• 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.

Color Classic VGA Mod. the CORRECT way

Franklinstein

Well-known member
Long-term project is an attempt at cramming a DOS-compatible board from a Performa 630 into a Mystic '040. I know it's missing the video port on the motherboard, but... I feel like there's a way this can be done without actually doing a full Takky mod. Everything else is there... it just feels so possible. Maybe someone has already done it? I couldn't find anything online. I've already got a Performa 630 DOS-compatible motherboard ready (purchased on eBay a couple of years ago when I first thought about giving this a try). Feels like a good project to take on.
The IDE-based boards like the 630 will not swap into a CC variant (or indeed any other AIO models except a 58x). For that, you'll need to do the Takky mod. While you can basically run the Takky mod off of the stock CC analog board, it's not recommended: the 68040 draws more power than the 68030, which is stressful enough; adding a high-res mod and expansion cards (let alone a DOS card) could cause stability and longevity concerns. Usually major mods like that use an analog board from a 5xx because they're already set to run high-res and can put out a bit more power. They need to be physically cut down, though, because they're too wide to fit in a CC chassis as-is. It's major surgery but props if you can do it.

Does the 630 DOS card fit in a 575 logic board? That may be interesting if it does. Also it would swap directly into a CC without a Takky mod, though you'd still need to do something about the resolution (either high-res mod or ResEdit hack). I'd still be concerned about the power draw of the DOS card; the 486 wasn't known to be a low-power chip.

 

liroku

New member
Ah, but the 630 DOS card is a really bizarre creature, much different than most of Apple's other forays into the territory; it plugs directly into the '040 socket. The '040 and the 486 reside on the same daughterboard, with additional connectors plugging into the PDS (interestingly, the PDS connection does not include the extended pins) and the small, odd video port near the motherboard's main connection pins. This small port is the only one the 575 board is missing, and this is the reason why I think I might be able to make the DOS card work.

Power draw is definitely a concern.

I'm not really big on the Takky mods, 'cause I don't wanna do that much work and I cringe at the thought of cutting up a CC chassis. I avoided the screen res mod simply because I didn't want to cut a trace! I'm probably just a weirdo. That's the reason, though, that I'd wanted to try the 630 DOS surgery on an earlier '040 board, rather than just do the full Takky mod.

Here's a pic of the 630 board (not mine; I found this online, but I'll get some close-up pics up tomorrow if I've got time). You can see the '040 next to the pins where the actual CPU socket is located. The small video connector is on the right in the foreground, blue on the daughtercard, brown on the motherboard. Some Quadras (not the DOS compatibles) seemed to use this port as a video-out port (a ribbon cable runs from it directly to the video-out port at the back of the chassis).

630_DOS_Compat.jpg

 

Franklinstein

Well-known member
I don't like major mods to something unless I have multiples or one is in bad shape/dead anyway. I bought a CC II where it turns out someone had done an improper high-res mod involving the trace cut/transformer jumper technique. I finished the mod but apparently I need to try another capacitor value or two because my screen is still too wide. Of course, being a CC II, it has a later analog board so the upgrade components and locations listed here don't align properly.

The 63x's is the only DOS card that connects in that manner, and as such is the only one I really wanted to find (I got a Performa 640 DOS). Interestingly, the 68040-based compact Mac PowerPC upgrade card connects in the same way, though I sincerely doubt you could run them concurrently.

The connector on the logic board is supposed to be for external video out, but it may be software-switchable to do the DOS card video input instead (I'm not 100% sure how these things work; it's not terribly detailed in the Dev Note). If so, you can probably modify an external video kit so that it permanently displays your DOS environment on a separate monitor in the same way that the other DOS cards do.

Of course this all assumes that the 575 and 6xx logic boards have the same layout; the DOS card is a very tight fit and a small variation in component placement could cause problems.

 

dkjones96

Active member
Tried to do this mod on a revision D board and got smoke. Oops! I went with the 640x480 posting at Powercc.org.

Here is a list of the parts mentioned to replace and what the Rev D board has:

CL9 and CL10 - 3.3uF - Changed to 2.2uF and 1.8uF respectively.
DL21 and DL22 - Changed to single 18V Zener at DL21(doesn't really matter as they are in parallel and the factory diodes are also 18V).
RP6 - 5.6K - I left this alone as the site says 4.52 to 5.11 and I found a different, higher, value.
RP8 - 15K - 33K
RF14 - 236K - 216K
RP23 - Unpopulated in Rev D - This is because a charge pump from the standby/crt supply feeds this circuit so I left it unpopulated since it only affects the logic supply.
RL54 - 18R instead of 33R - Changed to 15R.
RP61 - Already populated with 7.5K.
RL62 - 47R - Changed to 100R.
RL70 - 33K - 22K.
J104 - Jumper - 10R.

Sense line change is a bit more complicated. Revision D does away with J78 and J79 and there is a diode in that area. To make pin 20 connect as though it has J79 bridged you have to cut a trace from 20 to GND going towards the connector and run a jumper from 20 to 3.

I have not figured out where the smoke came from but I will borrow the FLIR from work and see if I can find it. Nothing looked obvious though as I turned it off the instant I saw it. When I turned the computer on the CRT breifly got HV and then lost it resulting in the normal static crackle. Screen stayed off and blank while the computer itself proceeded to boot.

 
Last edited by a moderator:

dkjones96

Active member
So, I don't know if I was just imagining smoke or what but there wasn't anything overheated or burnt when I very thoroughly looked over the board. I did have a wiring error on my part but it was signal level and what caused me to have no screen before.

It looks like I need to up the voltage a bit. I tried adding C26 from 1nF up to 5nF and all it appears to do is make the edges at the top and bottom extend to beyond what the horizontal deflection circuit can handle in the middle resulting in bad bowing and it makes the screen dimmer. It isn't a terrible aspect ratio, it is currently at 1.15 and I would like 1.33(4:3) It could be corrected by shrinking the vertical height but I would like it to take up more of the screen.

I took a thermal camera image of the AB after it ran for a bit but forgot to bring it for posting this. Hottest I found was 165F at the big guy near RF19 and second highest was the 7.5K at RP61.

Motherboard Signal.png

Screen after VGA.png

 

dkjones96

Active member
Alright, RP6 needs changed out, you can only get 72V max with it as-is. Swap RP6 to a 752(7.5K) 1206 resistor and you are in business. Looks beautiful! I had only an 822 1206 available but it requires PP1 to be turned down all the way to get 84V.

Only problem left is RF11 which runs over 300F as-is and when it's that hot you can't reboot or you'll lose sync and you have to shut down until the resistor cools and it seems to work fine again.

Hi-Res Mod.png

 

Paralel

Well-known member
Also, it shouldn't be possible to do 640x480 @ 16-bit. That would be 614.4k of VRAM. The most you can have in a CC is 512k. How does the system handle the difference in available vs. needed VRAM? Does it dip into system memory? Is it even capable of doing that? It doesn't seem likely, since the default video modes for the CC all fit easily within the maximum available VRAM.

 

dkjones96

Active member
Also, it shouldn't be possible to do 640x480 @ 16-bit. That would be 614.4k of VRAM. The most you can have in a CC is 512k. How does the system handle the difference in available vs. needed VRAM? Does it dip into system memory? Is it even capable of doing that? It doesn't seem likely, since the default video modes for the CC all fit easily within the maximum available VRAM.
I have upgraded this guy to a Color Classic II logic board(couldn't stand the gimped stock CC performance) and with the 256K expansion I have 768K VRAM.

As for screen distortion, it's the photo and I'll try to get a better one. I'm upping the dissipation rating of a resistor that is dissipating too much at the moment and when that is done I will get better photos. It does curve a bit at the top of the screen but it appears to be normal as it did it before on another analog board and is also something I notice in photos of other machines.

 

dkjones96

Active member
I thought I'd add this little update. This mod seems to be working just fine so far on the Rev D board outside of a warm reboot sync problem that I can't trace. None of the ICs or resistors, outside of the really hot area below, get above ~115F so it shouldn't be drift or anything like that. Leave the machine off for a few minutes and it syncs right up and works for hours without a problem. It just seems to need shut down, not powered off, so the problem is unrelated to RF11.

I upsized RF11 to a 3 watter and it runs at about 250F now but CF8 just runs hot I guess. That whole resistor/capacitor/zener diode area runs as long as the switch on the back is turned on so that electrolytic sits at about 65C(measured by attaching thermocouple). It's not comfortable to touch but well within specs, especially if you swap it out for a 105C during recap.

Considering undoing the mod as I don't like the lack of clarity at 640x480 compared to factory resolution. To each their own!

 
Last edited by a moderator:

Crutch

Well-known member
Forgive a naive question, I have never tried modding a CC but am thinking about it - why is this method preferable to cutting the analog board trace, which seems much easier? Is it because we don’t like dremeling the analog board, or does the older method have problems?

I’m referring to this of course, “option 1”: https://powercc.org/640x480/

 

techknight

Well-known member
Because this method was discovered on a prototype analog board that Apple's engineers themselves did to achieve 640x480

 

ttb

Well-known member
Echoing @Crutch a bit, but how is everyone feeling about this mod a few years on? It seems like @Bolle did it with "satisfactory" results (dark, unfocused image; better after warm up) and @dkjones96 did it and is considering undoing it (sync loss on warm reboot and lack of clarity). I get that it's as close to Apple-endorsed as possible, but that feedback is worrisome, especially given the amount of additional tweaking both of these guys had to do.

It seems really hard to directly compare this to the old preferred way (Option 1 on https://powercc.org/640x480/) since there are different AB revisions and perhaps even more variation at work. I guess I could try to do it the way presented here and then fall back to Option 1 as a plan B, but it would be nice to avoid excessive rework on a 30-year-old PCB.

 

mraroid

Well-known member
Hello ttb...

I also use the https://powercc.org

I think it is one of the best sites for original Color Classics, or mods. 

I did the 640 X 480 mod the old time way and used option one on the above web site.  That was maybe three or four years ago.  I use my CC about two or three times a week, for maybe between one and two hours each time.  I have only had one problem develop.  About 6 months ago I found two shadows on my screen.  Apparently, the built in degassing coil on the Sony CRT was not working.  So I bought am external degassing coil and it fixed the problem ASAP.  This issue did not appear until almost three years after I did the mod.  So I believe the mod had nothing to do with it.

Having said that, I think Techknight's solution is the best.  But my issue is that Techknight's solution applied to only one version of the board, and Apple had made several versions of the board.  The two boards I had where not the same one that Techknight used.  So rather then screw it up, I went old school. 

So far, so good.

mraroid

 

dkjones96

Active member
@ttb I did end up undoing it and then going with the old preferred way which is option 1 above. I kept having problems with a pair of diodes overheating an electrolytic capacitor in almost the dead center of the analog board. It runs crazy hot at 84v and stays that way even when the computer is off unless you flip the switch on the back. It does have a bit of wobble that I haven't tried to figure out yet but it isn't bad. It's much less work and way easier to reverse if you want.

 

ttb

Well-known member
Thanks, @dkjones96! I ended up doing it option 1 myself on one CC and am so far pleased. The geometry is good but the focus is a little off. Haven't tried tweaking the focus pot yet, though.

 

techknight

Well-known member
I suppose its highly dependent on the board revision, etc.... 

Still, its good for documentation purposes on how Apple approached it back then. 

 

Jagmn

Well-known member
Sorry to bump such an old thread, but it seems like a good place to ask questions about this mod.

I've got a 820-0360-B Color Classic analogue board that I've tried this mod on. I think it is the same revision as @Bolle 's. Unfortunately I'm seeing a very different issue after making the component changes, so I can only assume I've done something wrong. I have triple checked the component values, locations and solder joints and this is what I see on boot:

IMG_6287.jpeg

Cleaning and tweaking the adjustment pots don't improve the situation (but they do affect the screen).

Things I've tried:
  • Two different logic boards (575 and original CC)​
  • With and without capacitors at CL26​
  • Double checked the changed components, orientations and values, also checked my soldering.​
  • Measured the voltage at TP1 and adjusted PP1. It started at about 70V and I can get it to 80.1V at TP1.​
I've seen @dkjones96 's post about swapping RP6 to a 7.5k, but I'm not convinced that it is the voltage I have a problem with right now - but I'd be happy to be told I am wrong.

Does anyone have a suggestion? Thanks!
 

Jagmn

Well-known member
Gah, I think I found the culprit: RF14, I read one of my composite resistors backwards meaning I'm off 216k by a couple of orders of magnitude :confused: I don't have the right value to hand so I'll await the next parts delivery…
 

Jagmn

Well-known member
Part arrived and all looks good. I'm not quite at 84V and the CL26 capacitors just seem to make the display dimmer and more wonky. I've still got fairly thick horizontal borders that I'd like to remove.

What test points do you use to measure the horizontal scan voltage? I've been measuring it between LP5 and the ground that goes to the logic board, with the circuit disconnected.

Thanks
 
Top