1978 Commodore PET project

[ajgriff]

See attached.

Alan

[BrackenFix]

I am such a fool. I have that exact document printed out but couldn't marry up A6, A5 etc...the pin numbers are right next to them.

We are doomed!!

Thank you

[ajgriff]

Nothing foolish about it. It's just that you're on a learning curve. Stick at it.

Alan

[SiriusHardware]

Sorry, I was assuming you knew how DIP IC pins are numbered. With the 'notch' at the top, marked by the 'U' in this crude sketch below, pin 1 is usually the first pin anticlockwise to the left and marked with a 'dot' or 'dimple' and then the pin numbers carry on anticlockwise around the device, so for an 18-pin chip like the 6114 / 2114 the pin numbering is as below.

1   U   18
2       17
3       16
4       15
5       14
6       13
7       12
8       11
9       10

If you meant you wanted to find the pinout in order to see what the various pins do, then here is a llink to a 2114 datasheet.

https://www.princeton.edu/~mae412/HA...heets/2114.pdf

Edit: I see events have overtaken me.

[BrackenFix]

UF8 Pin 14 = 0.22v
UF8 pin 13 = 0.21v
UF8 pin 12 = 0.03v
UF8 pin 11 = 0.04v

UF7 pin 14 = 1.11v
UF7 pin 13 = 3.62v
UF7 pin 12 = 0.03v
UF7 pin 11 = 0.03v

[SiriusHardware]

That's a better result than I dared hope for with a meter.

To amplify what Mark said earlier in the thread, the operating system starts by filling the display memory with 20 (hex), which is the character code for blank / space. It then goes on to write a few other bits of text in the upper left area, the Commodore BASIC and Bytes Free messages.

As mentioned earlier, unlike most RAM, the display memory chips are selected 100% of the time and they are held in read mode for almost all of the time while the display hardware runs continually through all of the screen memory addresses.

What this means is that, if the display memory really does mostly contain 20 (hex) or 0010 0000 (Binary), then the average voltages being output by the display memory SD0-SD7 bus should look like this:

SD0 UF8 Pin 14 = 0.22 = '0'
SD1 UF8 pin 13 = 0.21 = '0'
SD2 UF8 pin 12 = 0.03 = '0'
SD3 UF8 pin 11 = 0.04 = '0'
SD4 UF7 pin 14 = 1.11 = '0'
SD5 UF7 pin 13 = 3.62 = '1'
SD6 UF7 pin 12 = 0.03 = '0'
SD7 UF7 pin 11 = 0.03 = '0'

Turning those binary digits 90 degrees to the right, we get 0010 0000 binary, or 20 hex.

So what this suggests to us is that (a) the CPU is successfully writing the value 20 hex to most of the screen locations initially, then (b), the screen RAM also outputs the correct values when it is being scanned by the display hardware.

However, instead of the expected blanks / spaces across the screen, you are instead getting a screen mostly full of '0'.

The difference between the codes for those two characters is:

Space =
0010 0000
'0' =
0011 0000

We can verify this again by looking at a few of the other characters, for example the first three on the screen should be
'###' = 0010 0011
but instead they are
'333' = 0011 0011

And there again we can see that the fifth bit in from the right is '1' when it should be '0'.

Somewhere between the output of the display RAMs and the video output, the state of the fifth bit from the right is being changed from a '0' to a '1'.

The next items in the chain after the display RAM outputs are the video latch UF9 and the character generator ROM UF10.

Could you now measure the voltages on these pins... this time you will be jumping all over the IC to get to the next pin so be careful to make the measurements on the right pin numbers.

UF9 pin 2
UF9 pin 19
UF9 pin 5
UF9 pin 16
UF9 pin 6
UF9 pin 15
UF9 pin 9
UF9 pin 12

[SiriusHardware]

Just an observation, the fact that the 5th bit from the left is averaging 1+ volts when all the other bits expected to be '0' are below a quarter of a volt could still signify a RAM fault, let's see what the outputs of UF9 look like.

[Mark1960]

That 1.11v on SD4 is a bit suspicious, I’d expect it to be lower for a definite mostly 0.

If its measuring high on the output of the 74ls373 it would not be certain that the 74ls373 was at fault.

I’m wondering if we should ask brackenfix to try a 470 ohm resistor from UF7 pin 14 to ground. Just for a moment to see if the display changes.

If it could be done while the pet initialises the display and then removed it might even tell us if the 2114 or the 74ls244 was the possible source of the issue.

[SiriusHardware]

Agree with all of the above, let's first see what we have on the outputs of UF9.

[SiriusHardware]

I have an idea about this, let's Phone A Friend.

ScottishColin? If you can get a moment, could you make the same measurements with a meter, not your scope, on the SD0-SD7 outputs of your video display RAM ICs, as per post #79?

Same conditions - machine turned on, running BASIC 2, idle and waiting for input.

[ScottishColin]

UF8 Pin 14 - 0.55v
UF8 pin 13 - 0.50v
UF8 pin 12 - 0.39v
UF8 pin 11 - 0.36v

UF7 pin 14 - 0.43v
UF7 pin 13 - 3.6v
UF7 pin 12 - 0.38v
UF7 pin 11 - 0.38v

I had to replace mine so they are marked as TMM314APL-1 2114L2 chips which I got for £3.00 ea off ebay. They don't seem to be there now - Cricklewood Electronics do have 2114 ICs for £5.70 each if you're needing them.

Colin.

[ScottishColin]

In anticipation, here's the next reading:

UF9 pin 2 - 0.8v
UF9 pin 19 - 0.75v
UF9 pin 5 - 0.52v
UF9 pin 16 - 0.41v
UF9 pin 6 - 0.53v
UF9 pin 15 - 3.5v
UF9 pin 9 - 0.31v
UF9 pin 12 - 0.32v

Again replaced on my adventure - IC is marked as a TI chip 74LS373N which is also available through Cricklewood if required for £1.02 plus P&P - search for 74LS373 if needed.

Colin.

[BrackenFix]

I am so grateful for all your help.

UF9 pin 2 = 0.54
UF9 pin 19 = 0.52
UF9 pin 5 = 0.33
UF9 pin 16 = 0.20
UF9 pin 6 = 4.00
UF9 pin 15 = 3.74
UF9 pin 9 = 0.14
UF9 pin 12 = 0.19

[ScottishColin]

Something up with UF9 pin 6/UF7 pin 14 which ties in with post 86.

Colin.

[BrackenFix]

Just rechecked and same readings (worried i messed up)

[SiriusHardware]

Thanks for the prompt action, Colin - yours is more like we would ideally expect, all of SD0-SD7 comfortably low except for SD5, which we would expect to be high. Thanks for also taking the time to do the outputs of UF9, where again there is one bit high, seven low as expected.

Just caught BrackenFix's UF9 measurements and there we can see that the 1.1V coming from the SD4 line is being converted to a logic '1' by the time it comes out through the other side of UF9, and this perfectly explains why all the character codes are being modified by +10 (hex) before they go to the character generator ROM which I believe we can now rule out.

BrackenFix, do you have any low value resistors, around 470R or thereabouts?

[BrackenFix]

I have so many resistors here and cant find anything like that

[SiriusHardware]

Do you have any 1Ks? Can you find two or three of those? Or 220R or 330R?

[BrackenFix]

Ok, I really don't understand resistors. I found what I think is a 47k but that's too much

[BrackenFix]

I have found what I think are 1k resistors though. I have loads

Brown - Black - Red - Gold

I also have a 10k