Strange Issue: Quadra 700 won't boot with monitor attached

[jmacz]

Summary: My Quadra 700 will not boot (stuck at no disk icon) when an Apple 16" Color Monitor is plugged into its monitor port.

Quadra 700: 68MB of RAM, loaded on VRAM, ZuluSCSI for internal drive, using on board video, AEKII with standard ADB mouse.

Apple 16" Color Monitor: I just finished recapping it (uh oh), the setup was working before the recap albeit with the monitor having multiple issues (hence the recap). The monitor has a built in DB15 cable.

What happens:

1.) Connect the monitor to the Quadra (built in DB15 cable from the monitor going to the Quadra's on board video).
2.) Power on the monitor (no issues), get a green light and can tell the monitor is powered on.
3.) Power on the Quadra.
4.) Get the boot chime.
5.) Monitor shows the usual gray screen (looks good).
6.) Monitor shows the mouse cursor.
7.) Mouse cursor moves when the mouse is moved, everything looks great.
8.) After a few seconds, the disc icon appears with the flashing "?" mark. It's not booting from the ZuluSCSI.

If I use a different monitor (Apple 13 RGB Monitor) with NO other changes, everything works great. Quadra boots from the ZuluSCSI with no issues. I have tried swapping keyboard and mouse, just in case. No change. I have tried swapping the ZuluSCSI with a standard hard disk. No change.

Basically, it won't work with the Apple 16" and I'm pretty sure it has something to do with the recap OR the reassembly after recap. But I can't for the life of me figure out how plugging in the monitor would impact the Quadra's ability to boot from its internal SCSI device.

How's the monitor?

Before recap, the monitor had a noisy display with very visible scan lines (it would ripple across the screen and you could see it with your bare eyes). It also had a creeping shadow appearing from the left side of the screen about a minute after turning the power on. The monitor was also dim.

After recap, as I'm looking at the gray screen with the question mark disk icon, the noise is gone. The visible scan lines are also gone. The shadow doesn't seem to appear anymore. And the monitor is brighter, looks sharper, and overall looks great!

Investigation

I found a link to the DB15 pinout to see if there's anyway one or more of the pins could be causing the Quadra to be unable to boot from SCSI. The link is here:

Apple Vintage Mac Monitor Connections

Looking at the 15 pins, I would think that most of the pins are working properly given I see a gray, focused screen, with cursor movement which is crisp, with no blue, red, or green tint of any kind. No sync issues. No geometry issues. I'm not sure how any of these pins could be causing the SCSI issue.

Any thoughts?

 

[volvo242gt]

Want to try a Nubus video card and see if the problem persists? Maybe also try the monitor on the IIci and see if the problem moves to that machine.

 

[Phipli]

Try removing the extra VRAM. The bigger screen will be using more VRAM.

 

[RetroTheryboy]

Is your zuluscsi bus powered ? Try powering it separately.
the 16" uses extended sense codes as opposed to the original sense pin system on the 13" , maybe its some weird interaction with 5V and pulldowns and diodes . Using a separate 5V could give a clue.

 

[jmacz]

Thanks for the suggestions! Will give it a try today.

 

[jmacz]

Quadra 700 Additional Tests

• Tried with a ZuluSCSI RP2040 (previously was a ZuluSCSI 1.1) - didn't help, same blinking question mark disk icon.
• Tried with a Nubus video card - didn't help, but this time no video at all. Nothing shows up.
• Tried with ZuluSCSI with aux power instead of termination power - didn't help, same blinking question mark disk icon.
• I did not try with VRAM reduction as this particular monitor was working just fine with this Quadra 700 prior to the recapping.

Macintosh IIci Additional Tests

• Tried it with my Iici (which has a ZuluSCSI 1.1 inside).
• Tried with on board video - doesn't work, no video at all.
• Tried with a Nubus video card - doesn't work, no video at all.

Was looking at some of Apple's hardware docs and noticed that at least on the Iici, the RBV which takes input from the video connector also is responsible for handling SCSI IO interrupts. Going to look at this a bit more.

Going to take the monitor apart and look more at the video board (which is where the monitor cabling goes into) and trace the individual pins to see if anything got messed up.

 

[olePigeon]

I assume you have a good battery installed? On the Centris 610 I was fixing up, without the battery the monitor would remain black no matter what unless I zapped PRAM and booted with extensions off. It absolutely needed a good battery to function.

 

[jmacz]

I assume you have a good battery installed? On the Centris 610 I was fixing up, without the battery the monitor would remain black no matter what unless I zapped PRAM and booted with extensions off. It absolutely needed a good battery to function.

Yup the battery is good, on both systems. And it's still coming up just fine using the 13" monitor.

 

[jmacz]

Some more notes.

Apple DB 15 Pinout

1. Red Ground
2. Red Video
3. Composite Sync (/CSYNC)
4. Sense 0
5. Green Video
6. Green Ground
7. Sense 1
8. Not Used
9. Blue Video
10. Sense 2
11. Composite and Vertical Sync Ground
12. Vertical Sync (/VSYNC)
13. Blue Ground
14. Horizontal Sync Ground
15. Horizontal Sync (/HSYNC)

The Apple 16" Monitor has a fixed DB15 cable attached to it. One side is DB15 utilizing the pinouts above. The other side (inside the monitor which attaches to the video board) is a 10pin connector.

Apple 16" Video Cable Internal-Side 10-Pin Connector Pinout (not sure about pin order on this connector, but based on continuity testing using a multimeter to the external DB15 end of the cable)

1. Ground
2. Pin-12 Vertical Sync (/VSYNC)
3. Ground
4. Pin-15 Horizontal Sync (/HSYNC)
5. Ground
6. Pin-9 Blue Video
7. Ground
8. Pin-5 Green Video
9. Ground
10. Pin-2 Red Video

So the 5 pins that are not represented on the internal 10pin connector side of the video cable are:

• Pin-3 Composite Sync (/CSYNC)
• Pin-4 Sense 0
• Pin-7 Sense 1
• Pin-8 Not used
• Pin-10 Sense 2

We can ignore pin 8. For the sense pins, this doesn't need representation on the 10pin connector because it's hardwired inside the cable to identify the monitor. A great explanation of the sense codes is found here: https://wiki.preterhuman.net/Macintosh_Monitor_Sense_Codes

For my Apple 16" monitor, it is using the extended type 7 sense codes. The cable itself has pin 4 and pin 10 tied together (checked via a continuity test). And all three sense pins are not tied to ground (floating). So via the original sense code scheme, all three sense pins are floating meaning you need to refer to the extended type 7 sense codes. Given my cable has pin 4 (sense 0) and pin 10 (sense 2) tied together, the result of the dynamic assertion polling is 10 11 01. Looking this up in that article linked above, that extended type 7 sense code identifies this monitor as Apple's 16" monitor. So these sense codes look correct here and again are hardwired so don't need to be represented in the 10pin connector.

So really the only pin missing via this 10-pin connector is pin-3 (/CSYNC). I don't believe this is used if both the VSYNC and HSYNC pins are used?

Based on that, the only 5 pins I need to look at are:

• Pin-2 Red Video
• Pin-5 Green Video
• Pin-9 Green Video
• Pin-12 Vertical Sync (/VSYNC)
• Pin-15 Horizontal Sync (/HSYNC)

Still don't see how these 5 above could be impacting the SCSI bus. I think these are all outbound signals (from computer to the monitor) so this doesn't make sense.

 

[jmacz]

My test with my Macintosh Iici's onboard video was doomed to fail from the start because duh, the Iici's onboard video doesn't support greater than 640x480 resolution.

My test with the nubus card (e-machines color page xl) I don't think worked because the switch on it is probably set to 640x480. I don't have docs for that switch but will probably try and see if the settings for the e-machines futura cards is the same (which I have the docs for). I'll have to test that out.

I should go back and try @Phipli 's suggestion of playing with the amount of VRAM. I discounted it because this monitor had worked before on this same Quadra. But who knows.

My test with a normal hard disk didn't work because it was formatted ... duh... I forgot I had erased it. So need to try that again. But before that, going to try a more basic test of booting from the floppy (another duh).

 

[cheesestraws]

Agree with @Phipli that in your position I would probably initially be peering suspiciously at the VRAM. This isn't an issue I've had myself, though, so that's a bit of a guess.

 

[jmacz]

I just tested via floppy boot and it worked. Booted using System 7 disk tools and I'm sitting on the desktop now. So for whatever reason it's just SCSI that doesn't seem to work when using this monitor.

Moving to VRAM next.

 

[jmacz]

Mucked around with the VRAM, trying different combinations, and it did not help.

BUT...

I took out the PRAM battery (CR2032 via MeowToast), let it sit, then put in a standard 1/2 AA battery, and somehow magically it worked! Booted from the ZuluSCSI, proper resolution, millions of colors, it's working. Then I shut it down and put the CR2032 back in and it's still working. Huh??

I had previously zapped the PRAM and it didn't help. Is pulling the battery all together different from zapping the PRAM? I'm confused.

 

[jmacz]

Something's still not exactly right ... given I pulled the PRAM battery, etc, I had to reset the date/time, etc. One of the settings I just changed was to enable 32 bit memory addressing. After that it won't boot again. Now instead of showing me a question mark disk icon, it shows me the happy Mac face and then locks up there. If I pull the battery (ie. lose all my settings again, 32bit addressing disabled), then it boots.

So I tried removing the CR2032/meowtoast and just leaving the 1/2 AA battery instead. Rebooted, worked. Changed to 32bit address enabled, rebooted, still works. Tried disconnecting the power supply, waiting for a bit, rebooting, still works and my settings are retained with 32bit enabled.

I'm at a loss. Could be a pure coincidence but what I'm seeing is that with a 13" monitor, either battery, with 32bit enabled, works fine. But with the 16" monitor, seems like the CR2032 works only with 32bit disabled whereas the 1/2 AA works in both modes with the 16". There's a voltage difference between these two batteries (3V on the 2032 vs 3.6V on the 1/2 AA).

In any case, it doesn't seem to be related to the recapping I did on the monitor. Seems like something with my Quadra when more VRAM is utilized? Right now, everything seems stable with the 1/2 AA.

 

[Phipli]

Have you tried booting with just the soldered RAM?

You might need to hold shift down to stop it loading extensions because it won't have enough RAM for them.

 

[lobust]

Which system version are you running? Does the Quadra 700 actually support 24bit addressing at all? I didn't think it did...

 

[Phipli]

Which system version are you running? Does the Quadra 700 actually support 24bit addressing at all? I didn't think it did...

They'll support 24bit. My 650 does, and so does the 475, as long as you're not running Mac OS 7.6 or newer

 

[lobust]

They'll support 24bit. My 650 does, and so does the 475, as long as you're not running Mac OS 7.6 or newer

Ah ok, makes sense that I have rarely seen it as I tend to use 7.6 or 8.1 on my quadras

 

[jmacz]

The Memory control panel defaults to 32bit addressing off after zapping the PRAM.

The system is stable and working great with the monitor after putting the regular 1/2 AA 3.6V battery in.

With the 3V CR2032 it wasn’t stable.

 

[jmacz]

Weird.. for fun, I decided to try different VRAM configurations. Pulled all of the VRAM out (leaving just the 512K fixed on the motherboard) and again I could not get past the blinking disk icon. I had to pull the PRAM battery, let the system drain, reinstall the PRAM battery, to get it to boot again. And was rebooting fine after that running at 832x624 @ 256 colors. Then tried 1MB of VRAM (the 512K on the board and two of the 256K modules). Again, got stuck with a blinking icon until I pulled the PRAM battery and go through that again. It's like every time I change the VRAM configuration or move from a 13" to a 16", I have to pull the PRAM.

While in the blinking disk icon stuck state, I did try seeing if the ZuluSCSI could tell me anything. I deleted the log file from the SD card. Then booted and waited for it to get to the blinking disk icon. Then powered the computer down and pulled the SD card. It had created a log file and looking at it, the ZuluSCSI did initialize, find all the drive images, and completed init. So the SCSI system is getting power and the ZuluSCSI is working just the Mac is choosing now to start from it. So I tried the various key combinations to get the system to boot from a particular SCSI ID and no go.

So weird.. never had a single issue with this Quadra at all until attaching this 16" monitor to it. If I use a 13" again, no issues. But I can't see how it's the monitor so the higher resolution must be exposing something wrong with this Quadra that I wasn't seeing at 640x480.