Announcement

After 5 years serving the vintage Apple enthusiast community, ThinkClassic has been marked for closure and is now in caretaker mode. Please see this thread for further information. Please direct any questions, comments and enquiries about the website, management and ownership to this thread.

You are not logged in.

#51 2016-05-16 23:21:28

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

Yeah, the HD20 and floppy are tied together fairly tightly. The current HD20 behavior is workable, it's just not ideal. smile

Any idea how much testing the romdrv0.9.6 driver has seen with the SE/30? I thought I was all finished with this, but then ran into 3 big problems when I tried it on an SE/30, and 2 of those 3 problems can also be reproduced with the "stock" driver that predates all my changes. I'm not sure if my SE/30 hardware is bad, or if there are some incompatibilities with the SE/30 that were never noticed before and date all the way back to 0.9.6 or earlier.

These two SE/30 problems happen the same way with either my modified BMOW rom image or the iisi+romdrv0.9+nomem+nosum+img.bin image:

- Booting from System 6.0.8 Tools floppy hangs at "Welcome to Macintosh".
- Booting from a Floppy Emu in HD20 mode (containing System 6.0.8 ) also hangs at "Welcome to Macintosh".

And with my modified BMOW rom image, any attempt to decompress or boot the rom disk freezes the computer. sad

From all accounts, I'd thought this should work on the SE/30 as-is with no changes or machine-specific code needed. I'm not sure if those hangs at "Welcome to Macintosh" are a pre-existing problem that just wasn't noticed before, or a symptom of some hardware problem on my SE/30.

Offline

#52 2016-05-17 01:45:08

uniserver
Member
From: Sf, Mi
Registered: 2014-05-15
Posts: 956
Website

Re: Mac II series ROM hacking testing

i do like the menu for choosing R or A.


#I Re-Cap √Mac √NeTX √Amiga Boards - A/B - PSU# (MacCaps.com)  Modern SCSI HD's - For Old Macs - Pre Cfg'd - 10k RPM! 73gb!! $50 + free shipping  -- Mac 128K Re-Ram kits (16 Chips) $35 + shipping, Floppy Issues?-> Bourns Filter Solution 128k - SE/30, $16 + shipping

Offline

#53 2016-05-17 02:17:37

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

Did you try it on an SE/30?

Offline

#54 2016-05-17 02:20:58

uniserver
Member
From: Sf, Mi
Registered: 2014-05-15
Posts: 956
Website

Re: Mac II series ROM hacking testing

Oh no…  haven't tried anything buddy.  my roms on the desktop have been untouched for like a full year or more.  i kinda made the 2mb 8meg ers the way i wanted them and left em.


#I Re-Cap √Mac √NeTX √Amiga Boards - A/B - PSU# (MacCaps.com)  Modern SCSI HD's - For Old Macs - Pre Cfg'd - 10k RPM! 73gb!! $50 + free shipping  -- Mac 128K Re-Ram kits (16 Chips) $35 + shipping, Floppy Issues?-> Bourns Filter Solution 128k - SE/30, $16 + shipping

Offline

#55 2016-05-17 14:16:22

bbraun
Member
Registered: 2014-05-29
Posts: 1,064
Website

Re: Mac II series ROM hacking testing

Huh, I hadn't heard of those issues before.  I did all my development on a IIx, which is pretty similar to the SE/30, and I know dougg3 has sold lots of these rom simms to SE/30 owners, so I think it has had a fair bit of use, although I'm not sure how much of that is with 6.0.8.  At least the 2MB simms probably used 6.0.8 exclusively, just due to space constraints.

Offline

#56 2016-05-17 21:17:31

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

Taking a step further backwards - I tried programming a stock IIsi ROM image (512K) to a 2MB simm, and putting it in the SE/30. It won't even boot - just shows garbage on the screen when you turn on the power switch. I thought the stock IIsi ROM was supposed to work on the SE/30, no? Is there some change to the rom image necessary, or jumper setting on the SE/30 logic board, or something else I missed?

Offline

#57 2016-05-17 23:03:39

bbraun
Member
Registered: 2014-05-29
Posts: 1,064
Website

Re: Mac II series ROM hacking testing

Yeah, stock IIsi ROM is supposed to boot fine.  One problem ISTR dougg3 saying was the SE/30 seems particularly finicky about the thickness of the ROM SIMM PCB, and even variations in manufacture could cause the SE/30 to be sad.  Maybe it isn't making reliable contact?

Offline

#58 2016-05-18 00:14:50

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

Yeah, he mentioned the thickness issue to me, but I'm kind of skeptical that's the issue here. I tried it with two different SIMMs, and both of them are from Doug's old stock 2MB PCBs that were a little thicker than the later 8MB ones. I also tried it with one of my new 4MB SIMMs. And I tried booting while physically pushing the SIMM against the socket's pins. And the nature of the problems seems so consistent and reproducible, it doesn't feel like a bad connection problem. Of course, I could be all wrong.

So anyone want to play armchair quarterback here and guess what could be wrong, or suggest another test I haven't thought of yet? I am well and truly stumped on this one. I kind of want to blame the hardware, except that everything is working perfectly as long as I use the original rom. Here's what I've got:

Mac SE/30, 20MB RAM, internal SCSI drive has System 7.1
logic board has been partially recapped in the past - some evidence of minor cap leakage
intermittent video problem - screen periodically blooms to near full white

With stock SE/30 ROM SIMM:
boots OK from SCSI, internal floppy, or Floppy Emu external floppy
everything seems OK, games work OK, no obvious problems or flakiness or crashes

With stock IIsi ROM image programmed to a 2MB SIMM:
plays normal startup sound, shows garbage video for about 5 seconds, then plays chimes of death. Still has garbage video.
This was so unexpected, I tried it three times with different SIMMs. I also re-read the contents of the SIMM and verified that it was programmed correctly, and also diffed the contents against the other IIsi-derived rom images to make sure I was using the right rom. The exact file is named 36B7FB6C - Mac IIsi.ROM and was taken from the Internet Archive collection of old Mac roms.

With iisi+romdrv0.9+nomem+nosum+img.bin downloaded from http://synack.net/~bbraun/macromboot.html and programmed to 2MB SIMM:
freezes at launch of Finder when booting 6.0.8 floppy from internal drive
freezes at Welcome to Macintosh when booting 6.0.8 floppy image from Floppy Emu (also tried a few other floppy images with similar results)
Sad Mac 0000000F, 00000001 when booting from SCSI
tries to boot from SCSI if both SCSI and HD20 are present
with SCSI disconnected: freezes at Welcome to Macintosh when booting 6.0.8 floppy image from HD20
boots OK from rom disk when holding R at startup, and successfully mounts SCSI and HD20

With my latest rom image http://www.bigmessowires.com/romdrv/2MB-fc2.zip and programmed to 2MB SIMM:
freezes at Welcome to Macintosh when booting 6.0.8 floppy image from Floppy Emu
didn't try it with SCSI
with SCSI disconnected: freezes at Welcome to Macintosh when booting 6.0.8 floppy image from HD20
pressing R or A at the splash screen (for rom or ram disk) freezes at a blank screen
pressing X at the splash screen shows "decompressing disk image" and then freezes

Offline

#59 2016-05-18 01:15:59

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

I forgot an obvious test:

With stock SE/30 ROM image programmed to a 2MB SIMM:
boots OK from SCSI, internal floppy, or Floppy Emu external floppy

This would seem to rule out any possible physical problem with the SIMM or its thickness.

Offline

#60 2016-05-18 03:37:58

bbraun
Member
Registered: 2014-05-29
Posts: 1,064
Website

Re: Mac II series ROM hacking testing

Just kind of spitballing because I don't have any real ideas, but it seems like RAM utilization causes a problem?  Try with less/different RAM?

Offline

#61 2016-05-18 18:06:31

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

I think I figured this out! As usual, bbraun's instincts were good, and it seems like a RAM issue. It looks like the IIsi-dervied ROMs are more picky than the stock SE/30 ROM about the type of RAM - either about having matched types/speeds, or about whether there's a 9th parity chip on the RAM simm.

RAM set A: four 1MB SIMMS, of 2 different types. All SIMMs have a 9th parity chip. Speeds are 70, 70, 100, 100ns
RAM set B: four 4MB SIMMS, all the same type. All SIMMs have a 9th parity chip. Speed 80.
RAM set C: four 4MB SIMMS, of 4 different types. No SIMMs have a 9th parity chip. Speeds of 70, 80, 80, and 100 (?)

Any RAM config involving sets A and B works. I tested 4MB, 16MB, and 20MB with the stock IIsi ROM and with my BMOW ROM. All is good.

Any RAM config involving set C with a IIsi-derived ROM does not work. It either fails hard with the chimes of death, or sort of works but then encounters all the weird problems I described above.

Any RAM config involving set C with the stock SE/30 ROM works.

Also, if I put set C in my IIsi, I get chimes of death.

Of course this could all simply be due to RAM set C being bad, but there's evidence against that:
- If Set C was bad, then it shouldn't have worked with the SE/30 ROM either
- Macdrone reported the same symptoms I had when putting a IIsi ROM into a "known good" SE/30

My guess is that this might be what was causing trouble for so many of Dougg3's SE/30 customers, and not flaky rom sockets or PCB thickness issues. Has anyone examined the DRAM controller on the SE/30, or compared the DRAM SPD configuration stuff between the IIsi vs SE/30 ROMs, and noticed anything that could explain this difference in behavior? 8 chip RAM SIMMs are fine, aren't they? I'm noticing that all my other Macs have 9-chip SIMMs or composite SIMMs, but I'm 99.9% sure the Mac never uses that 9th chip and it shouldn't matter if it's present or not. Which means the problem is likely due to the mix of types and speeds in Set C.

Last edited by bigmessowires (2016-05-18 18:12:42)

Offline

#62 2016-05-18 18:35:50

bbraun
Member
Registered: 2014-05-29
Posts: 1,064
Website

Re: Mac II series ROM hacking testing

Huh, that's pretty interesting.  My understanding of the memory initialization is as follows:
The GLUE memory controller used by the SE/30, II, IIx, and IIcx I believe, only has 2 banks: A and B.  There's no software configuration of the controller logic (the controller is pretty dumb from a software perspective), and the only thing software can do is tell it where bank B should be located.
When the ROM goes to configure things, it sets bank B to be located at 64MB mark, writes to memory at intervals to find out the size of bank A, repeats for bank B, then configures bank B's location to be contiguous with the end of bank A.
It then writes this information to a chunk table, which I don't have much understanding of.  The chunk table is later referenced by the MMU configuration, and is modified to support the .EDisk capabilities.  This is something I've wished I understood better, because if you really want to carve out a chunk of RAM that so the system won't see it or mess with it, the chunk table seems to be the place to do it.

So, I don't think you can configure the memory controller in the SE/30 to configure parity, RAM speed, etc.  The only thing I understand that it can do is configure the location of bank B and the size.  Does it make any difference which bank RAM set C is in?
There could be something about the RAM that confuses sizing or mmu configuration in the IIsi ROM?  I'm really not sure.

Offline

#63 2016-05-18 20:08:31

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

Yeah, it's weird. I guess I'll chalk it up to "use matched sets of SIMMs" since that's the advice I've always heard for Mac RAM upgrades. Although given your description of the memory initialization process, I'm not sure why it should matter, as long as all the RAM is fast enough.

I think you're right that something is confusing the sizing or mmu configuration in the IIsi ROM. There's a related clue that I didn't mention before, because I thought it was a red herring. I modified MESS to expect a larger ROM for the SE/30, and then debugged it with various IIsi-derived ROM images. In all my tests it detected twice as much memory as was actually present, if at least 8MB was present, or else it failed to boot at all if only 2MB or 4MB were present. But I wasn't sure whether this was real, or some artifact of MESS or my changes to it. I'm still not convinced it's real.

I don't think there's a choice about which bank the RAM goes in, except for a 32MB config, according to http://lowendmac.com/ram/se30.shtml. For the 20MB config, I tried swapping the 1MB and 4MB SIMMs between Banks A and B, and the machine wouldn't boot.

Offline

#64 2016-05-18 20:36:56

macdrone
Member
From: Rainier, Or
Registered: 2014-05-25
Posts: 246

Re: Mac II series ROM hacking testing

I will say that guide is wrong as I dont have mode 32 on my spare unit, it does not have a IIsi rom and shows 20 mb of ram.  So not seeing >8mb of ram is not exactly true.

Offline

#65 2016-05-19 01:06:38

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

I think without MODE32 (or its equivalent), the SE/30 will see the extra RAM, but can't use it.

Offline

#66 2016-05-20 04:22:42

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

It looks like my theory about a problem with SE/30 unmatched RAM SIMMs was wrong, or at least incomplete. sad Even with the "good" RAM configs, while the ROM disk stuff works OK, I consistently get a Sad Mac when booting from SCSI - not sure how I missed that. And while it first appears that HD20 is working, it always chokes with some kind of disk error after accessing a couple of files on the HD20, and corrupts the disk. So weird! These two problems both occur with the stock IIsi ROM image as well as my customized BMOW ROM image.

I think I have to wave the white flag on this, because I've sunk way too much time into it, and I probably can't find an answer without more test results from other SE/30 hardware. I'm left with a few possible explanations, none of which feel satisfactory:

1. This particular SE/30 is broken (but then why does it work fine with the stock ROM?)
2. The IIsi ROM code isn't really compatible with the SE/30 (but everyone says it is)
3. Something is physically wrong with my replacement ROM SIMM (I tried two different SIMMs)
4. I am haunted by evil retro-computer spirits (seems likely)

Offline

#67 2016-05-20 15:03:06

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

I'm wondering if the Sad Mac when booting from SCSI could be due to the specific system software on the hard disk, rather than a general problem. Would attempting to load MODE32 cause a crash like that? Or maybe if the system software were installed specifically for the SE/30, instead of the "System Software for any Macintosh" option?

I also started wondering what happens when the system software loads and applies ROM patches - won't this cause big problems with a customized ROM? It will think it's patching the SE/30 ROM when it's really patching a (modified) IIsi ROM. Or if it uses the ROM checksum to determine what patches to apply, that will also be bad news since the modified ROM doesn't have a valid checksum.

Offline

#68 2016-05-20 18:40:30

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

Yes, the Sad Mac when SCSI booting was due to the presence of the 32-Bit System Enabler in the System Folder on the hard drive (System 7.1). After removing that, the SE/30 boots fine from SCSI with the customized BMOW ROM image. I guess using a 32-bit enabler on a system that already has a 32-bit clean ROM causes problems? I would have expected it to just do nothing.

Now I'm just left with the mystery of why SE/30 HD20 support kind-of-mostly works, and can boot from HD20 or mount HD20 after booting from another disk, but then subtly corrupts the HD20 disk after open and closing a couple of files.

Offline

#69 2016-05-20 19:20:45

bbraun
Member
Registered: 2014-05-29
Posts: 1,064
Website

Re: Mac II series ROM hacking testing

Cool, glad you found the 32bit enabler problem!
As for the install being machine specific, I can't recall, but I think it's actually fine with the different ROMs.  It's mostly checking the actual hardware (using the VIA values in the SE/30's case), and not the ROM.  There have been some 3rd party products that get upset, I know the Turbo040 doesn't like the IIsi's ROM, and I believe the Radius Rocket software also gets upset about a IIsi ROM.  The latter case is just using the check to prevent using a Rocket in the IIsi.
For patching the ROM, there's different kinds of patching that takes place, and at least one form I'm not terribly familiar with, but one thing I do know is that the ROM is actually left in place at its normal location in the address space, and is still ROM.  They don't do any fancy tricks like copying it to RAM and then using the MMU to remap the RAM into the right location or anything like that.  I haven't heard of any stock Apple software having an issue with it (this is afterall how the ROMs were developed AFAIK), and only know of those two 3rd party products that have an issue.

Offline

#70 2016-05-20 20:56:30

cc333
Member
From: North S.F. Bay Area, CA
Registered: 2014-05-23
Posts: 600

Re: Mac II series ROM hacking testing

What about another ROM, such as, say, the IIfx? That one has been known to work on the SE/30, so it could be worth it to try a stock IIfx ROM to see if it'd work (I'm not sure if the IIfx has built-in HD20 support, though).

I have a FloppyEmu set to HD20 mode and a IIfx, so perhaps I could test my SE/30 with the IIfx ROM to see what happens?

c


Main Macs: Early '09 Mac Pro, Mid '12 MacBook Pro 13"
Secondary Macs: Early '08 Mac Pro, Mid '12 MacBook Pro 15"
Playthings: Mac SE/30, 3.0 GHz Mavericks-based HackServe, Many others....
Desired: Lisa, Kanga PowerBook G3, Apple IIc, Apple II, Spare parts, etc.

Offline

#71 2016-05-20 22:15:25

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

I believe the IIfx is the only other common II-era machine besides the SE/30 that lacks built-in HD20 support. Do you have a Dougg3 SIMM? You could try programming a stock IIsi ROM image to it, and put it in your IIfx and see what happens. Or put it in an SE/30 if you have one.

I did some more testing with the HD20 Emu and the SE/30, and it works fine for reading the HD20. If the HD20 disk is read-only, then everything works great. But if the HD20 disk is writable, it quickly gets corrupted after about 30 seconds of use, and disk errors pop up everywhere. With the same ROM SIMM and Floppy Emu hardware on a IIci or IIsi, the HD20 works fine. I'm going to get a second SE/30 to help determine if this is a hardware problem with the first SE/30, or a more general incompatibility problem of some kind.

Offline

#72 2016-05-21 18:39:31

bigmessowires
Member
Registered: 2014-05-15
Posts: 183

Re: Mac II series ROM hacking testing

Good news, I almost have closure on this SE/30 mystery. Thanks to olePigeon, I now have a second SE/30 to play with.

  • With a IIsi ROM or BMOW custom ROM, HD20 on the new SE/30 works great. Reading, writing, booting, mounting... no problems or disk corruption. So the old SE/30 must have a hardware problem, but it's a mysterious one. All the HD20 I/O is checksummed, so if it were randomly flipping bits somewhere I should see an error reported by the Floppy Emu. And I've found no other evidence of hardware problems. Very strange, but it's still good news. I'd rather have a mystery problem affecting one SE/30 machine than a mystery problem affecting all SE/30's with a non-standard ROM.

  • The bootup problems with certain RAM are consistent between the old and new SE/30, so I think that's a real issue. There's a particular set of 4 x 4MB unmatched SIMMs that work fine in both SE/30's with their stock ROMs, but fail at bootup with a IIsi ROM or a BMOW custom ROM. It shows garbage on the video, and I think it's playing the chimes of death, though my speaker isn't connected. So I'm back to my theory that the IIsi-derived ROM is somehow less tolerant than the SE/30 ROM of particular RAM type/speed mixes.

This is the problematic set of 4 x 4MB:
USA SY 9304, 4P1004JDJ-M8
81C1000A-70, 9418 F64Z, Malaysia PJ
AAA1M300J-08, NMBS 2091
TI-80, TMS4C1024DJ, HEI 114A R

The only thing undermining this theory is that Techknight's said he's seen some SE/30's boot the dougg3 ROM and some not, even when he's using the exact same RAM in both of them. So maybe there are other factors at work too.

Offline

Board footer

About ThinkClassic

ThinkClassic specialises in the use, maintenance, repair, restoration and modification of vintage computers and peripherals.