CPC 6128 - white screen, black border

[McArti0]

ok. Now test 74LS244.

When connect red wire whole CPC halted, screen too.

You have to be sure that you have drawn 0V on Pin1 LS244,
if not, you connect again.
Then check if D0-D7 on the left and right are the same.
Work for 6 hands.

[mmldrm]

ok. Now test 74LS244.

How do I test it? Do I start with the same ROMDIS and RAMDIS steps, then measure voltages?

[McArti0]

no, I added it above. Sorry 

[mmldrm]

ok. Now test 74LS244.

When connect red wire whole CPC halted, screen too.

You have to be sure that you have drawn 0V on Pin1 LS244,
if not, you connect again.
Then check if D0-D7 on the left and right are the same.
Work for 6 hands.

Should I do the same ROMDIS and RAMDIS steps first?

And 40031/IC118 is still replaced by 4 wires?

[McArti0]

NO

YES

[mmldrm]

Thanks. I had to have the red wire connected before switching on, or it didn't do anything. I hope that is correct.

Results are interesting I think. I will share the full measurements across the IC, and then my comparisons. Maybe I got them wrong, so it is better you see everything.

Code Select Expand

IC114 Pin Measurements
----------------------

    pin    voltage
    1    0
d0    2    1.4
    3    0.09        d7
d1    4    4.15
    5    0.09        d6
d2    6    0
    7    0.09        d5
d3    8    4.15
    9    5        d4
    10    0
d4    11    4
    12    5        d3
d5    13    0
    14    0        d2
d6    15    0
    16    5        d1
d7    17    0
    18    5        d0
    19    0
    20    5


Code Select Expand

Comparisons
-----------
    left    right
D0    1.4    5
D1    4    5
D2    0    0
D3    4    5
D4    4    5
D5    0    0
D6    0    0
D7    0    0
Everything matches, except D0 has a strange 1.4v value.

I double checked these to make sure. But I can do the same test again, if it is good to have two sets of results.

[McArti0]

Thanks. I had to have the red wire connected before switching on, or it didn't do anything. I hope that is correct..

This is an indication that after a certain point there are no writes to GA 40010.

[mmldrm]

This is an indication that after a certain point there are no writes to GA 40010.

Do I need to check traces and signals now?

You have done an amazing job guiding me with a multimeter, but I expect the oscilloscope to arrive in 5 days. I hope that helps to find the fault (or faults).

[McArti0]

much better will be to buy Dandanator with test of components.

now we have 1.4v on D0. test it again and see if READY is 0

[mmldrm]

much better will be to buy Dandanator with test of components.

now we have 1.4v on D0. test it again and see if READY is 0

I seem to be very good at buying the wrong things (RAM, oscilloscope) at the wrong times. 😅

Ran the test again. With 1.4v on D0, READY (pin22 on 40010) is 0.

[McArti0]

We won't stop where we should.

Look Fig. 8 here...   https://bread80.com/2021/06/03/understanding-the-amstrad-cpc-video-ram-and-gate-array-subsystem/

N244EN have 0 two times. We find the first meaningless one. 

[McArti0]

Ok first We must short IORQ and READY , and turn on CPC.

pin 20 Z80 should be 0

First instruction is:

LD BC,#7F89
OUT (c),c

Our Z80 is stopped at write to GA

next we add RED connect and check if Pin1 244 is 0. If not, RED to reconnect.

GA is stopped.

now 244 is open and Left and Right side should be 89h 1000 1001.

[mmldrm]

After I short IORQ and READY, then turn on the CPC, pin20 of Z80 is 0 (well, 0.14v so I guess 'low' is correct).

Should I continue to short IORQ and READY? Or stop?

I removed IORQ->READY and added the RED connection. But voltage was 5v.

I expect you will say I must keep IORQ->READY so the Z80 remains stopped, while I do RED connect and measurements. But I wanted to confirm.

[mmldrm]

More simply, should I continue to short IORQ->READY while I also do IC115->R143 (red wire)? So both wires at the same time. (4 hands!)

(Also, I now have an oscilloscope, if that is useful during these tests.)

[McArti0]

Yes permanetly stop Z80 by continue iorq-ready and next permanetly stop GA by continue Red.

[mmldrm]

Thanks. I thought that must be correct.

Is there a way of doing this with less than 3 people? 😅 I did see that I could use a little solder to fix the wires to the pins, but I'm not doing this because I don't own this 6128.

If not, it's ok, I have two teenagers that can help.

[McArti0]

MacGyver could do it himself with one hand.

You can do It on stiff wires. z80 wait line (Ready) and iorq pin 20

[mmldrm]

Ok first We must short IORQ and READY , and turn on CPC.

pin 20 Z80 should be 0

First instruction is:

LD BC,#7F89
OUT (c),c

Our Z80 is stopped at write to GA

next we add RED connect and check if Pin1 244 is 0. If not, RED to reconnect.

GA is stopped.

now 244 is open and Left and Right side should be 89h 1000 1001.

I'm back! Sadly no MacGyver here, but a few kind family members are better than nothing.

Left and right sides were:

Code Select Expand

    Left  Right
D0  1.5  5
D1  0    0
D2  0    0
D3  0    0
D4  5    5
D5  0    0
D6  5    1.5
D7  0    0
Maybe that is right and left? Left=pin2-17. Right=pin3-18.

Anyway, 10001010 I think. 8Ah? 🤔 Not quite right.

I will try again when my helpers are home, to get two sets of data .

[McArti0]

Have you often confirmed that pin1 244 is 0?

[SerErris]

I am not sure I do understand what you mean by stopping GA, GA will only stop if you stop the clock. Otherwise it will continue to output catch clock an will send RAS CAS to memory. And that will influence what you see on the right side of LS244.

You may hunt a ghost here esp. with the only tool available being a voltmeter.

Also if you see 1.5V on the left side, you have either not stopped the CPU or something is influencing the D7 line.

And again a Voltmeter is not good to troubleshoot digital logic. 

[Bryce]

I'm starting to think that this is some sort of never ending April fools joke. The initial failure you showed in the very first post is a classic symptom of a failed RAM and by far the most likely culprit. But instead of just swapping them and getting on with life, you've now spent over a month doing ends tests of questionable usefulness, which don't really tell you anything for certain and some of which were definitely totally pointless.

So I really have to ask. Why don't you just swap the RAMs?

Bryce.

[mmldrm]

I'm starting to think that this is some sort of never ending April fools joke. The initial failure you showed in the very first post is a classic symptom of a failed RAM and by far the most likely culprit. But instead of just swapping them and getting on with life, you've now spent over a month doing ends tests of questionable usefulness, which don't really tell you anything for certain and some of which were definitely totally pointless.

So I really have to ask. Why don't you just swap the RAMs?

I genuinely apologise if this comes across as trolling or something. It isn't. I'm just trying to follow instructions from whoever is willing to take the time to help me.

As it isn't my 6128, nor my budget, I've been trying to avoid purchasing anything unnecessarily. For example, I'd love to go buy a Dandanator, but I can't justify it for this one-off investigation.

I did buy 64Kb of RAM (and sockets), but I think McArti0 managed to diagnose that the fault lies elsewhere (IC114/IC115 maybe?).

I could replace the RAM. I just hoped that we could get a definite diagnosis of where the fault lies first, rather than risk my soldering skills introducing more doubt.

I did buy a oscilloscope, so if that in any way helps to diagnose at this point then I'm more than willing to justify the purchase by using it.

[mmldrm]

I am not sure I do understand what you mean by stopping GA, GA will only stop if you stop the clock. Otherwise it will continue to output catch clock an will send RAS CAS to memory. And that will influence what you see on the right side of LS244.

You may hunt a ghost here esp. with the only tool available being a voltmeter.

Also if you see 1.5V on the left side, you have either not stopped the CPU or something is influencing the D7 line.

And again a Voltmeter is not good to troubleshoot digital logic.

I do have an oscilloscope now (the one that you suggested), if that gives more options to try.

(I might also buy those mini grabber clips, rather than trust the assistance of my kind but inexperienced family members.)

[mmldrm]

Have you often confirmed that pin1 244 is 0?

I just bought some hook probes (now that I know that the are suitable) and evicted my family members from the team.

These should arrive by Friday, and will allow more reliable results.

[McArti0]

I'm starting to think that this is some sort of never ending April fools joke..

Then prove to him that his memory is faulty. So far I have found out that the 3rd byte of the Rom has the value 61 instead of 89. If at all he did it well with 3 people. For now, it turns out that z80 writes to frames. And that the ram has the correct contents. Maybe you could stop being mean and start helping.