Print

[rayarachelian]

If you have a large logic analyzer it may help you debug futher

I have a USB logic analyzer, but the vast amount of data that it produces when connected to the Lisa makes it impossible to go through it all. The text and CSV files that it outputs are so large that I can't open them in any program without it crashing.

Yeah, those HP ones understand 650x and 68000 busses and can show you what the CPU is doing, what address it accesses, read/write and value and timing.
With a USB one, you'd need specialized software to analyze it.

[AlexTheCat123]

So I played around with the logic analyser some more and it revealed something interesting. I hooked it up to the BD0-BD7 lines as well as the address lines A1-A7 that feed into the muxes that allow the CPU to access the shared floppy RAM. I also connected it to the select signal (S) on the muxes so that I can see when they're allowing the CPU to address the floppy RAM. I then went into service mode and tried to read the I/O ROM version number from location FCC031. Interestingly, every time it successfully read the version number, the select signal went low and stayed low for around 600ns, but each time it failed the select signal only stayed low for around 100ns. You can see this behavior in the two screenshots that I've attached. What could be causing this?

[rayarachelian]

So I played around with the logic analyser some more and it revealed something interesting. I hooked it up to the BD0-BD7 lines as well as the address lines A1-A7 that feed into the muxes that allow the CPU to access the shared floppy RAM. I also connected it to the select signal (S) on the muxes so that I can see when they're allowing the CPU to address the floppy RAM. I then went into service mode and tried to read the I/O ROM version number from location FCC031. Interestingly, every time it successfully read the version number, the select signal went low and stayed low for around 600ns, but each time it failed the select signal only stayed low for around 100ns. You can see this behavior in the two screenshots that I've attached. What could be causing this?

Looking at: https://lisaem.sunder.net/LisaSchem/Lisa1SysIO4.gif

By Select you mean !CS on pin 8 of the two SRAMs, yes?

Could be R28 (s/b 47Ohm) or U4C LS157 pin 12 (4Y), or it's input lines 4A, 4B.  4A should go to ground, so that's easy to check.

4B goes to Y0 of U5A S139, so one of it's inputs could be the cause.

That could be the XNOR S02 gate at U8A (pin 10), A11, A12 of the 6504, but likely these aren't it as those come from the 6504 itself.

There's also some fun components in the center left of the schematic, right under the SRAMs, there's a transistor Q7 with 3 resistors going to capacitor 28.
Not sure if those can pull down the signal or mess with it, doesn't look like it, but who knows, if they're messed up in some way they might.
(R54 is acting as a pullup from the looks of it, and R26 too)

[Lisa2]

Guys,
Just to clarify, the drawing that Ray linked is actually the 2/10 I/O board with the IWM chip.

I have attached the correct drawing for the 2/5 I/O board.

[rayarachelian]

Just to clarify, the drawing that Ray linked is actually the 2/10 I/O board with the IWM chip.
I have attached the correct drawing for the 2/5 I/O board.

D'Oh! Thanks for that. That came from here: https://lisaem.sunder.net/cgi-bin/bookview2.cgi?zoom=0?page=8?book=6?Go=Go and is labeled "Lisa 1 I/O Board 4" - are the ones labeled "Lisa210Sys" swapped with the "Lisa1" ones?

[AlexTheCat123]

By Select you mean !CS on pin 8 of the two SRAMs, yes?

No. I'm talking about the S signal that feeds into the three LS157 muxes that select between the shared RAM being addressed by the 68000 and the RAM being addressed by the 6504.

The /DTACK signal that comes out of the 8T97 on the schematic also seems to suffer from the same problem where it goes low for a much shorter period of time when the Lisa fails to properly access the shared RAM compared to when it succeeds.

[Lisa2]

D'Oh! Thanks for that. That came from here: https://lisaem.sunder.net/cgi-bin/bookview2.cgi?zoom=0?page=8?book=6?Go=Go and is labeled "Lisa 1 I/O Board 4" - are the ones labeled "Lisa210Sys" swapped with the "Lisa1" ones?

I/O sheet four for both the 2/5 and 2/10 appear to be the same on the sunder site....

[AlexTheCat123]

I think I fixed it! It looks like the 8T97 that buffers the /DTACK signal was flaky and after replacing it the Lisa boots from floppy just fine!. Unfortunately, the ProFile started dying at around the same time. It takes forever to complete the self-test and the LOS installer will fail to initialize it. Would a low-level format be likely to fix this?

[rayarachelian]

I think I fixed it! It looks like the 8T97 that buffers the /DTACK signal was flaky and after replacing it the Lisa boots from floppy just fine!. Unfortunately, the ProFile started dying at around the same time. It takes forever to complete the self-test and the LOS installer will fail to initialize it. Would a low-level format be likely to fix this?

Yeay! Glad you found it.

Maybe. Does NeoWidex tell you how many spared blocks there are? if there are just a few it may help, if there are a lot, it might not help at all (indicating physical damage).
But, if it's not working, and it's not an electrical issue, not much left to lose, right?

Before you do that though, is that VIA ok? Remind me again, before the power issues, was the ProFile working before, or is that an unknown?

So far you've replaced the 8T97, and the 6504, right? that indicates issues on the I/O board, what could have damaged these guys? Any chance the parallel port VIA has gone bad too? Or something from it to the port?

If at least NeoWidex is able to read blocks off the profile, especially the spare table, it should say the name is "PROFILE" at minimum and tell you it's 5MB, etc. and how many spared blocks there are, then there likely isn't a communication issue, but rather a media issue.

[AlexTheCat123]

Maybe. Does NeoWidex tell you how many spared blocks there are?

It says that all 20 (I'm assuming that this in in hex) spares are in use and that there are 1D bad blocks on the disk. Would a low-level format potentially help or is this drive too far gone?

Before you do that though, is that VIA ok?

I know that everything on the I/O board is just fine because my Cameo/Aphid works perfectly with it.

[rayarachelian]

Maybe. Does NeoWidex tell you how many spared blocks there are?

It says that all 20 (I'm assuming that this in in hex) spares are in use and that there are 1D bad blocks on the disk. Would a low-level format potentially help or is this drive too far gone?

...

I know that everything on the I/O board is just fine because my Cameo/Aphid works perfectly with it.

Ok, perfect so the port is known good - that helps a lot. Sadly if the drive has that many bad blocks, you won't be able to install LOS - I think that's the max or near the max.
Since you have the Aphid, you could have installed BLU on the Aphid before you managed to fix the 6504 and 8T97 - then powered off the Aphid and attached the ProFile, but doesn't matter now.

So, yes, your only recourse is the low level format procedure. You'll need to burn the piggyback LLF EEPROM for the Z8 and then you can use  BLU to Low Level Format it.

But be ready in case it won't fix the issue as that's a whole lot of spared blocks and could indicate a lot of physical damage. Let's hope they're only bad because the power supply was a bit weak and it caused the overwrote the formatting markers, etc. In that, case the LLF will absolutely fix the drive. Keeping my fingers crossed for you.

Another thing you might do if the formatting fails is to attempt to replace the ST506 with whatever the 10MB drive is and I think it might need a new EEPROM for the Z8 for that, but that's a bigger project.

Once you're ready to do the LLF, make sure you have really good power, high recommend you plug both Lisa + ProFile in a big UPS just incase. Most likely nothing will go wrong, but you'll want to make sure power is nice and clean and no brownout caused by an HVAC or whatever ruins your day.

[AlexTheCat123]

I'll go ahead and try a low-level format, but there's one problem: I don't have one of the special ROMless Z8s and I can't seem to find any on eBay. @Stepleton, I saw in a previous thread that you had several of these chips that you were spreading around the Lisa community. Do you still have any of these left and if so, how much do they cost?

[stepleton]

Hello, the chips are free as long as they circulate around the community from person to person. So far I've only sent out one of them, and I have several more that I'd be very happy to distribute via post. The only hang-ups these days are that (a) shipping is from the UK so it could take a while unless you live nearby (b) thanks to the new, improved Delta COVID variant and the UK's vaccination scheduling, I'll be aiming to stay out of post offices for about another month or so. There is the option of at-home pickup so that might not matter so much. Feel free to shoot me a PM with your address, and we can at least get the ball rolling.

[AlexTheCat123]

Update: The ROMless Z8 that @Stepleton kindly sent me just arrived in the mail and a low-level format cleared up all of the ProFile issues that I was having! The drive now works perfectly and doesn't seem to have any defects at all!

[compu_85]

Have you checked the reset signal on the CPU board? Sometimes capacitor C8 gets leaky and the NE556 will stick low.

As always, Patrick's hints point in the right direction. My 2/10 had started being "lazy" when turning on, having to run for a minute or so before it would reset. Changing C8 on the cpu board solved the problem.

-J