SE/30 video redraw problem

[Y030]

Hi folks, this is my first post here and I’m looking for some help diagnosing yet another video issue with a Mac SE/30 - hoping that someone could point me in the right direction.

Usual story, dug the Mac out of storage to find it didn’t boot, display full of stripes and noise. Removed the caps, cleaned out the electrolyte, dirt and oxide, replaced caps. The board looks great, no visible broken traces and squeaky clean. 

Connected to power, the Mac chimes, POSTs and boots all the way to the Finder normally, the mouse and floppy work. The issue is only the left quarter of the screen is being drawn, as if divided into four columns, leaving horizontal strips across the rest.

In the Finder I can make out the trash and menubar but being continually overwritten in the same quarter screen. It’s as if only 25% the VRAM is being written, but the whole memory is being read back out to the video circuit.

With the RAM and ROM SIMMs out the Mac shows a stable full screen stripe pattern. Swapping out both RAM/ROM with a known good set yields the same results.

I’d appreciate it if anyone could offer any advice on what could be happening. Before I buzz out the entire board or start replacing chips I’d like an idea of what’s going on. 

 

[jimjimx]

That’s the weirdest thing I’ve heard of...

Can you post pictures. 

 

[Y030]

Yes it’s an odd one...

Hmm couldn’t upload images. (Am I doing something wrong?)

So instead here are a couple of links:

During POST

Waiting for boot disk

Board is plugged in using an atx extender but behaves the same plugged directly in.

 

[techknight]

Your missing address lines/signals to the video memory from the main system bus. the framebuffer output is able to see the video memory as its drawing the standard bars as uninitialized framebuffer space. 

So you need to trace all that down, as well as checking the PALs. You could have a bad one. 

 

[Bolle]

Could be bad video address muxes UA8 to UD8 as well.

But I'd start checking traces like techknight already said.

 

[Y030]

Your missing address lines/signals to the video memory from the main system bus. 

Could be bad video address muxes UA8 to UD8 as well.

Thanks guys very helpful. I traced D24-D31 back to the ROM and PDS slot, everything fine there. Also fine were RA0-8 and the other lines RAS/CAS, etc back to the PAL ICs.

Were there any others I should check?

Unless there is something else - then I guess either the 253 mux’s or a bad PAL. Are PAL ICs known to die? Long shot but there a reliable source for replacements? seems unlikely 30+ years later anyone Is burning fresh ones...

Further info, my board is p/n 820-0260-10. It looks like an early revision as it has a factory cut trace to UI6.

PALs are marked:

UE6 341-0637-A 918NNDKB

UE7 341-0688-A 918YZLAB

UG6 341-0635-A 921ZIMOB

UG7 341-0633-A 918NLTJB

UI6 341-0665-A 917NJFCB

 

[Bolle]

Check the CPU address lines to the muxes UA8-UD8. Most of the time one of them is broken.

 

[Y030]

Check the CPU address lines to the muxes UA8-UD8. Most of the time one of them is broken.

OK - tested lines A0-A15 from the mux inputs back to the PDS, all is well there too. Looking more and more likely its an IC at fault. Is there a good method to test the PALs? I was going to get a working SE/30 and compare the signals on each pin, unless there's a more obvious/simple procedure.

Thanks again.

 

[Bolle]

Ok, so to sum up again...

-RA lines mux to VRAM - ok

-CPU address lines to muxes - ok

-data lines obviously ok, because the part of the picture that you can see is ok

-UE8 shift register ok, because you get a correct output of what's in the VRAM

Please also check the VADR signals from the counter at UF8 to the muxes. Those addresses are used when data is read into the shift register.

Next thing I would do is desolder the four muxes and two counters and test them out of circuit.

I personally never had any of the PALs go bad. When ICs go bad it's always been one (or more) of the U*8 gang.

Testing the PALs will be hard. A known good set of PALs to swap in is the only simple and workable way I would say.

BTW: did anyone ever reverse engineer what's inside those PALs? Being able to reproduce them in a GAL would be helpful in the case that one really goes bad.

 

[Y030]

Thanks Bolle. I grabbed some more time on this and checked the VADR lines, all good, so setup a second good se/30 on the bench for comparison. This has what looks like a later revision board and different and socketed PALs. Otherwise, ICs are labelled the same (except the PALs curiously) and at a cursory glance circuit looks identical.

Showing the same ‘waiting for boot’ startup screen i could identify that outputs (P 7&9) of UB8 had a very different reading - on the bad board I could clearly see a clock signal with 50% duty cycle, and on the good board this was a very narrow 0v spike and the remainder of the pulse cycle was high.

Traces were good so I figured UB8 was at fault so pulled it, then found I didn’t have any spares.... replacement is now in the mail.

Will update once I’ve got the UB8 back in.

 

[techknight]

Ive seen the PAL sets go bad a couple times, usually resulting in missing vertical, or horizontal drive. 

If your address lines to and from the muxes are good, chances are the muxes are bad. They are failing left and right in the standard RAM stages too. 

 

[Y030]

My package of deadstock 74 series chips  arrived today from Poland and sure enough, it was UB8. During the wait I got a new fan in there to replace the noisy squirrel cage which made a huge difference, and a SD to SCSI card - yet to be installed.

https://imgur.com/JQnyCgr

Just need to find a better tube without the burn in. Does anyone know if the tubes used in the Plus and or Classic/II were the same or compatible? Spares for these seem more abundant.

Anyway, another Mac liberated! Thanks for your help guys!

 

[bibilit]

the tubes used in the Plus and or Classic/II were the same or compatible

Tubes are the same.

 

[Scott Baret]

The tubes are the same, but the connectors may vary. A Plus CRT will work. Not all Classics will. There's a thread on this forum about compact Mac CRTs--read it over to see the difference between the A and B connectors on the Classic.

All Pluses and SE models (both regular and /30) have the same CRT.