PowerBook 540c Restoration Project

[jmacz]

It shouldn't be the processor. Most likely the version of 7.1 you're using since I tested only on the 7.5.5 I am running. I will try booting my PB540c under System 7.1 and see if I can reproduce it.

What SCSI emulator are you using? It's possible it might need more delay also. If it didn't crash that means the patch itself is probably fine. There was no "X" through the icon when you booted right?

 

[jmacz]

Repro'd the issue on 7.1.2... it's because of my check to lock out PowerPCs. System 7.1.x doesn't support that gestalt check. I will fix it. The boot icon is crossed out properly on System 7.1.x denoting that the patch did not install. Anyhow, will fix.

 

[jmacz]

Update: continues to work fine on my PB540c running either System 7.5.5 or System 7.1.1 PB update. I cannot test on System 7.1 because my PB540c can't run it (too old).

I provided a modified version to @3lectr1cPPC and that one now installs but seems to only work with application access to the disk, but it crashes when the Finder accesses it after power down. I provided two additional modified versions to him to test to see if those changes work.

Otherwise, I have procured a beat up PowerBook 160 (that works but with an external monitor only since its internal screen is dead). Will try installing System 7.1 on it and see if I can repro.

I'm being told the PowerBook control panel under System 7.1 does not provide HD spin down time configuration. Whereas System 7.1.1 PB update does. I'm wondering if there's some unfinished Power Manager in System 7.1 that is causing the issues. Will explore more once I get this PB 160 up and running.

 

[jmacz]

@3lectr1cPPC 's issue appears to be timing .. the delay wasn't long enough for the BlueSCSI he's using. A longer delay seems to work on his system as well.

I've converted SD Aide into a Control Panel so that the delay is configurable. Will release for people to use once I get some more testing done and @3lectr1cPPC gets a chance to test the control panel version on his 170.

 

[croissantking]

Interestingly, I don't have such wake from sleep/drive spindown issues in either of my 500 series machines and both have BlueSCSI V2s.

I haven't even seen the problem in my PB165/180.

 

[jmacz]

It's interesting. I have a SCSI2SD in my PB 160 and it doesn't have the issue. But it looks like that SCSI2SD is ready very quickly. From debugging it looked like the system software (7.1 through 7.5.5 anyway) restores power to the device and within 500ms sends the IO request that caused it to spin up the device.

On the ZuluSCSI (with the latest firmware and default options) it looks like power on to device being ready takes approximately 1000 - 1800ms. So it misses that first IO request and just sits there waiting for new requests while the system software is waiting for a response from the SCSI device. So that's causing the lock up and it seems to clear after exactly 180 seconds (again on System 7.1 through 7.5.5, I have not tested any other versions).

My guess was that SCSI2SD uses the SD as a raw single device and gets ready very quickly in time to see that first IO request. But the ZuluSCSI uses the filesystem on the SD card and does some work to prepare image files to be exposed as individual SCSI devices. That plus the other features in ZuluSCSI are probably why it's taking so long to get ready and accept its first UI request? That's my working theory anyway.

I don't have a BlueSCSI device but @3lectr1cPPC does in his PowerBook 170 (not sure which version of the BlueSCSI it is) and it also seems to be hanging after spin up. And at least on his copy with his settings, it's taking even longer to get ready (more than 3 seconds). Perhaps you have a newer version? or newer firmware? or less features enabled?

Perhaps a better fix might be to have those devices get ready faster? But this 2-3 second delay after a spin up seems to do the trick.

 

[jmacz]

Also possible newer firmware on the BlueSCSI is actually doing something smarter and getting ready quickly to see IO requests and then holding them until the rest of the device (images, etc) are ready to go?

 

[jmacz]

Memory upgraded, CPU overclocked again, cool.

Now need to figure this out...

 

[MacUp72]

cant say something about the delay thing but I notice a strange behaviour on principle with SD card in my machines ( 3400c/Kanga/2400c). it works, but there is always problems with copying larger files from the pcmcia to the sd hdd or even install. I never have this with regular ide hdds in the machines.

 

[jmacz]

cant say something about the delay thing but I notice a strange behaviour on principle with SD card in my machines ( 3400c/Kanga/2400c). it works, but there is always problems with copying larger files from the pcmcia to the sd hdd or even install. I never have this with regular ide hdds in the machines.

Do those have internal IDE instead of SCSI?

 

[croissantking]

Do those have internal IDE instead of SCSI?

Yes

 

[croissantking]

cant say something about the delay thing but I notice a strange behaviour on principle with SD card in my machines ( 3400c/Kanga/2400c). it works, but there is always problems with copying larger files from the pcmcia to the sd hdd or even install. I never have this with regular ide hdds in the machines.

I would strongly recommend a CF card rather than an SD card. A CF card adapter is just passive/straight through since CF cards work natively on the IDE bus. Far less complicated than an SD card adapter and far more reliable.

 

[jmacz]

There are so many of those IDE/CF/SD adapters, curious whether it might be an issue with one the one you're using @MacUp72 ?

 

[MacUp72]

Do those have internal IDE instead of SCSI?

of course, a very basic design



I used CF/IDE adapters too, but my CF card isnt the best I guess, and correct formatting is key
from normal use I couldnt tell a big difference in usind a SD or CF
I'll do some more testing.

 

[jmacz]

Just received another 540c... it works but the screen has a blue tint when looking straight at it. Hoping it's a ribbon problem. Otherwise the screen is in good shape and I'm hoping to be able to move it into my main 540c given that one's display has backlight bleed.

Also got two batteries with it... one has corroded contacts so that's not a good sign. The other one looks ok from the outside but who knows.

This one is going to be used for parts. I also got some additional parts (it's the main chassis of a PB 500 series but no CPU card, no memory, partial broken keyboard, non working track pad, non working floppy drive. Just motherboard and primarily chassis parts.

 

[jmacz]

The power adapter I recently received works but was working but it had the leaking capacitor fishy smell so I stopped using it. I just opened it up this morning (what a pain in the butt to do while keeping the case nearly intact for a rebuild).

The two 330uF 25V caps were leaking. The rest were not. Luckily I don't see much in terms of damage. Will be ordering caps now.

 

[jmacz]

Recapped the stock power supply today and it works. Yay! Just sealed it up using a combination of superglue and hot glue.

I've got two working power supplies now, one stock, one a custom one with a 16V 4A VMAIN and 16V 1A for VBATT.

 

[jmacz]

I swapped out my Sharp LCD (which had the backlight bleed shown earlier in this thread) for a Hosiden LCD (came with the 540c I got recently). Both are active matrix.

The sharp is the right color from the get go but has the bleed issue.

The hosiden seems to be a little darker and slightly red tinted for the first 10-15 seconds on a cold start but then brightens up and gets back to the right color and is good after that. Weird. But no backlight bleed which is great.

 

[3lectr1cPPC]

The red tint typically indicates a high-hour CCFL tube. If it’s working fine and still bright then I wouldn’t worry too much about it, just know that it could need replacement not too far in the future.

 

[jmacz]

Ah.. well, I guess that will be coming at some point.

Do you know if those CCFL tubes are changeable across the display types/manufacturers or whether they are unique/different per display type?

I guess when the time eventually comes, I can also look at building out an led replacement for that backlight.