Random characters

[razza]

I have another fault on my new toy, I have done some Googling and it appears RAM/ROM are common faults that cause random characters. I thought before I start faulting I would see if anyone has seen this issue and can point me?


When I type I get a second corrupted set of characters...

 

[razza]

Sorry, I forgot to say, it's a 2001N

 

[daver2]

It is most likely the 2114 video RAM at either F7 and/or F8.

However, not only do you have a data issue, but also an address issue. The Commodore message is repeated, but in a corrupt data format - in general with a data bit fault.

This fault is amplified by your video. If you type all of the keys on the keyboard in order, what 'shadow' character do you obtain vs what you typed? If you convert the characters typed and displayed into binary (remembering that the PET uses PETSCII) you may be able to identify which bit(s) are affected.

Likewise, if you look at the positions on the screen and convert these into addresses (in binary) you should be able to identify the address offset.

A faulty 2114 video RAM can also present these types of fault.

Are F7 and F8 already in IC sockets or not?

There can be other reasons for this fault, but this is the most likely cause...

Dave

 

[razza]

Hi Dave,
Thanks. F7 and F8 are not socketed. I don't like removing ICs unless I have to, partly as my desolder station never does a great job! That said I have a couple of 2114's on order for a VIC-20 I am working on, I can use those on the PET as that gets priority

I have mapped the codes with a few keyboard/other errors and will convert to binary when I have time tomorrow using this resource - https://www.pagetable.com/c64ref/charset/

A = !
B = "
C = #
D = $
E = %
F (came up as &) = &
G = '
H = (
I =)
J = *
K = +
L (key must still be broken) = N/A
M = -
N (must still be broken) = N/A
O = /
P = <space>
Q = !
R = "
S = #
T = $
U = %
V = &

 

[daver2]

It looks like the codes are all in order. I suspect @ should give you a <SPACE>.

Dave

 

[razza]

I suspect @ should give you a <SPACE>.

Interesting, in my original test to determine the shadow character response yielding the table I posted previously, the shadow text never overwrote existing text, so I might have given you duff info. The shadow text seems to depend on the existing character the shadow response is overwriting...

I also noticed that there is a defect with a flickering unknown character...


Video here...

It seems to me the best approach is to replace the two 2114 memory ICs.

Thanks for your ongoing support

 

[razza]

Two new 2114's and boom!



Thanks Dave!

 

[daver2]

No problem...

Dave

 

[Hugo Holden]

Two new 2114's and boom!

View attachment 1294927

Thanks Dave!

I pointed out on a number of PET repair posts that practically every PET repair has involved faulty 2114 SRAM. It appears most of the original chips have got problems by now. And some stocks of vintage ones are not wonderful either. Though of course you need to have a known good stock of 2114's to replace them with. I collected some very good ceramic package types (equivalents) made by Intel. And I have pre-tested them all too, so there is no doubt when I fit them I know that I have good SRAM.