold out

[qrulf]

Originally Posted by Schoenfeld

...and a more complicated one:

http://eab.abime.net/showthread.php?p=753433#post753433

Just to sum up components I would need to order the following components for the complicated part (I'm sorry about the stupid questions, but I'm not that used to reading those specs).

3x 22pF capacitors (size 0805) in places E121C and E122C and one extra if there's nothing in E125C:
http://export.rsdelivers.com/product...p/0110223.aspx

2x 0805 ferrite beads, in places E121R and E122R:
http://export.rsdelivers.com/product...v/0148222.aspx

[FOL]

Originally Posted by qrulf

Just to sum up components I would need to order the following components for the complicated part (I'm sorry about the stupid questions, but I'm not that used to reading those specs).

3x 22pF capacitors (size 0805) in places E121C and E122C and one extra if there's nothing in E125C:
http://export.rsdelivers.com/product...p/0110223.aspx

2x 0805 ferrite beads, in places E121R and E122R:http://export.rsdelivers.com/product...v/0148222.aspx

That sounds correct, check your board first though as the first timing fix may already be applied. E121C and E122C

[qrulf]

Originally Posted by Retrofan

qrulf what Hard Drive are you using? Anyway my only problem is with ACA 1230/56 and Fast Ata. It seems an "slower" Cf Hd is the best, as it solves it (at least for a time), I have to try with the slowest one.

Hi, I've been using a Sandisk Extreme III (x133) 2GB flashcard, that has worked in my A500 with an Alfa Power controller for some time. I've also been trying out a Kingston Elite Pro 4GB (also x133) card in the A1200, but apparently it has a bad reputation with accelerators I can see. The reason for those cards is that they've been lying around from an old linux box.

As my local dealer stocks them, I might consider getting another Sandisk card (Sandisk Ultra 30mb), I can see there are no reports of any of the Sandisk cards not working. Or maybe or a Verbatim card, but there are no reports on them. Hmm.

Until then it might be safest just to use an old (real) IDE harddrive for testing purposes?! ;-)