• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Backlit Portable: No boot, lines on display

SuperSVGA

Well-known member
I've got a Macintosh Portable M5126 here that doesn't boot. With power applied via the Molex, it will briefly show random lines on the display when a key is pressed, before shutting off again.

I did a full recap of the board but the issue still persists. Sometimes not even the lines show up.

Here's what I've been able to determine so far:
  • Feeding 6.4v directly to the Molex connector, it draws around 50mA.
  • When hitting a key, it draws around 300mA for a few seconds before shutting off again. This is with nothing else connected to the board.
  • While the computer is "on", the clock signal being delivered to the 68000 looks fine.
  • FCLK on the SWIM is always high, even when off. Though this may be normal.
  • The high signal voltage seems a little low at 4.6v.
  • There's no activity at all on the SCSI /CS line.

I tried removing the SWIM, but that hasn't changed the symptoms at all. Now I need to remove the gunk from my desoldering iron again.

The board looks much better than my M5120 board, I haven't been able to find any damaged traces yet.

Anyone have any ideas? I suspect the low voltage may be part of the problem, but I'm not sure what to look at for that. Are there any good points to probe to see where it's failing?
 

moldy

Well-known member
I've got a Macintosh Portable M5126 here that doesn't boot. With power applied via the Molex, it will briefly show random lines on the display when a key is pressed, before shutting off again.
Are any drives (flopy/hard drive) connected? If not, probably makes sense to just keep the mainboard + speaker + screen connected.

Feeding 6.4v directly to the Molex connector, it draws around 50mA.
That seems reasonable as far as I remember.

When hitting a key, it draws around 300mA for a few seconds before shutting off again. This is with nothing else connected to the board.
That also seems reasonable. Does it shut off on its own? Does the current draw return to 50mA then?

There's no activity at all on the SCSI /CS line.
This is imaginable. If it cannot access RAM/ROM for example, it never reaches the point where it should read from any drive.

I tried removing the SWIM, but that hasn't changed the symptoms at all. Now I need to remove the gunk from my desoldering iron again.
I would leave it desoldered for now. During my Portable recaps I confirmed (also thanks to the users of 68kmla) that without SWIM you should reach a clean 50% grey screen - no point in mounting SWIM to make the system more complex now.

Have you seen my thread btw?
I had a somewhat similar case when one address line was hanging. You also have quite a bit of references of SYS_RST and SYS_PWR signals. Generally, the schematics of M5120 are spot on apart from RAM and the power circuit so you could easily check continuity between CPU, GLU, VIA and the PWR_MGR. AFAIR, the CPU talks to the PWR_MGR over VIA.

Probably makes sense to check continuity of all address/data lines between RAM/ROM, the buffers above the CPU and the CPU.

To me, this shutdown signal might sound we also have a problem with the power circuit. One thing that comes to my mind is to check the ADC input of the power manager chip (see the M5120 schematic) - if it's too low, the power manager will just shutdown the system.

SYS_RST is also managed by the PWR_MGR, if something is wrong over there, it might never bring up the system.
 

Johnnya101

Well-known member
I could be missing it, but do you have a good battery installed? You absolutely need a good battery installed for it to work.

-edit sorry, guessing by molex you mean you are bypassing the battery? If so, make sure the contacts in the battery box are pressed down, I think one of them or both are switches.
 

moldy

Well-known member
-edit sorry, guessing by molex you mean you are bypassing the battery? If so, make sure the contacts in the battery box are pressed down, I think one of them or both are switches.
Good point! Also based on @techknight advice it's important to power both positive pins in the 4-pin Molex if you haven't done so, otherwise something might get fried.
 

SuperSVGA

Well-known member
Does it shut off on its own? Does the current draw return to 50mA then?
Yes, it does shut off on its own, and was doing that consistently.
Though for some reason, it's not always doing that anymore. Now it seems like it will usually stay "running". I have no idea why this is though, because I haven't changed anything.

Have you seen my thread btw?
I hadn't seen that one, but I took a look at it.
As for the screen mine doesn't look the same as yours. I don't get any backlight, and no repeatable patterns. Sometimes I get a few horizontal and vertical lines on the left and bottom sides of the screen. A few times I got a black background along with them, and lately it seems like I get nothing at all.
I haven't found any address lines shorted to ground at least.

I've been looking at individual signals, but haven't found much I can trace down to anything.

Here I looked at CS1 and /CS2 respectively at the VIA:
SDS00006.png
While the address strobe at /CS2 appears to be working, CS1 never goes high at any point, so it seems the VIA never gets accessed.

Looking at the main reset and halt lines, it appears almost like it tries to do something, but resets after several seconds and does this in a loop.

Looking at the ROM select lines, it does appear that they are briefly accessed.

I'm probably going to have to see if I can figure out how to work this logic analyzer so I can see everything happening at once. It's going to be a bit tricky since the M5120 schematics appear to be drastically different since several large chips have completely different pinouts.
 

moldy

Well-known member
Looking at the main reset and halt lines, it appears almost like it tries to do something, but resets after several seconds and does this in a loop.
For me the behavior was similar because the address line was shorted to ground. The ROM could not be accessed properly and the system was stuck in a reboot loop.

As far as I remember, it's the PMGR that controls the SYS_RST line so it could very well be that there's something wonky about the power regulation circuit in your Portable.

In my case, the backlight was consistently turning on and the system kept staying on, which could also (my guess) suggest that the power regulation doesn't always start up properly in your case.

I'm probably going to have to see if I can figure out how to work this logic analyzer so I can see everything happening at once.
I can recommend it - the cheap 8 channel LAs are already a good start to understand what's going on on several important pins.

It's going to be a bit tricky since the M5120 schematics appear to be drastically different since several large chips have completely different pinouts.
I had a different impression - while the power circuit is very different and the RAM as well (SRAM vs PSRAM), everything else is pretty much the same. I could trace all signals between SWIM, SCSI, CPU_GLU, VIDEO, CPU, VIA and PMGR based on the M5120 schematic and haven't really seen a difference.
 

SuperSVGA

Well-known member
I had a different impression - while the power circuit is very different and the RAM as well (SRAM vs PSRAM), everything else is pretty much the same. I could trace all signals between SWIM, SCSI, CPU_GLU, VIDEO, CPU, VIA and PMGR based on the M5120 schematic and haven't really seen a difference.
So far the biggest issues for me are the GLU chips. For example what should be A17-19 on U12G is actually just ground. I actually wasn't able to find those lines at all on the M5126 MISC GLU. They are there on the M5120 board though.
Most of the other chips that haven't changed between revisions seem fine however, at least as far as I can tell.

I can recommend it - the cheap 8 channel LAs are already a good start to understand what's going on on several important pins.
Unfortunately mine is a bit more complicated. I haven't quite figured out the advanced triggering, and it runs Windows 2000 which doesn't help.
 

SuperSVGA

Well-known member
Well, that was a pain. I had to setup a Windows XP VM just to get the file to open properly.
Windows XP Professional-2022-05-04-20-03-17.png

It looks like it's resetting every 2us or so.

PMINT is pretty much always high, except for a single instance for 10.35us as in the image. Not sure why that happened across all the time it was on.

A23 is always high, but this might be normal.

I tried to test some other points, but now for some reason I'm back to the original issue. It "turns on" for a brief few seconds, and then immediately goes off. Very little activity seems to happen.
 

Johnnya101

Well-known member
I know you've probably already checked, but are the battery box switch(es) connected still even if you are powering the board directly?

Did you do what Moldy said with powering both pins?

It could very well be something stupid and simple.

Could also be a bad power management chip. That seems to be a popular thing to go. Luckily, a new replacement is being worked on. Seems like it might not be getting enough voltage to turn on, like you said.

Might be worth picking up a new battery and and feeding it the correct voltage and amps for charging to see what happens?
 

SuperSVGA

Well-known member
I know you've probably already checked, but are the battery box switch(es) connected still even if you are powering the board directly?

Did you do what Moldy said with powering both pins?
The battery box isn't connected at all.

I have power connected directly to the two pins.

Could also be a bad power management chip.
I'm assuming the version from the M5120 board is compatible? I have no idea if the one I have is good either, but I could probably swap it in and see.
Maybe I'll try building a replacement hybrid for my M5120 really quick just to see what that does before I start swapping chips.
 

SuperSVGA

Well-known member
Well I swapped the power management chip from the M5120 board, but the results seem to be exactly the same. It's possible both are bad, but it's also possible there's an issue somewhere else.
Also I did test it in the M5120 (after building a new hybrid) before I removed it, and while the symptoms were similar to the M5126 it wasn't quite the same problem. I would get some lines on the screen, then that standard "grey pattern", then it would shut off.

I just can't figure out what makes it switch between staying on and immediately shutting off. I can't seem to get it consistently reproducible.
  1. I took some readings with the logic analyzer. It stayed on.
  2. Later, I attempted to test it again at different points, and it shuts off.
  3. I removed the 74AC244 at U13H - stays on.
  4. I removed the VIA - shuts off.
  5. I replaced the power manager - shuts off.
  6. I put the VIA back on (U13H is still removed) - shuts off
  7. I put U13H back on - shuts off.

Maybe I look into what line is controlling power and see if I can force it to stay on.
 

mg.man

Well-known member
I would get some lines on the screen, then that standard "grey pattern", then it would shut off.
This is a bit of a wild guess... but I saw similar behaviour when trying out @Siliconinsider's Hybrid bypass:

When I tried the above, I was able to boot up, but then got a "low battery" and shutdown. I've not had time to work on my Macs much since, but wondering if there might be some commonality here?

Do you have a "good" battery in the loop? And if so, could it be there's a break somewhere resulting in the Portable's power circuit wizardry not being able to "sense" that you have a good battery? As I said, just thinking out of left field here...
 

mg.man

Well-known member
Oops, just realised you have a backlit = no Hybrid. Still, wondering if it might not be able to sense a "good" battery and is simply shutting down. In addition, have you tried a PRAM reset, and connecting a FloppyEMU? A "static" 'gray screen could be because it's trying to spin up the HDD - as that's set as the boot device... (once again, more left field thoughts...)
 

SuperSVGA

Well-known member
Oops, just realised you have a backlit = no Hybrid. Still, wondering if it might not be able to sense a "good" battery and is simply shutting down. In addition, have you tried a PRAM reset, and connecting a FloppyEMU? A "static" 'gray screen could be because it's trying to spin up the HDD - as that's set as the boot device... (once again, more left field thoughts...)
Unfortunately it's not the backlit model that's giving the grey screen. The SWIM is currently removed as well.

As for power, it's connected directly to a bench power supply, so I'm guessing connecting a battery won't help as it will detect the battery as too low even at a full charge.
 

mg.man

Well-known member
I'm guessing connecting a battery won't help as it will detect the battery as too low even at a full charge.
There are ppl more experienced than me, but I'm pretty sure the Portable needs to "see" a good battery, or it'll shut down. Not sure the "battery sense" is on either of the 5V rails - hence mine also shutting down when only powered via the molex.
 

SuperSVGA

Well-known member
There are ppl more experienced than me, but I'm pretty sure the Portable needs to "see" a good battery, or it'll shut down. Not sure the "battery sense" is on either of the 5V rails - hence mine also shutting down when only powered via the molex.
How does it see the battery though? That molex connector is the only thing that connects to the battery, so it shouldn't be able to differentiate between the battery and the bench supply.
 

Johnnya101

Well-known member
Doesn't matter what's in there, as long as it sees voltage on a couple of those pins, that's good enough.
 

SuperSVGA

Well-known member
Well I haven't made much progress, but it's at least back to the state where it tends to "stay on" for a bit.

I checked most of the address and data lines, and none of them appear to be stuck.

I managed to get a SadMac (00000014 0000CD36) for the first time by forcing SYS_RST high, but it's not consistent so I'm not really sure that means anything. At least I know the system is capable of displaying a proper image.
 

SuperSVGA

Well-known member
I finally got it booting. I got frustrated and removed most of the RAM and logic ICs, cleaned, soldered, reflowed, and then reflowed again. Now it seems to work fine.

Except...

Now I have a strange issue with the 5V. I have to bring the battery voltage up to around 7.1V to get a stable 5V, otherwise at 6.4V battery I'm only getting around 4.3V. Guess I'll have to start looking at signals between the op-amp and Q16.
 

techknight

Well-known member
I finally got it booting. I got frustrated and removed most of the RAM and logic ICs, cleaned, soldered, reflowed, and then reflowed again. Now it seems to work fine.

Except...

Now I have a strange issue with the 5V. I have to bring the battery voltage up to around 7.1V to get a stable 5V, otherwise at 6.4V battery I'm only getting around 4.3V. Guess I'll have to start looking at signals between the op-amp and Q16.

I ran into a backlit portable that had this same regulation problem. Never did figure that one out when all the traces were good.
 
Top