II+ "Color Killer"

[Bobbi]

My 'new' Apple II+ has horrible color fringing in TEXT mode. I uncovered the following nugget from The Apple II Circuit Description:

Color Killer-Early Rev. 7 boards do not have IC A14. These boards
use transistor Q6 as a color killer. When TEXT MODE is high, base current through R27 drives Q6 into saturation. This shunts COLOR BURST to ground. This use of Q6 is not fully effective and some tv sets still display tinted text characters. In later Rev. 7 boards, A14-1 and A14-4 are added to force COLOR BURST to a low when TEXT MODE is high. For these boards, the last waveform of Fig. 4-3 is always low. Note that A14 is inserted into the circuit differently on Rev. 7 boards than it is on RFI boards. Also note the use of COLOR REF versus COLOR REF'.

My rev 7 board does not have ICs A14-1 and A14-4 so maybe the fringing is just a fact of life (unless I want to do a mod to the MLB). Maybe I should just go find a *mono* monitor!

I also discovered this (from an old 1999 posting here):

"Another thing to check is the 74LS259 (IC F14). If it is making poor
contact on certain pins (mainly, Pin 4) the color kill will not work. Try
scoothing [sic] it up and resocketing it."

In my machine, F14 pin 4 had a patch wire inserted into the socket, so maybe the socket got damaged. I will pull the chip, clean it up and resocket it to see if it makes any difference.

[Michael J. Mahon]

Bobbi <bobbi....@gmail.com> wrote:
> My 'new' Apple II+ has horrible color fringing in TEXT mode. I uncovered
> the following nugget from The Apple II Circuit Description:
>
> Color Killer-Early Rev. 7 boards do not have IC A14. These boards
> use transistor Q6 as a color killer. When TEXT MODE is high, base current
> through R27 drives Q6 into saturation. This shunts COLOR BURST to ground.
> This use of Q6 is not fully effective and some tv sets still display
> tinted text characters. In later Rev. 7 boards, A14-1 and A14-4 are added
> to force COLOR BURST to a low when TEXT MODE is high. For these boards,
> the last waveform of Fig. 4-3 is always low. Note that A14 is inserted
> into the circuit differently on Rev. 7 boards than it is on RFI boards.
> Also note the use of COLOR REF versus COLOR REF'.
>
> My rev 7 board does not have ICs A14-1 and A14-4 so maybe the fringing is
> just a fact of life (unless I want to do a mod to the MLB). Maybe I
> should just go find a *mono* monitor!

Or just turn down the "saturation" control all the way to force mono.

> I also discovered this (from an old 1999 posting here):
>
> "Another thing to check is the 74LS259 (IC F14). If it is making poor
> contact on certain pins (mainly, Pin 4) the color kill will not work. Try
> scoothing [sic] it up and resocketing it."
>
> In my machine, F14 pin 4 had a patch wire inserted into the socket, so
> maybe the socket got damaged. I will pull the chip, clean it up and
> resocket it to see if it makes any difference.

If there's a problem, it's probably with the socket. You can try to bend
the spring contacts back, or, to try things out, insert a short length of
the patch wire--that should restore good contact.

--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com

[Bobbi]

I didn't have any luck trying to improve the contact of the 9334 at F14. However, playing around with the machine some more reveals an interesting symptom.

In TEXT mode, the colour fringing is pretty horrible and there is quite a bit of noise in the picture in terms of vertical lines. The background level is well above black. This is a pre-RFI Apple II+, with one chip elevated in a wirewrap socket, so it is probably super awful for RFI. However, switching to a mixed mode such a GR markedly improves the quality of the text in the bottom four lines. There are still colour fringes, but text looks normal (for Apple II) in mixed mode. It is TEXT mode which is so horrible, but why?

Why would the video level be different in TEXT mode versus GR, I wonder?

[Bobbi]

It is almost like the color killer is activating in TEXT mode but is making the signal worse (*more* fringey.) Is that even possible I wonder? Back to reading 'The Apple II Circuit Description' for me.

[James Davis]

On Friday, October 13, 2017 at 5:11:09 PM UTC-7, Bobbi wrote:
> I didn't have any luck trying to improve the contact of the 9334 at F14.

You could put another wire wrap socket into the F14 socket, too.

[Bobbi]

I think F14 is okay. Ssytem switches between TEXT / GR / MIXED modes without any problem. In general the TEXT mode video looks much noisier than text in the bottom four lines of the MIXED mode.

Is there an easy way to measure whether or not color killer transistor is on or off? Or even to force it on or force it off? That would help my debugging.

[Patrick Schaefer]

Am 14.10.2017 17:29 schrieb Bobbi:

> Is there an easy way to measure whether or not color killer transistor
> is on or off? Or even to force it on or force it off? That would
> help my debugging.

To force it on, just short its collector to ground. You can do this at
the center leg of the variable capacitor C3, where the trace goes to Q6.
Connect this point to ground to disable the color burst.


Patrick

[Jorge]

On Saturday, October 14, 2017 at 5:29:14 PM UTC+2, Bobbi wrote:
>
> Is there an easy way to measure whether or not color killer transistor is on or off? Or even to force it on or force it off? That would help my debugging.

Lift F14 (LS259) pin 4 out of the socket, then feed +5 or GND to B13 pin 12 to turn it on/off.

The color burst signal comes from B12 (LS11) pin 12, if you lift that pin out of the socket then there won't be color burst regardless of the transistor.

Many II/II+ have a problem with the GND connection of the video out RCA female connector. If that's the case you just need to solder the round part to the bracket, from behind. If you grab the rounded part of the RCA video out female plug and can rotate it even slightly then you've got the problem.

Another frequent problem is the video out level pot/adjustable. It's invariably of a shitty/cheap brand with the carbon exposed (unlike the spanish PIHERs :-) and tends to fail quite often. Turning it back and forth a few times usually makes it work again.
--
Jorge.

[Bobbi]

Thanks for the info Patrick and Jorge!! I haven't been home all day so I haven't had a chance to investigate further.

After some reflection, I think the issue may be that my early Rev 7 board relies on the transistor Q6 to switch the colour burst and this is not 100% effective. Enough colour burst leaks through to make text mode ugly on a modern LCD monitor (which seems to be pretty sensitive to colour burst.)

Probably not much to do unless I want to do a mod to add the later more effective colour killer circuit which uses a TTL gate to switch the colour burst off (according to "Understanding the Apple II")

[Bobbi]

I have been meaning to post a couple of pics.

This is TEXT mode. http://picpaste.com/IMG_20171015_202324-RkXyVjFj.jpg

And this is the four lines at the bottom of HGR mode. http://picpaste.com/IMG_20171015_202310-0h7ThWVE.jpg

As you can see the fringing is way worse in TEXT mode, and in general the picture is more noisy. It looks better in the photos than in 'real life'.

I have been studying Jim Sather's book and there does not appear to be many possibilities for this type of issue. This has to be an issue with the analog video circuitry but that is pretty much the same in TEXT/GR/HGR modes. It pretty much *has* to be that the colour killer is making things worse.

I need to get in there with a scope and see what is going on. Analog is not my strong suit though.

I think the TEXT mode picture quality gradually improves as the system warms up. It is pretty horrible after a cold start.

[James Davis]

Bobbi,

Have you checked the intensity/brighness settings on your TV? It looks like its is way too high for colors, but not background. Of course, contrast may need to be adjusted, also.

What kind of connection are you using? Are you using a "Sup'R'Mod" RF converter (to Channel 3 or 4)? Or a direct feed to a color monitor (Not a TV)?

Just curious,

James Davis

[Bobbi]

This is composite into a LCD monitor. Works great with my IIe. Also works great in HGR / GR modes. It is just TEXT mode that is icky on this II+. Very mysterious.

I checked the obvious things - level potentiometer, ground connection on RCA jack. I have tuned the 'color trim' and the colours look great in GR mode. It's just TEXT mode looks awful.

[Bobbi]

Brightness, contrast and saturation are set to 50% (default) which works fine for other sources and also other modes on the same Apple II.

[James Davis]

On Sunday, October 15, 2017 at 7:00:13 PM UTC-7, Bobbi wrote:
> This is composite into a LCD monitor. Works great with my IIe. Also works great in HGR / GR modes. It is just TEXT mode that is icky on this II+. Very mysterious.
>
> I checked the obvious things - level potentiometer, ground connection on RCA jack. I have tuned the 'color trim' and the colours look great in GR mode. It's just TEXT mode looks awful.

MJM had something to say about composite into LCD awhile back in another thread/post (or this one).

[Michael J. Mahon]

But I don't think it is particularly relevant to this case.

I think Bobbi has got it right--the color killer on his board is not
attenuating the color burst enough to keep his monitor from switching into
color mode. (This is the very problem that the later revs were meant to
fix.)

Silicon bipolar transistors have a hard time shunting a signal to ground,
since there's always a diode drop of 0.6v involved. Perhaps a JFET would do
a sufficient job.

[Bobbi]

Still, the mystery remains - why is TEXT mode *worse* than GR / HGR mixed mode? It's not just that the colour killer seems ineffective - it seems to actually be making things worse somehow.

[Bobbi]

BTW Bobbi -> 'she' not 'he' :)

[I am Rob]

On Monday, October 16, 2017 at 6:58:30 AM UTC-6, Bobbi wrote:
> BTW Bobbi -> 'she' not 'he' :)

Way to go. I think you just burst everyone's bubble. Most probably thought tinkering with vintage hardware was a guy thing.

BTW, welcome to the club.

[Raymond Wiker]

I'm not sure if this has been mentioned before, but it seems that some
people have improved the colour killer by replacing a resistor: see
http://www.applefritter.com/node/24650

Actually, I just had a look at "Understanding the Apple II", and it
appears that the colour killer is not supposed to be active in MIXED
mode (page 8-5)... curiouser and curiouser.

[Bobbi]

Colour killer should be active in TEXT mode only. Usually you would expect colour fringing in the bottom four lines of MIXED, but not in TEXT.

In my case I *do* have fringing in MIXED as expected, but I have *worse* fringing in TEXT mode.

For whatever reason my machine seems to have nasty analog video in general in TEXT mode, but okay in GR/HGR. Doesn't make much sense.

[Michael J. Mahon]

Bobbi <bobbi....@gmail.com> wrote:
> BTW Bobbi -> 'she' not 'he' :)
>

My apologies, Bobbi! ;-)

[Michael J. Mahon]

It sounds like the color killer transistor is not properly connected to
ground. Try bridging the emitter to ground with a short wire.

[Michael J. Mahon]

Michael J. Mahon <mjm...@aol.com> wrote:
> Bobbi <bobbi....@gmail.com> wrote:
>> Colour killer should be active in TEXT mode only. Usually you would
>> expect colour fringing in the bottom four lines of MIXED, but not in TEXT.
>>
>> In my case I *do* have fringing in MIXED as expected, but I have *worse*
>> fringing in TEXT mode.
>>
>> For whatever reason my machine seems to have nasty analog video in
>> general in TEXT mode, but okay in GR/HGR. Doesn't make much sense.
>>
>
> It sounds like the color killer transistor is not properly connected to
> ground. Try bridging the emitter to ground with a short wire.
>

I said that without checking the schematic!

Better to bypass emitter to ground with a 0.1uF capacitor.

[Bobbi]

I have been meaning to get back to this -- maybe later this week!

[James Davis]

On Monday, October 16, 2017 at 5:58:30 AM UTC-7, Bobbi wrote:
> BTW Bobbi -> 'she' not 'he' :)

Who referred to you as a he? Although, MJM apologized for it, I did not see in his posts, where he did it.

I hope I didn't. I've always thought you were a she (from the spelling of your name).

[Bobbi]

"On the Internet no-one knows you're a dog." :)

I looked up the circuit in 'The Apple II Circuit Description' - it is Fig C-20 'Video Generator (part 2)'. Note that the circuit shown in the book is the later version where a NOR gate is used to disable COLOR BURST in TEXT mode. My board is an earlier rev 7 which relies on transistor Q6 only.

The diagram convinced me that I wouldn't set anything on fire if I connected the centre pin of varicap C3 (Color Trim) to ground. When I do this the COLOR BURST is shorted to ground and I get a wonderfully clean mono TEXT mode display. It seems the COLOR BURST is the only issue - without it all the dot crawl and artifacts went away.

My conclusions are as follows:
1) COLOR BURST is inadequately disabled by transistor Q6 (as Apple themselves recognized with later changes.)
2) My particular monitor gets very upset when COLOR BURST is present but attenuated / noisy which results in TEXT mode looking much *more* fringey than HGR for example. Noisy colour burst also causes other screen artifacts / dot crawl type effects.
3) COLOR BURST itelf is fine - colors are clear and true in GR mode, for example. I tuned C3 to get the hue correct.

So I guess I will live with it. I may tack-solder a wire to the middle of varicap C3 and connect it to ground via a COLOR KILL switch. Is there a nicer way to do this, I wonder, without just adding gate A14 ...

[Patrick Schaefer]

Am 18.10.2017 15:41 schrieb Bobbi:

> 1) COLOR BURST is inadequately disabled by transistor Q6 (as Apple
> themselves recognized with later changes.)

As a next step you can check if the transistor is working properly.
Connect a 2,2k resistor from base to +5V. This will turn the transistor
fully on. If nothing happens, Q6 is dead.

If this works, you can reduce the value of R27 to drive it harder into
saturation. Use 2,2k or 1k (i.e. connect 4,7k or 2,2k in parallel to the
existing one).

The last step would be replacing Q6 with something more efficient. A
low-saturation type or N-channel MOSFET. But you want to keep it
original, won't you?


Patrick

[Bobbi]

I know the transistor Q6 is doing something because TEXT mode is way more fringe-y than the bottom four lines of GR or HGR MIXED mode (see the pics I posted further up in the thread.) If the transistor was dead then the colour fringes would be identical in TEXT and MIXED modes.

My working theory is that Q6 is attenuating but not eliminating the colour burst and my (modern LCD) monitor still detects the colour burst and tries to do colour, but winds up with a very noisy display due to the unstable attenuated colour burst. Something like that.

I would like to keep the machine original but I am open to reversible mods, such a a wire tack soldered to a point on the board, for example.

[ritt...@gmail.com]

You can't really test this with and LED. The composite out of the original II is "weird" enough to mess with them pretty badly.

Try it on an old color composite monitor before you go chasing problems you may not even have.