Since my Apple II+ is modded for the PAL region and I don't have a PAL encoder card, it can't display color. So I decided to restore it to NTSC, but the result is not so successful.
As you can see in the video, the display is far too bright and unstable, with visible vertical bars. It is mainly in text mode that the image flickers, in color mode it is slightly more stable (but still too bright with visible bars).
To restore it to NTSC, I have made the following changes:
- replaced the crystal with a 14.318 (from ReActive Micro)
- removed the solder blobs from the circle jumpers on the motherboard
- joined the triangle jumpers on the motherboard
- removed the wires from B10 and B11 to slot 7 on the backside of the motherboard
I've adjusted the composite trim pot and it is set as good I can get it. I've also adjusted the color pot and I think the colors look as they should.
Does anyone have any suggestions on what the issue might be?
I should add that I've tried to do the same modification once before, but with a crystal bought at the local electronics store. The result was pretty much the same (with some additional issues). What I then noticed was that it was when I changed between bridged circle jumpers to bridged triangle jumpers that the problem with the washed out and unstable image started. The replacement of the crystal itself had no negative effect on the image quality.
Have you tried adjusting the video output level potentiometer at J-13?
The one shown here: https://www.nightfallcrew.com/wp-content/gallery/apple-ii-europlus/IMG_8076.jpg?4baa53
Yes, I've adjusted it as good as possible. :( I can lower it, and darken the image a bit, but that will make the signal too unstable and text/colors too dark. If I turn it up, the main issue just get worse.
Twice now has the computer booted with a good, stable and dark image, but with color disabled. You can see in the attached picture that the black background is much darker, and the vertical bars are less visible. Not perfect, but good enough. If I then start adjusting the video potentiometer color will eventually come back. But also the unstable image, grey background and vertical bars...
I have very little knowledge about the color burst signal, but I'm guessing that when color burst is completely off = stable image? Color burst on = unstable image (also in text mode, as I understand it, color burst is not completely off in text mode?).
It's possible that you are dealing with frequency problems.
Color burst accuracy only affects the ability of a monitor to display color. The fact that sometimes the monitor displays color and other times does not indicates to me that it is possible that your crystal frequency might be slightly off. The Apple II is very sensitive to exact crystal installed. A while back I bought 6 different crystals that were all listed as the correct frequency and only found 1 out of the 6 that worked consistantly.
Can you swap a crystal from a known good NTSC Apple II and see what happens?
... try to find a real NTSC TV with a real CRT (Cathode Ray Tube) and try how the picture looks there.
IMHO the vertical bars may be an artifact of the LCD TV/Monitor re-sampling the video to its pixel rate. I've worked in the design of mostly digitized TV ICs for the last generation of CRT based TV sets (pre year 2000 ... time flies) and I remember that our digital design team had huge efforts to weed these artifacts out. It's much like a Moire effect. The human eye is very sensitive to that. Having crisp pixels as on a LCD only exacerbates the issue. Otherwise (on a CRT) you could "cheat" by adding noise to the signal or the pixel clock which would "smear" the edges and the eye would perceive a nicer looking, smooth, but a little less sharp picture. Which looks much less offensive than the crisp one ! (Psychology of human vision involved).
The color fields on your screen as such seem OK to me but without a vectorscope I could not tell if there are deeper issues with the color signal.
Another pitfall may be that this TV/monitor may use a comb filter in the color decoder. The Apple-II does not produce a ITU standard conforming NTSC signal where the phase of the color burst flips by 180 degrees on each other horizontal line. The Apple-II violates this rule intentionally to make color artifact colored vertical lines possible to appear as straight lines and not as jagged lines. Woz even got a patent on the Apple-II video system due to this trick. But back in the day, there were no TVs with comb filters and it worked. Nowadays, not so much. However, many other 1970s consumer products also violated the NTSC standard for the same reason. Other than the Apple-II, all Atari VCS did it, all Atari 8-bit computers did it, and I suspect most color TV games of the era did it. So when TVs/monitors were digitized in the 1990s, and comb filters came up, various levels of auxiliary circuits were put in to still make these chip sets work with these older consumer products, and still have color, but the level of success of these added tricks varies from chip set to chip set, and, frankly, idiocracy is at work and much of that knowledge about quirks and tricks for decoding ancient obsolete video signals nowadays is lost, and younger engineers which were not even born when the CRT disappeared don't have that knowledge and now may design video chip sets that can't cope with these ancient signals anymore, especially if these signals don't conform to the published standards. I could tell you stories ...
So you best bet is to use a much older, perferably CRT based TV/monitor of the cheapest possible kind (no expensive comb filter inside). If your picture there is good, you don't have a problem with your Apple-II, and what you don't have you can't find regardless of how much money you throw at it by buying replacement components.
As for the crystals, even the cheapest color subcarrier crystals you can find anywhere (except manufacturing fallouts / rejects) are good enough for any TV. These were consumer products that were made by the tens of millions and everything had to be made as cheap as possible. All the difficulties some people report come from a given crystal not seeing the proper load capacitances specified in its datasheet and, sometimes, being overdriven with a signal being too strong. The "signal being too weak" case does not exist - it would not oscillate at all.
The primitive Apple-II crystal oscillator with its two transistors is highly suspect and Apple (the corporation) certainly had a custom crystal made which would work in this circuit. This means if you purchase any random crystal, even of the correct frequency, all bets are off. There is a reason why most reputable textbooks on electronics discourage designing your own crystal oscillators. These are "building bricks" you better buy ready-made in form of the ubiquitous DIL-14 or DIL-8 (or their SMD equivalents) crystal oscillators. If made by a reputable manufacturer these will guarantee proper start up, proper mode of operation of the crystal, and proper crystal drive power for a long life, over all the specified temperature range.
All this can be achieved by a competent electronics engineer in a well-equipped lab, too, if you have the datasheet of the crystal, a climate chamber, and maybe a week of time. It may be worth the savings in mass production over ready made DIL oscillators if you intend to produce millions of these cards with your homegrown crystal oscillator. Apple did it, apparently. But they knew the secret sauce on how to spec their crystal (or learned the lesson the hard way, I don't know how much trouble they had with that oscillator).
But this said, if you have any doubt about the oscillator in your Apple-II, buy a ready made 14.31818 Mhz DIL oscillator and hook it up in lieu of the 2-transistor contraption. Keep in mind that all of them need a power supply bypass capacitor as close to the power pins as possible (100nF ceramic works well) and that most do have an "enable" pin which needs to be hooked up the proper way. (Datasheets for the part in question always help)
I don't have any known working crystal unfortunately. The crystal comes from ReActiveMicro, who claims all units are tested before shipping. I can of course nor guarantee it's working but should be alright. I purchased a set of two so I can always try the other one if I find no other solution.
I've now noticed that the computer almost always cold boots with color disabled. I then have to turn the video level pot a couple of turns before color is possible again. And then picture becomes grey/unstable. If I quickly just turn the computer off and on, color will remain.
Thank you for your detailed answer! I also suspect that the LCD TV I use may be part of the problem. However, I live in Europe and finding an NTSC CRT is very difficult. Then I think it's maybe better to restore the PAL modification and start looking for a PAL encoder card instead.
" I live in Europe and finding an NTSC CRT is very difficult."
Don't give up so quickly ! There were lots of TVs and video monitors sold in 1980's and 1990's Europe which have a NTSC decoder inside. They also would do PAL and SECAM, of course. You just need to look into their datasheet / technical specs, though. A little bit of detective work with google needed. I have a few of those sitting in my basement a could look at their model / manufacturers, if you are interested.
Some of the most unexpected candidates are these small portable TVs with stick antennas popular with truckers. Most of these cheap TVs were made in Asia, based on multistandard color decoder chips. If they can do PAL and SECAM, and are based on a multistandard chip set, most likely they also can do NTSC. It's very cheap to make a PAL decoder switchable to NTSC. But it must be a multistandard TV. A plain vanilla single standard TV will only be able to do that single standard.
Another option is to try finding a LCD based TV which works well with those nonstandard NTSC signals. There are lots of websites / forums out there who deal with this. Try sites dedicated to Atari 400/800/600XL/800XL computers. www.abbuc.de perhaps. Many PAL Ataris were later converted to NTSC to get the artifact colors in hires video mode used by many US made games. So these collectors had to find suitable NTSC TV/monitors and I know for certain they did find some, and even LCD / flatscreen ones, which work fine. But myself, I use CRTs. The light gun needed for some games only works with real CRTs.
I'll keep my eyes open! I actually didn't know that ordinary cheap CRTs can have NTSC support. I thought it was a feature you paid $$$ for. Right now, however, I have too little space at home for another gadget, like a CRT...
Something new that I have discovered, I don't know if it can give any clue to the cause of the problem:
When I cold boot the computer, like I said before, it always starts with color off and stable image. But every now and then, for very short moments, it flickers with color. I took a photo when the color was displayed and can see that the image actually looked like it should, with completely black background and no artifacts:
Then when I restart the computer and it's "warmed up", the same screen looks like this:
So for a few frames, color is actually displayed correctly, without artifacts. Well, until the next restart... :)
Edit: I can also reproduce the "cold boot" behavior this way:
1. Start with color enabled / unstable image with artifacts
2. Switch to 80 column mode (PR#3)
3. Switch back to 40 column mode (Reset)
4. Color is now disabled / stable image with no artifacts (good color will flicker now and then)
...actually it's not the software switching between 40 and 80 columns (PR#3 and Reset) that restore the monochrome mode. It's when using the hardware switch between internal graphics and the 80 column card. When I switch back to internal grapichs it somehow resets the graphics mode.
Perhaps you need to chase an intermittent fault in the video generation circuitry.
It's a tricky area of the motherboard, so you'll need to be persistent.
Look in this document: https://www.willegal.net/appleii/apple-service-notes.pdf
The part on "Motherboard Page 4" describes maybe what you're seeing - vertical bars...indicating possibly bad chip(s) at C12, D12, F14, B4
Thanks! I have focused on those chips before, but not more than cleaning and re-seating them. I tested them now with my EPROM programmer and actually B4 (74LS194) tested bad. I don't have any spare 74LS194 so I swapped it around with B9 and A10, but that didn't change anything.
But I also tested some other chips related to the video generator: A12, B12, B13, C11 and C13. And also C13 (74LS51) tested bad. If I understand it correctly that's the video sync multiplexer. Don't know if a faulty C13 can cause issues like the one my Apple has? I don't have any other 74LS51 chip to test, maybe I should order some new and try.
So I continued to chase a solution to this issue. I replaced the 74LS194's, 74LS51's and the 2N3904 transistors (Q3 and Q6). Also replaced the R9 resistor and changed R6 to 4,7 KOhm instead of 2,7 KOhm. To no avail.
I also purchased two different VGA scalers, in hope they would produce a better video signal that my LCD TV would accept. One cheapo generic from Amazon and one a little more high end. Both produced really low quality video signal, worse than the already bad composite signal. I also bought a cheap AV2HDMI adapter, but it was the same crap quality signal.
I also got another LCD TV (mainly for the smaller form factor). But same video issues as with the other TV.
So I finally gave in for a modern solution and purchased the VidHD HDMI card. Together with the 19" LCD the quality is now great! I'm a little bit disapointed that I couldn't solve it without using modern hardware, but the VidHD is great, so I'm happy.
I think I will now revert the computer back to PAL again, to keep it as it once was sold.
Sometimes you have to kill the fly with a hammer. The VidHD is great. And it gets rid of the annoying colour fringing like a good RGB card/monitor combination.