CPC 6128 - white screen, black border

[McArti0]

no way of knowing if their test is running as expected or not and the results can't be trusted.

Test is Pin 24 z80 WAIT is permanent Lo after Step 3

[McArti0]

Test 74LS373 by spider hardware

[Bryce]

@Bryce.
Notice that the clock inverter gets a heavy zero at the input - it has no chance of ticking even once. The freeze is sure and strong.On the address lines there is &B100 or &B101. I use rectifier diodes so they have a 60ns propagation time, they may not stop at the first LDIR address.

Hi McArti0,
      you may have misunderstood what I meant (or I explained it badly?).
Yes, the clock will definitely stop as planned with your setup, that's certain. My query was whether it runs properly before that point
and it seems you've just answered that question yourself with the screenshot. If you take a look at the diagram, the clock glitched
a few cycles earlier. So the clock was unstable and it can't be said with certainty that the CPU acted correctly from that point onwards
and the entire system was in an unknown state there after.

Bryce.

[McArti0]

it can't be said with certainty that the CPU acted correctly from that point onwards
and the entire system was in an unknown state there after.

This diagram is not mine. It is borrowed from Bread80. 

https://bread80.com/2021/06/03/understanding-the-amstrad-cpc-video-ram-and-gate-array-subsystem/
Fig.5

I understand what you are talking about, but for now it is not important to us because even if I have disturbed the program code, for now we want GA to stop with READY Lo and RAMRD Hi or Lo. And that's exactly what's happening. even the address is correct. You can statically check whether 373 is open and output responds to input and 373 save state. The next test will be to catch 373 in latch. i.e. the first RAM reading, READY Lo, RAMRD Lo without steps, using an external OR gate made on the 74LS02. the reading should be 00h.

[McArti0]

the clock glitched
a few cycles earlier. So the clock was unstable

I tried this. My tests show that the clock is only wrong when touching the contacts with a bare hand and rubbing the wires against each other when the clock is active. there is no interference from the wires as the antenna MW radio.

[McArti0]

never connect the pins this way.

I pushed the IORQ pin apart and now it doesn't connect to the socket without it.

[Bryce]

Any update? Did the RAM ever arrive and get installed?

Bryce.

[McArti0]

We are wating for 74LS02 and Breadboard for test 373. We have diods to test 244. We are waiting for some time.

mmldrm is reading z80 manual Fig 4-8.

[McArti0]

Did the RAM ever arrive and get installed?

If mmldrm fixes this CPC, he will have to give it back. Would you give away your first CPC in your life?  Besides, here mmldrm learns digital electronics.

[Bryce]

By the time it's fixed (when he gets around to replacing the RAM), he'll have spent so much money and time, that he could have bought his own one!

Bryce.

[McArti0]

@Bryce 

He started by buying an oscilloscope, I'm afraid to ask what kind, but I think it was RIGOL, so the CPC costs went first. 

[Bryce]

Rigol is a very good mid-range brand. I'd highly recommend them for hobby electronics. I use some of their devices too.
(but I already had my CPC when I bought them ).

Bryce.

[McArti0]

A new memory test occurred to me.

RAMDIS HI, ROMDIS HI (of course after plugin wired version PAL)

and grounded D5,D4,D1.

Z80 always takes the CDh,CDh,CDh,CDh...  i.e. CALL #CDCD (push to stack CDCD+3=CDD0)

Whole of memory fills by CDD0

1 1 0 0  1  1 0 1
1 1 0 1  0 0 0 0
_________________
1 1 0½ ½½0½

Its test other bits than FF (RST 38h)

I'll check if it works 

[McArti0]

@mmldrm

Try one think. Cross Pin21 Z80 RD to other RD like RD expansion port.

[mmldrm]

Finally the 74LS02 has arrived!

I will try the above tests too, if I understand them. 

My oscilloscope is a Zoyi ZT-702S: https://zoyitools.com/product/zt-702s/

I found it for £43, and figured it's something I can use for years after this. This is true for the breadboard and components I bought for the tests - I like playing with electronics and would like to do more hardware projects in the future.

If these tests from @McArti0 can reduce the number of chips that I desolder and replace, I am less likely to destroy the motherboard in the process. Unorthodox, but I'm happy to proceed with caution on this.

Thanks for staying with me this far on the journey. I know there is a variety of opinions on the path we have taken.

[McArti0]

Ok. Make OR Gate on breadbord. (From 2 NOR gate in 74Ls02).

[mmldrm]

Ok. Make OR Gate on breadbord. (From 2 NOR gate in 74Ls02).

I could do this when I was 18. 😅 I'll go read up on it and get back to you. 🫡 

[Bryce]

Finally the 74LS02 has arrived!

I will try the above tests too, if I understand them. 

My oscilloscope is a Zoyi ZT-702S: https://zoyitools.com/product/zt-702s/

I found it for £43, and figured it's something I can use for years after this. This is true for the breadboard and components I bought for the tests - I like playing with electronics and would like to do more hardware projects in the future.

If these tests from @McArti0 can reduce the number of chips that I desolder and replace, I am less likely to destroy the motherboard in the process. Unorthodox, but I'm happy to proceed with caution on this.

Thanks for staying with me this far on the journey. I know there is a variety of opinions on the path we have taken.

The scope is ok for basic "signal present" checking, but only up to about 3MHz for proper diagnosis. That's a bit on the low side even for retro computer repair.

Bryce.

[mmldrm]

Bryce.

Ouch. That would be a bit disappointing. It's the one suggested on page 2 of this thread. 

The spec shows 10MHz. Am I misunderstanding it? (Quite possible.)

[McArti0]

48Msps
Although zt-703 and ... I dreamed about Rigol

[McArti0]

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

Fig.7
See when n244EN and nCPU ADDR are LOW for the first time together.

[mmldrm]

Thanks. Ok I got the 74LS02 set up on the breadboard. (Ignore 74HC02 - TinkerCAD had no 74LS02.)

Where is nCPU ADDR on the Z80? Not sure where to connect that.

Let me know if 

1. IC118 should be replaced (i.e. not wires)?
2. should Z80 pin20 and pin24 still be connected?

[SerErris]

Bryce.

Ouch. That would be a bit disappointing. It's the one suggested on page 2 of this thread.

The spec shows 10MHz. Am I misunderstanding it? (Quite possible.)

That is what you can get for that price and it is on the low end of usable, however still massively better than a multimeter for digital cirquits. You can actually see what the lines are doing (up, down etc). And exact timing (e.g. fractions of a µs) is not important here.

Also the next better suited would be a proper oscilloscope like the Rigol 1054z (or any other entry level brand and model). But you know, it is even used for 200-300€ .. So I did not recommend it here, but that would be actually my normal recommendation to get into an Oscilloscope.

Nowadays you can get the new Rigol DHO800 series, with the DHO804 the cheapest option. That goes for ~400 Euros. This is a huge upstep with 12bit resolution vs 8bit of the old Rigol and 100Mhz vs. 50Mhz.

[McArti0]

Thanks. Ok I got the 74LS02 set up on the breadboard. (Ignore 74HC02 - TinkerCAD had no 74LS02.)

Where is nCPU ADDR on the Z80? Not sure where to connect that.

Let me know if

1. IC118 should be replaced (i.e. not wires)?
2. should Z80 pin20 and pin24 still be connected?

Have You DM74 or SN74?

/CPU (nCPU ADDR)  Pin14 GA 40010 or Pin2 LS153 or Pin15 AY3 or Pin11 PAL 40030/1

1) wires is better.
2) NO. Now the Force is with us.

When CPC stops on address lines A15-A0 you will have address #7F89
Z80 Pins 5,4,3,2,1,40,39... is 30

Next on Pins
0,1,1,1,  1,
1,1,1,  1,0,0,0, 1,0,0,1
RD Hi, WR Lo, IORQ Lo, nCPU Lo, Pin1 244 Lo.

You can tests D0-D7 Pin2-Pin14 RAM chips

[Bryce]

Bryce.

Ouch. That would be a bit disappointing. It's the one suggested on page 2 of this thread.

The spec shows 10MHz. Am I misunderstanding it? (Quite possible.)

The quick explanation is that the quoted 10MHz is for a sinewave signal. To measure and display a squarewave the bandwidth of the scope needs to be around 5 times the frequency of the signal. I won't go into further details here, because I wrote a more detailed article about this a while ago, which you can read here:  https://www.cpcwiki.eu/forum/amstrad-cpc-hardware/understanding-retro-electronics/msg144677/#msg144677

However, as @SerErris quite rightly mentioned above, a 50MHz Scope will cost at least €200, even for a simple budget device.

Bryce.