Model 1 Video RAM issue(s)

[Biff]

Hi,
when I originally received this "AS-IS" Model 1 from eBay - the screen acted like bit6 sticking in places - but not everywhere.
Thus the "@" signs sprinkled throughout when the screen should have been cleared.
As shown here: https://www.bifftech.com/model1/Before.jpg


So - I removed Z62, installed a socket and inserted a "new" 2102 from a pack that I bought from someplace a long time ago.
As shown in the second screen shot - the random "@"'s have gone away - but now it seems to have a new problem:

https://www.bifftech.com/model1/After.jpg

In the second picture - it is supposed to be displaying "MEM SIZE?"
Along with the usual

"R/S L2 BASIC"

The following screenshots show what happens on a single line, after I clear the screen, and then type "1" through "0".

https://www.bifftech.com/model1/1.jpg
https://www.bifftech.com/model1/2.jpg
https://www.bifftech.com/model1/3.jpg
https://www.bifftech.com/model1/4.jpg
https://www.bifftech.com/model1/5.jpg
https://www.bifftech.com/model1/6.jpg
https://www.bifftech.com/model1/7.jpg
https://www.bifftech.com/model1/8.jpg
https://www.bifftech.com/model1/9.jpg
https://www.bifftech.com/model1/0.jpg

I've tried swapping in another 2102 from my collection, and it had the same results.

Any suggestions?

thanks!

Jim

 

[gp2000]

Looks like there's some kind of problem in the lower address lines generated by the video circuitry.

Instead of generating a nice series of lower bits counting 01234567 it is doing 45454567:

01234567
R/S L2 B

45454567
L2L2L2 B

Form a binary perspective, bit 0 is fine. Bit 1 is mostly 0 but gets itself right for the last two. And bit 2 always stuck at 1. In fact, it looks like bit 1 is the AND of the correct bit 1 and 2.

This could well be due to a bad Z64 chip (a 74157). Or maybe it is Z65 (a 7493) which is generating bits 1 and 2 for the video circuitry. Given than bits 1 and 2 are the bad ones it does point the finger at Z65.

But I'm just working the numbers here and looking at the schematic. I've never replaced a chip in my life so do verify my theory with your own board and schematics before trying anything.

Doing some screen PEEKs might give more information but it'll be kinda hard to interpret the output.

 

[ldkraemer]

Let's start with the first screen and understand it. It is as follows:

MEM@SIZE       @@@@@@@@

which has 7 spaces between the two character strings

MEM@SIZE1234567@@@@@@@@

So, this is an accurate presentation of the characters in the first line.

Notice that the first character that is incorrect is the "@"
"@" = 0x40 = 64 Decimal, but you want the SPACE not "@".

The SPACE is " " = 0x20 = 32 Decimal, so what is different?

0100 0000 = 0x40
0010 0000 = 0x20
-------------------------
0^^0 0000 So, these bits are changing D5 & D6.


Continuing on, the string is

MEM@SIZE1234567@@@@@@@@

If we count the Characters that are mostly correct we have 15 characters {1..15}
and the 16th character is incorrect and repeating for 8 Characters. What power
of 2 gives us 16?

2^0 = 1
2^1 = 2
2^2 = 4
2^3 = 8
2^4 = 16 <------- So, A4 is STUCK HIGH causing the repeating pattern.

You can prove this by removing the CPU and using a PULLUP Resistor of 1K
with one end tied to +5 VDC. The other end of the resistor needs to go to the
Z80 CPU's Socket A4 Pin. This allows you to use another jumper to jumper
the 5ma SIGNAL at A4 of the Z80 CPU's IC to GND and create a LOW on A4.
You can measure this with a volt meter and follow A4 to the problem IC.
When you get that fixed, one of the two problems will be fixed. Then just fix
the D5 & D6 problem.

Larry