• Please review our updated Terms and Rules here

Looking for some help to solve an Apple IIe video weird problem!

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
Hello

I bought an Apple IIe PAL version with a 80col/64 k that has a very strange behavior:

Actually when in 80col mode, in text mode, one character out of two is drawn without its last pixel column.

in 40 col mode everything is ok, the problem only occurs in 80 col mode.


Actually, in 80 col mode, either in text or double hires, bits 5 and 6 (counting from 0) of main memory are handled on video as if they were replaced by bit 0 (counting from 0) of Aux memory!


Here are screenshots that illustrates the issue at http://zakor.dyndns.org:8095/appleII/videobug.zip

1.jpg is the aux video set to 1 : as you can see there is more than 1 pixel set! But 3 pixels

41.jpg is the aux video set to 41 (first and 7th bit set to 1): again the first bit is much larger than the 7th one (because there are here 3 pixels)

7F.jpg is the aux video set to 7F (all first 7 bits set to 1) with the whole main video set to 0: as you can see there is less free space for the main video bank!! As you can see on the TEXT below the graphics part, the right part of each second character (taken from the main memory) is truncated.


I discovered that in double hires mode the main memory video only uses the first 5 bits (instead of 7)For example, setting 2000:20 or 2000:40 makes no pixel appearing ! in normal hires mode everything is normal though!



Setting the AUX $2000 address to 1 makes appear three pixels on screen (as a wider pixel) AS IF main memory(to the left) has its 6th and 7th pixel set to 1. But if I read main memory in the monitor, it is still set to 0 !! (please refer to the 1.jpg screenshot). Besides there is NO left main memory to $2000 (since the first column is in AUX memory) and we have the same artifact.



So the problem seems to be a video signal problem, NOT a memory problem.



It seems to me that the video signal that is generated behaves as if bits 5 and 6 (counting from 0) of main memory were overwritten by bit 0 (counting form 0) of AUX memory. This should lead us to the faulty component I guess.


I also attached in the ZIP several files of a hires demo that shows the problem:

- ext80demo2.dsk disk image that contains a double hires program tests I used

- emulator.gif : the perfect result of the double hires 'cubes' test on the famous AppleWin emulator

- real.jpg: the result of this same test program on my apple IIe. As you see there IS a double hires problem



I hope to find someone who would be able to give me hints which component may be involved. Could you help me please?



Thanks a lot



cyril
 

RWallmow

Veteran Member
Joined
Oct 19, 2006
Messages
1,791
Location
Anoka, MN USA
...Here are screenshots that illustrates the issue at http://zakor.dyndns.org:8095/appleII/videobug.zip...

Yeah, no disrespect meant, but I don't know of anyone here who is going to download a ZIP from a BRAND NEW member, if you want us to help troubleshoot, please post your photos using the forum or a known photo sharing site like picasa or photobucket and someone here can probably help. It would be WAY too easy for someone to get a virus from an unknown zip file, just saying.
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
screenshots to illustrate to video problem

screenshots to illustrate to video problem

Ok no offense

here are the screenshots:

in this order:

1.jpg is the aux video set to 1 : as you can see there is more than 1 pixel set! But 3 pixels

7F.jpg is the aux video set to 7F (all first 7 bits set to 1) with the whole main video set to 0: as you can see there is less free space for the main video bank!! As you can see on the TEXT below the graphics part, the right part of each second character (taken from the main memory) is truncated.

41.jpg is the aux video set to 41 (first and 7th bit set to 1): again the first bit is much larger than the 7th one (because there are here 3 pixels)


emulator.jpg : what it should look like (using the applewin emulator)
real.jpg: what it looks like on the computer








1.jpg7F.jpg41.jpgemulator.jpgreal.jpg

thanks !!
 

RWallmow

Veteran Member
Joined
Oct 19, 2006
Messages
1,791
Location
Anoka, MN USA
Ok no offense

here are the screenshots:

in this order:

1.jpg is the aux video set to 1 : as you can see there is more than 1 pixel set! But 3 pixels

7F.jpg is the aux video set to 7F (all first 7 bits set to 1) with the whole main video set to 0: as you can see there is less free space for the main video bank!! As you can see on the TEXT below the graphics part, the right part of each second character (taken from the main memory) is truncated.

41.jpg is the aux video set to 41 (first and 7th bit set to 1): again the first bit is much larger than the 7th one (because there are here 3 pixels)


emulator.jpg : what it should look like (using the applewin emulator)
real.jpg: what it looks like on the computer








View attachment 11513View attachment 11514View attachment 11515View attachment 11516View attachment 11517

thanks !!

I have not seen that exact pattern, but I would start with likely RAM issues, seems to be common point of failure on Apple IIs.

Do you have any RAM chips you can try to piggyback on top of the existing ones, that usually helps to quickly track down the failed one.

You could also be looking at bad solder joints, I had a //c which was somewhat garbled, and with my experience with bad RAM in a //e, I was figuring it was a bad chip, but I found in wiggling/flexing the board near the RAM it would work. Soldering iron in hand, I re-flowed all the solder on the RAM chips and it fixed it.
 

RWallmow

Veteran Member
Joined
Oct 19, 2006
Messages
1,791
Location
Anoka, MN USA
I have not seen that exact pattern, but I would start with likely RAM issues, seems to be common point of failure on Apple IIs.

Do you have any RAM chips you can try to piggyback on top of the existing ones, that usually helps to quickly track down the failed one.

You could also be looking at bad solder joints, I had a //c which was somewhat garbled, and with my experience with bad RAM in a //e, I was figuring it was a bad chip, but I found in wiggling/flexing the board near the RAM it would work. Soldering iron in hand, I re-flowed all the solder on the RAM chips and it fixed it.

Sorry to follow up on my own post, but I forgot to address that you don't think its RAM. Almost all of the Apple II video signal is generated through main memory (though with the 80-collumn card I believe it has some memory on it too, I dont actually have one, so not 100% sure on this).
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
Hi

According to me it is not a ram issue since the missing lines are on the main memory (not the aux 80col /64 k board) locations and these same ram chips perfectly work in 40 col mode.

moreover if it was a ram issue setting some values in AUX memory would fill main memory ram to create this problem. BUT as I wrote the video signal is generated AS IF the main memory was overwritten by the AUX mem but actually it is not modified.

writing
2000:0 N 2001<2000.3FFEM --> fills the main video ram with 0
C001:0 N C055:0 N 2001:1 --> fills the aux ram at location 2001 with 1

this lines makes 3 pixels appearing as if MAin RAM 2000 bit 5 and 6 were switched to 1 !!

but if you read $2000 you will get ....0!


this is why I guess this is a video signal generation issue when merging the Aux ram and Main ram to create the 80 col / double hires image.

but then... what should I do ?

thanks
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
Hi

Somebody just lent me two others 80col / 64 k boards.. but the problem is the same

I now have to find video chips to test a rom replacement or do you know how I can flash the video rom with a PC?

any other lead ?

thanks !!
 

RWallmow

Veteran Member
Joined
Oct 19, 2006
Messages
1,791
Location
Anoka, MN USA
Hi

Somebody just lent me two others 80col / 64 k boards.. but the problem is the same

I now have to find video chips to test a rom replacement or do you know how I can flash the video rom with a PC?

any other lead ?

thanks !!

If you have access to a ROM burner, burning a new chip should be pretty trivial. Reactive micro used to sell alternate video ROMs pre-burned with enhanced character sets (like mousetext, and others). However I don't believe they are taking new orders, you could email them to see.

I would still take a hard look at main RAM though, mine seemed to be just fine, even passing all RAM diagnostics, seriously like 99.9999% of Apple "video glitches" are main RAM. I have seen the exact opposite symptoms as yours (yours has last column of pixels missing, saw one with last column of pixels ALL lit) and it turned out to be bad RAM after the person mucked with trying to replace ROMs thinking their RAM was just fine. Its almost always RAM, lol.
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
Ok I will inspect the RAM chips (not so many on this computer :) ) but how do you understand they behave perfectly well in 40 col mode?

and how do you understand that in 80 col mode the video image shows pixels from main ram set to 1 although in reality the main memory is still 0 ?

Doesn't it seem more realistic that it comes from the video signal that is a corrupted combination of the main memory and the aux memory ?

thanks

cyril
 

RWallmow

Veteran Member
Joined
Oct 19, 2006
Messages
1,791
Location
Anoka, MN USA
Ok I will inspect the RAM chips (not so many on this computer :) ) but how do you understand they behave perfectly well in 40 col mode?

and how do you understand that in 80 col mode the video image shows pixels from main ram set to 1 although in reality the main memory is still 0 ?

Doesn't it seem more realistic that it comes from the video signal that is a corrupted combination of the main memory and the aux memory ?

thanks

cyril

I never said it makes sense ;-) I just know what I have seen in the past. RAM that passes all tests, however when acted on in VERY specific ways, fails, in very odd ways.

There is LOTS of trickery in the Apple II design to make it do as much as possible with as little hardware as possible, and a lot of that trickery is in the video path, using main ram, versus dedicated video ram like most other computers, and using software timing and NTSC witchcraft to make colors without extra hardware. Really a genius design, but also kind of weird.
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
OK

I inspected the RAM on the motherboards. I watched the 8 chips very closely on both side of the mainboard but found nothing visible.

Should I unsolder two of them and switch them? according to the screenshots I sent here, which ones should I switch ?

thanks

cyril
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
Someone just gave me another video rom... but the problem remains

so what's next

- motherboard ram?
- LS166
- LS374

I have all these chips on another motherboard

Since I now have to unsolder chips, I want to proceed with caution.

According to you which chip (ram, ls166,ls374) should I switch first ?

thanks

cyril
 

RWallmow

Veteran Member
Joined
Oct 19, 2006
Messages
1,791
Location
Anoka, MN USA
Someone just gave me another video rom... but the problem remains

so what's next

- motherboard ram?
- LS166
- LS374

I have all these chips on another motherboard

Since I now have to unsolder chips, I want to proceed with caution.

According to you which chip (ram, ls166,ls374) should I switch first ?

thanks

cyril

I would start with RAM, but make it easy on yourself in the future and solder in sockets, rather than soldering the replacement chips right into the board, that way the next time RAM issues pop up its a simple pull & swap.
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
Hi

I switched the LS166 and the LS374... same problem!!!

To sum up:

- KERNEL OK
- switch the 80 col /64 K board
- switch the video ROM
- swith the LS166
- swtich the LS374

now I have to switch some RAM chips: do you know which one corresponds to bit 0 , bit 1 etc... ?

thanks

cyril
 

Arcady

Experienced Member
Joined
Jan 1, 2013
Messages
69
Location
Louisville, Kentucky USA
When you do the system test and you get RAM error results like this: 0 0 0 1 0 0 0 0

The 8 results correspond to the 8 RAM chips across the motherboard, but in reverse order. The chip at the right is bit 0, while the chip at the left is bit 7. The chip marked 1 above would be the 4th chip from the right when looking at the motherboard from the front of the computer. I figured this out recently when I had to replace a faulty RAM chip in one of my //e's.
 

cyril3d

Member
Joined
Jan 3, 2013
Messages
25
Hi

The trouble here is that the internal test gave NO ERROR but a KERNEL OK, so the RAMs should be ok.

So it may be another component involved in the display process

cyril
 

nige the hippy

Veteran Member
Joined
Apr 7, 2006
Messages
1,282
Location
Luton UK
Is the machine as originally built? it just reminded me of the problem we had running original apple software on our ITT2020 Apple II clone. Whenever data was directly written to the video area of RAM, we always used to get vertical lines missed out (or added) as if the ITT used all 8 bits & the apple 7 (or vice versa). Just wondering whether it's a firmware issue.
 

tezza

Veteran Member
Joined
Oct 1, 2007
Messages
4,712
Location
New Zealand
This is a longshot but does the computer have the correct IOU chip? I had a case of weird video behaviour in a IIe once and the cause was an incorrect replacement IOU. I'd been using one designed for machines in North America rather than those in the PAL/240v world. As has been mentioned, colour in the Apple IIe depends on very subtle manipulation of NTSC signals. I didn't the specific symptoms mentioned (my problem was the video was offcentred and shimmied) but then again I wasn't looking for anything else.

I wrote an article about it which is here
http://www.classic-computers.org.nz/blog/2010-03-04-diagnosing-apple-ii-platinum-screen-sway.htm

Just check your number on the IOU chip and make sure the right one is installed.
 
Top