CPC 6128 - white screen, black border

[mmldrm]

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.

Hi. I do appreciate the suggestion of the ZT-702S/ZT-703S. I'm just clueless regarding what level of oscilloscope is sufficient for this kind of work.

So I was a little disheartened to be told that the ZT-702S isn't really up to the job of helping to diagnose the issue here, after buying it for that purpose.

I've requested a return on the ZT-702S. Hopefully that will work out and I can spend the money on a Dandanator instead.

I can also almost hear the shouts of "replace the damn RAM chips" and to be honest it does regularly cross my mind to do this, but I would love for a positive confirmation via McArti0'd tests before desoldering anything.

I hope the above doesn't sound too ungrateful! Thanks again for taking the time to reply here.

[mmldrm]

Have You DM74 or SN74?

SN74 (SN74LS02N).

[McArti0]

McArti0'd tests before desoldering anything

Let's GO! 

[Bryce]

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.

Hi. I do appreciate the suggestion of the ZT-702S/ZT-703S. I'm just clueless regarding what level of oscilloscope is sufficient for this kind of work.

So I was a little disheartened to be told that the ZT-702S isn't really up to the job of helping to diagnose the issue here, after buying it for that purpose.

I've requested a return on the ZT-702S. Hopefully that will work out and I can spend the money on a Dandanator instead.

I can also almost hear the shouts of "replace the damn RAM chips" and to be honest it does regularly cross my mind to do this, but I would love for a positive confirmation via McArti0'd tests before desoldering anything.

I hope the above doesn't sound too ungrateful! Thanks again for taking the time to reply here.

No issue. I won't shout anything  . It's your machine and it's your decision how you want to fix it. If you intend always using your Scope where a computer is nearby, there are some very good USB Scopes around, that will give you a lot more bandwidth and all the features for under €200.

You can get a Hantek PC 6074BC  (4 Channel 70MHz USB Scope) for €180.

Bryce.

[mmldrm]

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

Thanks, I used pin15 of AY3.

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

I keep getting high on the address lines. A0-A14, all 1. Strangely, A15 was 0. I tried this three times and got the same results.

Not every startup seemed to halt. I saw the white screen and black border occasionally, but presumably this kind of system halting is not exact.

RD Hi, WR Lo, IORQ Lo, nCPU Lo, Pin1 244 Lo.

I confirmed nCPU Lo and Pin1 244 Lo. Are RD/WR/IORQ all Z80 pins? Or expansion port? Or... are those the same thing anyway? I guess I can confirm all of them tomorrow.

I will also check the wiring again, but I think it is all correct. Please let me know if my breadboard diagram above is incorrect.

[mmldrm]

No issue. I won't shout anything  . It's your machine and it's your decision how you want to fix it. If you intend always using your Scope where a computer is nearby, there are some very good USB Scopes around, that will give you a lot more bandwidth and all the features for under €200.

You can get a Hantek PC 6074BC  (4 Channel 70MHz USB Scope) for €180.

Thanks for the info, but even €180 is hugely more than I can justify. Maybe if I survive this process and fix a few Amstrad, Spectrum, or Commodore computers then I will treat myself to a Hantek USB scope.

[McArti0]

Are RD/WR/IORQ all Z80 pins?

You are no longer allowed to ask such questions.
https://www.cpcwiki.eu/imgs/4/4a/CPC6128_Schematic.png
Signal name is constant.

Please let me know if my breadboard diagram above is incorrect.

Have too many joins. +5V should join one wire near Pin14, one hole above. Groung one wire near Pin7,one hole under.

Please, show me photo of all setup.

[mmldrm]

Are RD/WR/IORQ all Z80 pins?

You are no longer allowed to ask such questions.
https://www.cpcwiki.eu/imgs/4/4a/CPC6128_Schematic.png
Signal name is constant.

😆 Good answer. I think I am still learning that the state is constant across the board. No co-processors, etc. I sometimes found pins with the same name but on different ICs, and not sure which is the 'right' one. Now I know that there is no wrong answer.

Have too many joins. +5V should join one wire near Pin14, one hole above. Groung one wire near Pin7,one hole under.

I found my mistake. I connected pins 2->3 and 3->4, making them all the same input. It should have been 2->4 and 3->4.   Will try again with it corrected.

Please, show me photo of all setup.

It seems difficult to show everything in enough detail, but I will try (in a separate reply).

[mmldrm]

Please ignore 2x green hooks on Z80. Those are from our previous test and not connected to anything.

All clips are connected when I power on, including the pin1 LS244->R143.

Occasionally on startup I get the white screen and black border, like the Z80 didn't halt. Other times the screen remains 'dead', so I continued.

I checked Z80 IORQ/RD/WR, LS244, and R143 for hi/lo, and they were all correct when reading A0-A15. If not correct values then I powered off, waited 5s, and powered back on to try again.


I took GND/+VCC from IC108 because the pins were easiest to connect hooks to.

A0-A15 all hi.  Still possible that I am doing something wrong, but it feels like I checked everything twice...

[McArti0]

Check whether when you short both 74LS02 inputs, pin 5,6 to ground, the CPC freezes.
You can add a second gate to the output by connecting PIN13 to Pin1 and pin11,12 to the group of pin2,3,4. 74LS02

If this does not work stably, we will add support with a diode pin 20 to pin 24 as before.

[McArti0]

found my mistake. I connected pins 2->3 and 3->4, making them all the same input. It should have been 2->4 and 3->4.   Will try again with it corrected.

It really doesn't matter at all.

[mmldrm]

found my mistake. I connected pins 2->3 and 3->4, making them all the same input. It should have been 2->4 and 3->4.  Will try again with it corrected.

It really doesn't matter at all.

Only now do I realise that I was visualising current going from left-to-right across the 74LS02 diagram and not from right-to-left. 

[mmldrm]

Check whether when you short both 74LS02 inputs, pin 5,6 to ground, the CPC freezes.
You can add a second gate to the output by connecting PIN13 to Pin1 and pin11,12 to the group of pin2,3,4. 74LS02

If this does not work stably, we will add support with a diode pin 20 to pin 24 as before.

Thanks I will try with the second gate and maybe the diode. Probably tomorrow - too late tonight.  

Also, I got a refund for the oscilloscope. So if I manage to break you with my lack of knowledge/skills/tools, I can order that Dandanator Mini. 

[mmldrm]

I just found something interesting on the LS02. Two lo inputs (0v and 0.3v) are giving me hi output. Pins 5 and 6 input, pin 4 output.

I checked the tolerances for my SN74LS02N here. Max lo is 0.8v, min hi is 2v.

Not sure if this is causing the problem, but it sure surprised me.

(I forgot to add, the second NOR gate was on pins 11,12,13 when I measured this.)

[McArti0]

Very good! You're starting. understand what's going on here. But what you are writing about is the correct operation of what we want to build. When Pin1 244 and nCPU are low, only then we want the diode and pin1 LS02 to be low.
Maybe you should add a capacitor between pin 14 and pin 7, i.e. on the LS02 power supply .

[mmldrm]

It worked!! 

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

Here are the test results.

Apologies for use of 0, 1, and 3. The '3' readings were unexpected and I wasn't sure how to represent them in the table.

Notes:
0 = lo (0-0.2V)
1 = hi (4-5V)
3 = 3.67V except pin14 which was always 3.23V

Code Select Expand

    ---------------------pins---------------------
IC.  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16
127  0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
128  0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
129€ 0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
130  0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
131  0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
132  0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
133  0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
134  0  3  1  0  3  3  3  1  3  1  3  3  3  1  1  0
I was working from IC134 up to IC127. At the point marked (€) the screen suddenly brightened. I retested other pins and the 3.2V-3.6V pins became 4V when retested. I didn't have time to test all of them before the voltage dropped again.

After noting some of these new values, just as I touched ground again with multimeter probe (and nothing on +ve probe), screen dulled and back to 3.2V-3.6V. Probably a coincidence, but maybe not.

I wondered if this shows a power problem. I checked some random pins on other ICs and found other 3.67V values.

Hmm, now I checked the data sheet for for the 4164 chips and minimum hi value seems to 2.4V, so maybe this isn't the "smoking gun" that I originally thought it was...

Anyway, lots of consistent values across the RAM chips.

[McArti0]

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

Ps. The address lines in memory change 4 million times per second. So there's not much to measure on them.

[mmldrm]

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

I'm 99% sure these were the values I saw, because I only checked the memory ICs after passing the CPU A0-A16 address values 'test'. But I'll do it again today to confirm.

Ps. The address lines in memory change 4 million times per second. So there's not much to measure on them.

Oops, I thought everything was halted during this test.

[McArti0]

Have you seen the picture frame? Because that means it wasn't stopped.

[mmldrm]

No picture frame. Just a blank screen.

It looked like static, but with horizontal lines. Then I connected LS244-R143 and get a fixed bright screen. (White? Difficult to tell on the monitor.)

[mmldrm]

Tried again today. No static screen or white screen, just blank/dark screen the whole time.

1. Switched on (with all connections as described previously, except LS244-R143 is disconnected)
2. Connect LS244-R143
3. Check Z80 pins 20/21/22 are lo/hi/lo
4. Check Z80 address pins
5. Measure RAM values

So Z80 address (A0-A15) are correct.

10010001 11111110

DRAM values are the same for all 8 ICs.

Today I measured DRAM pins 1-16 in correct order. Yesterday I accidentally measured 1-8, then 16-9, and had to rearrange the table. Maybe I made a mistake there, because now Pin13 is lo, not hi.

Code Select Expand

     ---------------------pins---------------------
IC   1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16
127  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0
128  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0
129  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0
130  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0
131  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0
132  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0
133  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0
134  0  3  1  0  3  3  3  1  3  1  3  3  0  1  1  0

[McArti0]

This is OUT #7F89  , .... #FF

D2 lines shows that z80 does not break through with value #89

Check pin1 LS373 should be Hi.
Check R160,R161 should be Hi

Ps. You can all connect before turn on cpc

[mmldrm]

Pin1 LS373 is Hi
R160 is Lo
R161 is Hi

I do still have wires replacing IC118, if that makes a difference.

[McArti0]

Check MREQ should be Hi

R160 is Lo , very bad.  Check pin16 40010 should be Hi and wire 5v to pin16 in socket for PAL 40030 (i belive is not present)

[mmldrm]

MREQ = hi
pin16 of 40010 = hi

Correct, there is no IC below the 40010.

Should I solder a wire from +5v to the pin16 point on the board?