Supermac Spectrum/24 Series III and Series IV Issues

[jmacz]

Recently procured two cards, a Supermac Spectrum/24 Series III + Daughterboard and a Supermac Spectrum/24 Series IV. Both are having some issues and would love to get some second opinions on what else I can try to get them to work.


Supermac Spectrum/24 Series IV

Setup: Quadra 700 and IIci, two Apple 13" RGB Monitors, all machines and monitors have no issues and run fine with onboard video and other video cards.
Card Details: Arrived with ROM version 1.6, card has 14.32MHz and 30.24MHz oscillators.

Issue: At 640x480 @ 24bit color, the display is very feint and very blue in color. At any other bit depth including 8bit color, no image, screen is dark.

Debugging Efforts:

• ROM Versions: tried ROM version 1.6 which came with the card, also tried ROM version 3.0 (got it from this forum in another thread), both ROM versions exhibit the same issue. Looked at the checksum on the v1.6 ROM and compared against a ROM image of v1.6 from this forum and they match so no corruption.
• Extensions: disabled all extensions and control panels except the Supermac driver and that didn't help.
• Boot Time Configuration: held the Option key down during boot and tried to have the card cycle through the various configurations (the press space bar to select thing) and that didn't help.
• PRAM Reset: reset the PRAM multiple times via command-option-P-R and that didn't help.
• Multiple Machines: tried on both my Quadra 700 and a IIci, and same problems on both.
• Multiple Monitors: tried two different Apple 13" RGB Monitors and same problems on both.
• Multiple DB15 Cables: tried two different DB15 video cables and same problems on both.
• Driver Versions: tried SuperVideo driver versions 2.07, 2.49, 2.74, 2.75, 3.02, 3.1 and same issue on all of them.
• 24bit/32bit Modes: tried in both 24bit and 32bit address modes, didn't help.

Next step I will try is to attach to a multimeter or oscilloscope and see what's coming out of the various pins in the DB15. I'm guessing something's wrong with the red/green video signal.

Some Pictures:



The above is what it looks like at 24bit. Very blue. It's also not entirely in focus, you can see bleeding in the text. If I change to 8bit, 4bit, B&W, I get no image at all.



You can see above it's running ROM version 3.0 (again, also tried v1.6). The version 3.0 was for the Spectrum/24 Series V but the other thread on this forum about ROM versions for the SuperMac cards showed that the v3.0 ROM does work on the Series IV and it certainly looks like it does. And the right monitor is selected. Also tried both Custom Gamma and Uncorrected Gamma. Doesn't make a difference.



The card. The ROM is my own (with 3.0 on it). Again, both 3.0 and 1.6 have the same issue. The card looks pristine except the white component near the DB15 port. The board has "F12A" printed next to it, and the component itself is labeled "1 1/2 A 125V F 459". Not sure what it is, but there's some board markings near it and it doesn't look to be in good shape. Not sure what it is though. Need to look it up. If someone who has the Series IV can look at their card and tell me what they see, that would be great.



You can see it's not in good shape (pointed at by the red arrow). What's weird is if I go online and find a close up picture of a series IV, I see something different in those pictures:




So not sure what to make of it. Of course the card I have is a revision B and the card in the picture above is a revision C.



Back of the card looks clean.


Supermac Spectrum/24 Series III + Daughterboard

Setup: Quadra 700 and IIci, two Apple 13" RGB Monitors, all machines and monitors have no issues and run fine with onboard video and other video cards.
Card Details: Arrived with ROM version 1.0, card has 30.24MHz, 64.00MHz, and 80.00MHz oscillators.

Issue: Running at 640x480, and at B&W, 4bit, and 8bit color, the card works great. Seems fast/accelerated. And no issues. But if I flip to 24bit color, the display stops refreshing. The background of windows gets color distorted and if I move windows around, the screen doesn't update. The cursor still shows and can be seen moving, but the windows don't update on screen.

Debugging Efforts:

• ROM Versions: tried ROM version 1.0 which came with the card, also tried ROM version 1.1 (got it from this forum in another thread), both ROM versions exhibit the same issue.
• Extensions: disabled all extensions and control panels except the Supermac driver and that didn't help.
• Boot Time Configuration: held the Option key down during boot and tried to have the card cycle through the various configurations (the press space bar to select thing) and that didn't help.
• PRAM Reset: reset the PRAM multiple times via command-option-P-R and that didn't help.
• Multiple Machines: tried on both my Quadra 700 and a IIci, and same problems on both.
• Multiple Monitors: tried two different Apple 13" RGB Monitors and same problems on both.
• Multiple DB15 Cables: tried two different DB15 video cables and same problems on both.
• Driver Versions: tried SuperVideo driver versions 2.07, 2.49, 2.74, 2.75 and same issue on all of them.
• 24bit/32bit Modes: tried in both 24bit and 32bit address modes, didn't help.
• Daughterboard: tried it with the daughterboard connected and disconnected, didn't help.

Not sure what else to do on this one.

Some Pictures:



The above picture shows it's a Spectrum/24 series III and this picture was from when it's running the 1.1 ROM. The SuperVideo is v2.7.5 in this picture. As you can see it's fine and running at 8bit (256 colors). Once I click on Millions, I see the below:



The background changes, window title bar disappears. And if I click and move the window, the screen doesn't update. The window is moved, but the display didn't update so the window is no longer where the image suggests it is. I have to carefully click around the screen until I "find" the window and can switch it back to 256 by guessing where the "256" color option is.



The above is to show that the proper monitor has been selected for my Apple 13" RGB. The screenshot also shows the oscillators installed.



Card is super clean. The picture shows my ROM 1.1 installed in it.



The back is also clean. Card looks great. Don't see any trace issues, no corrosion, etc.

Again, not sure what else to do on this one. If anyone has additional suggestions, I would welcome it. Probably going to try tracing all the connections and checking the solder joints for connectivity next.

 

[jmacz]

The card looks pristine except the white component near the DB15 port. The board has "F12A" printed next to it, and the component itself is labeled "1 1/2 A 125V F 459". Not sure what it is, but there's some board markings near it and it doesn't look to be in good shape. Not sure what it is though. Need to look it up. If someone who has the Series IV can look at their card and tell me what they see, that would be great.

Ok, so looks like this is a fuse. A surface mount 125V 1.5A fuse. And there's connectivity from one side to the other. So looks like it's intact. But still bothered by the area around it... doesn't look like it's in good condition.

 

[Phipli]

"F12A"

Not quite, that's "F1 : 2A" not "F12A".

Fuse number 1, rating : 2 Amps.

Looks like it blew and someone replaced it with a 1.5A fuse. I suspect it is related to your issue, but not the cause. Perhaps an old repair attempt.

 

[jmacz]

Not quite, that's "F1 : 2A" not "F12A".

Fuse number 1, rating : 2 Amps.

Looks like it blew and someone replaced it wit a 1.5A fuse. I suspect it is related to your issue, but not the cause. Perhaps an old repair attempt.

Ok got it. The current fuse (the incorrect 1.5A one) isn't blown so it's not my current issue... but if this is the incorrect replacement for one that did blow, then I guess something ugly must have happened to this board at some point.

 

[Phipli]

Ok got it. The current fuse (the incorrect 1.5A one) isn't blown so it's not my current issue... but if this is the incorrect replacement for one that did blow, then I guess something ugly must have happened to this board at some point.

Something bad enough to take out a 2A fuse was likely to have caused additional damage, yeah.

 

[jmacz]

Something bad enough to take out a 2A fuse was likely to have caused additional damage, yeah.

Yeah, not looking great... the north side of that F1 fuse touches a set of resistors, capacitors, and two pins on the nubus slot. The south side of that F1 fuse goes to Pin 8 on the DB15 (which is supposed to be not connected) and a resistor. The other side of that resistor touches 4 of the large chips on the video card. Three of those four chips are the Supermac label BSR03 chips. The fourth chip is on the other side of board (has a label on top which I haven't peeled yet so not sure what it is exactly).

 

[jmacz]

I am guessing there's three of those BSR03 custom chips (bit shift registers) as there's one for each color? If so, the fact that I'm seeing only blue probably means two of them are damaged somehow? And given there's some focus/bleeding issues, something else is probably wrong as well. So this Spectrum/24 series IV card is probably shot as you can't get replacements if indeed those chips are toast. Might do some more work to prove it but looks like potentially a lost cause.

For the Spectrum/24 series III card though... I'm curious why just the 24bit color has issues? Maybe some of the memory chips are not in a good state?

 

[jmacz]

Played with the Spectrum/24 Series III card some more today.

Millions of colors works only if I disable acceleration or remove the daughterboard (which effectively disabled acceleration). But if I enable acceleration with the daughterboard in place, only 256 colors works... millions of colors shows the issue shown in the pictures earlier in this thread. So something's wrong with acceleration. I don't understand it.

So basically:

• BW, 16 Colors, 256 Colors - works with acceleration enabled/disabled.
• Millions of Colors - works only with acceleration disabled.

I have checked the daughterboard and connectors on the main board for issues. It looks pristine. Don't have a way of testing the chips on the daughterboard but all the resistors check out. I don't see any solder joint issues. Thing looks fine.

 

[jmacz]

Well, finally some good news on the Spectrum/24 Series III. Given it wasn't working with acceleration enabled, I figured it had to be something on the daughterboard. I used a hot air rework gun to heat up all the solder joints on the daughterboard as well as the two connectors on the main board. After that, tried it again and surprisingly things started working. I can get millions of colors with acceleration enabled now, with no display issues. Also ran some benchmarks to confirm it's faster than with acceleration disabled.

As far as the Spectrum/24 Series IV, it's definitely not working and has lots of issues. I was going to try replacing the fuse but given there's connectivity with the incorrect fuse on it right now, I can't see how the fuse could be contributing currently. Oh well.

 

[jmacz]

Need a favor... if you have a Spectrum/24 Series IV card, can you put up a high resolution photo of the bottom of the card, and a close up of the bottom near the video port? Thanks.

 

[David Cook]

Need a favor... if you have a Spectrum/24 Series IV card, can you put up a high resolution photo of the bottom of the card, and a close up of the bottom near the video port? Thanks.

I realized after I took the photos that my board needs cleaning. Let me know if you need any specific shots other than these.

 

[jmacz]

Weird, I have a REV B board but my R32 and R33 are populated. I replaced the fuse with a proper 2A fuse, and as expected, still no go. I had thought previously that the BSR03 chips might have gotten hit by whatever took out the original fuse but I think I might be wrong. R34 isn't populated on my board (it isn't there on your board either) so I don't think the fuse is in the path to those chips. But I did however find that two of the resistors on the bottom of the board R44 and R52 seem out of spec... they should be 10K ohm resistors but are reading super high. R52 does connect to the fuse. R19 also has a high reading. Going to try replacing those and see if it's any better.

Really appreciate the pictures @David Cook !

 

[David Cook]

Weird, I have a REV B board but my R32 and R33 are populated. I replaced the fuse with a proper 2A fuse, and as expected, still no go. I had thought previously that the BSR03 chips might have gotten hit by whatever took out the original fuse but I think I might be wrong. R34 isn't populated on my board (it isn't there on your board either) so I don't think the fuse is in the path to those chips. But I did however find that two of the resistors on the bottom of the board R44 and R52 seem out of spec... they should be 10K ohm resistors but are reading super high. R52 does connect to the fuse. R19 also has a high reading. Going to try replacing those and see if it's any better.

Really appreciate the pictures @David Cook !

Happy to help.

In place, my R44, R52, and R19 are all spot on 10K. So, you're definitely right to replace those.

 

[jmacz]

Replaced R44, R52, and R19, and they are now spot on 10K. But the card still has issues. Only getting an image with millions of colors, and it’s primarily blue. I checked out the video signals for RGB and as expected, red (DB15 pin 2) is not sending anything. Blue (DB15 pin 9) is sending the strongest signal. Green (DB15 pin 5) is sending but weak. Red when checking for continuity is grounded. When measuring the peaks, red is at 0V, blue is at 0.26V, and green is at 0.11V. So something’s wrong with the red signal, and green doesn’t seem very healthy either. Every so often the screen gets a bit brighter and more teal, and sure enough, the green pin shoots up to 0.22V but then drops back down to 0.11V. Keep in mind the image being displayed is an empty desktop with a gray background so I would have expected all three signals to be close.

I’ve checked all the surface mounted resistors and they all seem to be in spec now. I still need to use an esr meter to check the surface mounted capacitors.

D1-D6 diodes near the DB15 port, I need to figure out their specs. D2/4/6 in particular seem tied to the red/green/blue signal pins (D2 tied to pin 2 red, D6 tied to pin 9 blue, D4 tied to pin 5 green). D2 doesn’t seem to be behaving the same as D4 and D6, but I am not positive yet.

Need to trace some more tomorrow to find out whether the chip generating the signal is not sending anything in which case the card is probably shot since I am not going to be able to replace the BSR03 chips. Keeping my fingers crossed that some component along the way is the issue.

 

[Phipli]

I’ve checked all the surface mounted resistors and they all seem to be in spec now. I still need to use an esr meter to check the surface mounted capacitors.

Start by checking the caps (and diodes) for shorts.

 

[MacOSMonkey]

Edit -- oh -- I didn't see the replies above...I see that you reflowed...but -- also make sure you didn't change systems or reset the slot, as below. I recall it being a software issue. But...in reading about your debugging, it sounds like you already tried software/slot stuff. So, I guess ignore my comment - I just deleted it from this post.

I don't understand the debug about trying it with and without the daughtercard. I have seen this problem in the past and it was a software/addressing-related acceleration issue. So, if you pull the daughtercard, acceleration no longer works. Also, the Spec24 III mostly accelerates 24-bit mode. Any 8-bit acceleration was minor and there were some general software improvements to QD32 that helped.

There was a 32-bit addressing INIT -- might have been related to that. Anyway -- I will try to see what I remember.

 

[MacOSMonkey]

On the Thunder IV, the final output is coming from the ADV473, which I think was a cheaper, drop-in replacement for the Bt473 -- 24-bit CLUT/DAC. The video signal should be sync on green (for the old trinitrons and hitachi clones), so if there is no green output, then there won't be any image. Check the output on the ADV473...and that is something you may be able to replace. The DAC may be blown/ESD'd from the connector side.

 

[Phipli]

On the Thunder IV, the final output is coming from the ADV473, which I think was a cheaper, drop-in replacement for the Bt473 -- 24-bit CLUT/DAC. The video signal should be sync on green (for the old trinitrons and hitachi clones), so if there is no green output, then there won't be any image. Check the output on the ADV473...and that is something you may be able to replace. The DAC may be blown/ESD'd from the connector side.

Radius used them on their PrecisionColors. Got one in the post over the weekend for a project to upgrade a PrecisionColor 24xk into a PrecisionColor 24x (80MHz RAMDAC to 110MHz).

Radius used a mix of Analog Devices and Bt parts.

 

[jmacz]

Start by checking the caps (and diodes) for shorts.

Yup, will do.

 

[jmacz]

Edit -- oh -- I didn't see the replies above...I see that you reflowed...but -- also make sure you didn't change systems or reset the slot, as below. I recall it being a software issue. But...in reading about your debugging, it sounds like you already tried software/slot stuff. So, I guess ignore my comment - I just deleted it from this post.

I don't understand the debug about trying it with and without the daughtercard. I have seen this problem in the past and it was a software/addressing-related acceleration issue. So, if you pull the daughtercard, acceleration no longer works. Also, the Spec24 III mostly accelerates 24-bit mode. Any 8-bit acceleration was minor and there were some general software improvements to QD32 that helped.

There was a 32-bit addressing INIT -- might have been related to that. Anyway -- I will try to see what I remember.

Sorry, I should have made two separate threads, one for my issue with the III and one for my issues with the IV.

Regarding the III and the daughter card, I was trying to rule out the main board vs the daughter card, although in hindsight, all it did was prove that the issue was with acceleration. The card finally started working without redraw/erase issues after I reflowed the daughter card (and the two daughter card connectors on the main board). Was working great for the last week until yesterday it started happening again (albeit now nearly as bad as before). Previously, I couldn't get any part of the screen to refresh/erase/redraw properly after switching to 24bit. Now, most of the screen is fine but I get random artifacts after a cold boot, and eventually it goes away. Happens in both 8 bit and 24 bit modes now (whereas it was 24bit only before). Again, only when acceleration is enabled. Going to try reflowing the entire mainboard as well.

Regarding the INIT, I think I read in one of the threads that there were some artifact issues due to A/ROSE. But I don't have that particular one installed in the computer this III card is in.