CPC 6128 - white screen, black border

[Bryce]

How set 0 color to purple.
All ground on z80 side. (Pin2 RAM)
A) Ready/Wait.  -----|>|------  IORQ and keep (Connect by diode, catode at IORQ)
B) start cpc or reset.
C) ground D0 and keep , next ground D3 and keep, quick short ground D7 and remove all ground
D) ground D0 and keep, Hi on D6 (best way by serial 4x diodes ) and keep, quick short ground D7 and remove all ground and Hi set.

Background screen should be purple.
But for this to work, the 244 must be operational.

You are aware, that grounding a data pin will mean that the IC driving that data line will then have a direct short to ground with absolutely no current limiting if it's trying to pull that line high (trying to send a "1") and will most likely destroy the output? It may survive if there's a pull-up resistor or if the short is only for a few seconds, but otherwise you've fried a chip.

Bryce.

[McArti0]

You are aware, that grounding a data pin will mean that the IC driving that data line will then have a direct short to ground with absolutely no current limiting if it's trying to pull that line high (trying to send a "1") and will most likely destroy the output?

Bryce.

I totally agree. That's why I wrote "best way by serial 4x diodes" 5v - 4x0.6v = ~2.5v. 
I checked that the 5v to 0v short circuit current for cmos is 20mA. 

[Bryce]

You are aware, that grounding a data pin will mean that the IC driving that data line will then have a direct short to ground with absolutely no current limiting if it's trying to pull that line high (trying to send a "1") and will most likely destroy the output?

Bryce.

I totally agree. That's why I wrote "best way by serial 4x diodes" 5v - 4x0.6v = ~2.5v.
I checked that the 5v to 0v short circuit current for cmos is 20mA.

Ok, but just for my understanding... When he has done all of this, what exactly will this have shown? Or how will it bring him any closer to knowing what the route cause of the issue is? I have to say, I'm struggling at this point to understand the goal of many of these experiments. Testing the 244 is a simple matter of comparing the inputs and outputs when the /E pin goes low and this can be done with any cheap multimeter without any jumpers (in fact back when I didn't have the possibility to own fancy equipment, I used to do this with an LED and a resistor). Also, the chances of a 244 failing are 1000's of times less likely than a RAM failing.

Bryce.

[McArti0]

@Bryce

Test whether memory works. And memory writes and reads. Test if 244 works. Test whether you can set the color in GA. RAM works but may be unstable, then mmldrm will buy dandanator mini, which we have agreed on for a long time by PM. But we must be sure that GA shows colors. We should be done with the 373 test, but mmldrm doesn't have control over the 74LS02 gates yet. His increase in experience is also important here.
Now the interesting thing is that z80 writes memory, GA reads memory and z80 does not read through 373.

[Bryce]

@Bryce

Test whether memory works. And memory writes and reads. Test if 244 works. Test whether you can set the color in GA. RAM works but may be unstable, then mmldrm will buy dandanator mini, which we have agreed on for a long time by PM. But we must be sure that GA shows colors. We should be done with the 373 test, but mmldrm doesn't have control over the 74LS02 gates yet. His increase in experience is also important here.
Now the interesting thing is that z80 writes memory, GA reads memory and z80 does not read through 373.

So to put it in perspective... If his car breaks down and an experienced mechanic says "It's your fuel pump, they're always failing", he should still start by testing every single other component on the vehicle before he goes and buys a cheap pump? Even if some of those tests could easily result in him damaging something much more expensive on the car?

As I tend to document everything, I've just gone through the 100+ CPC repairs I've done over the last few years. I have never once needed to swap either the 244, nor any other TTL chip, other than a multiplexer once or twice and the keyboard driver.

Bryce.

[McArti0]

If his car breaks down

Its NOT his car.

Even if some of those tests could easily result in him damaging something much more expensive on the car?

YES! because z80 is in socket. RAM not.

If I confirm that the ROM contains the correct procedure for using RAM and it does not work because of RAM and not 373, it will solder 256 RAM pins.

Think of it as gathering mmldrm experience and my fun in coming up with tests without an oscilloscope.

[McArti0]

CPC stepper.

Stepper can go to next M1, next IORQ, next WR.

At least one crocodile must always hold an diode.

First hold M1 crocodile. turn on CPC. CPU halt at addr 0000

next hold IORQ crocodile to other diode

remove M1 crocodile.  CPU halt at WR IORQ after addr 0004

hold M1 crocodile .

remove IORQ crocodile. CPU halt at addr 0005

hold IORQ crocodile .

remove M1 crocodile.  CPU halt at WR IORQ after addr 0595

e.t.c. 

[Bryce]

If his car breaks down

Its NOT his car.

Even if some of those tests could easily result in him damaging something much more expensive on the car?

YES! because z80 is in socket. RAM not.

If I confirm that the ROM contains the correct procedure for using RAM and it does not work because of RAM and not 373, it will solder 256 RAM pins.

Think of it as gathering mmldrm experience and my fun in coming up with tests without an oscilloscope.

It's not your CPC either and your "fun" is at the expense and risk of someone else's property.

As for learning. mmlmrd has now learnt in detail how NOT to diagnose and repair a CPC.

At the moment, the status of the CPC is less clear than when he started. On page one, it was blatantly clear that the RAM had failed. Good advice would have been to tell him this. During any of these hair-brained experiments he may have touched a wrong pin by mistake and fried an output without knowing. He now has a CPC that may have additional failures and damaged parts and it will be much more complicated to confirm this.

He would have soldered all 256 pins and had a working CPC weeks ago. It would have been less work, less expense and he would have a working CPC now. Instead he still has a non-working CPC, now with possible additional failures. If that's your idea of good repair practise, you also have a lot to learn.

Here's some stuff he, you and others should learn:

1 - Most electronics will have a common weak point. If many other owners had the same issue and the cause was almost always the same, then follow that advice. Any additional poking around is risking damage to parts that are much more difficult to find and/or expensive to fix.

2 - Don't spend time or money testing a part that is cheaper than the equipment needed to test it.

3 - Learn what parts rarely fail and assume they work until you have ruled out the weakest parts.

4 - NEVER inject signals into a circuit at a point that has outputs connected to it. Signal injection is useful for repairing power supplies, audio equipment and test gear. It's not common or good practice to manipulate bits on a running computer.

5 - Electronic repair isn't just about voltages, currents and components. It's about recognising what most likely failed.

Bryce.

[Gryzor]

If his car breaks down

Its NOT his car.

Even if some of those tests could easily result in him damaging something much more expensive on the car?

Think of it as gathering mmldrm experience and my fun in coming up with tests without an oscilloscope.

Oh ok, last time I went to the mechanic we had lots of fun tearing the car down instead of using some foam to keep the fuel line in place, and I learned so much in the process!

[McArti0]

He now has a CPC that may have additional failures and damaged parts and it will be much more complicated to confirm this.

Bryce.

OMG "realy" ? No risk no fun. 

4 - NEVER inject signals into a circuit at a point that has outputs connected to it.. It's not common or good practice to manipulate bits on a running computer.

Like all extension RAM for 464. And examples of pin "blocking" in this topic.

5 - Electronic repair isn't just about voltages, currents and components. It's about recognising what most likely failed.
Bryce.

Ugly approach to the matter. Blind replacement without diagnosis.
I am very glad that this mainboard is not mechanically damaged by desoldering the frames and there are no traces that have been detached.

[Gryzor]

He now has a CPC that may have additional failures and damaged parts and it will be much more complicated to confirm this.

Bryce.

OMG "realy" ? No risk no fun.

4 - NEVER inject signals into a circuit at a point that has outputs connected to it.. It's not common or good practice to manipulate bits on a running computer.

Like all extension RAM for 464. And examples of pin "blocking" in this topic.

5 - Electronic repair isn't just about voltages, currents and components. It's about recognising what most likely failed.
Bryce.

Ugly approach to the matter. Blind replacement without diagnosis.
I am very glad that this mainboard is not mechanically damaged by desoldering the frames and there are no traces that have been detached.

You totally fail to take into consideration the risk/benefit ratio, as well as the cost (monetary and time) involved.

From personal experience, when it comes to fixing something, I always made it worse when I tried to bite more than I could chew. And I'm willing to bet I'm not the only one.

[McArti0]

You totally fail to take into consideration the risk/benefit ratio, as well as the cost (monetary and time) involved

Please, no demagogy.

From personal experience, when it comes to fixing something, I always made it worse when I tried to bite more than I could chew. And I'm willing to bet I'm not the only one.

And it's completely the opposite for me. It always ends with something revealing.

[Bryce]

You totally fail to take into consideration the risk/benefit ratio, as well as the cost (monetary and time) involved

Please, no demagogy.

From personal experience, when it comes to fixing something, I always made it worse when I tried to bite more than I could chew. And I'm willing to bet I'm not the only one.

And it's completely the opposite for me. It always ends with something revealing.

The bus contention used by external RAM expansions is controlled, not a random wire with no current limiting. And still it's very bad practise.

Demagogy would suggest that you are somehow elite?

Having "fun" at the expense of others and suggesting (without warnings) risky experiments to someone without the experience to understand the risks involved is not ok.

mmldrm didn't come asking for fun. He came looking for advice on the best way to repair the CPC and up to now you have brought him no closer to his goal.

Bryce.

[Gryzor]

You totally fail to take into consideration the risk/benefit ratio, as well as the cost (monetary and time) involved

Please, no demagogy.

Ok, I actually laughed on this one.

[McArti0]

mmldrm didn't come asking for fun. He came looking for advice on the best way to repair the CPC and up to now you have brought him no closer to his goal.

Unfortunately, he doesn't sit on cpc all the time, he doesn't have much time and he doesn't learn very quickly.

Setting color 0 without the risk of forcing the 4 pins states is even longer.

[McArti0]

You totally fail to take into consideration the risk/benefit ratio, as well as the cost (monetary and time) involved

Please, no demagogy.

Ok, I actually laughed on this one.

https://www.cpcwiki.eu/forum/hardware-related/cpc-6128-white-screen-black-border/msg237939/#msg237939

Are you laughing at this post too?

[Gryzor]

You totally fail to take into consideration the risk/benefit ratio, as well as the cost (monetary and time) involved

Please, no demagogy.

Ok, I actually laughed on this one.

https://www.cpcwiki.eu/forum/hardware-related/cpc-6128-white-screen-black-border/msg237939/#msg237939

Are you laughing at this post too?

No, only laughing at you and your inability to keep up with a discussion.

[McArti0]

No, only laughing at you and your inability to keep up with a discussion.

Ah ok. Greetings for your wife.

[Gryzor]

Yeah, insulting family of others - that's a first around here. Much as I'll miss the fun, he's banned for a week.

[Bryce]

mmldrm didn't come asking for fun. He came looking for advice on the best way to repair the CPC and up to now you have brought him no closer to his goal.

Unfortunately, he doesn't sit on cpc all the time, he doesn't have much time and he doesn't learn very quickly.

Well at least you took into consideration that he doesn't have much time and helpfully suggested the quickest way to get the CPC working again... 

Also nice that you consider him a slow learner. I'm sure he'll be pleased to hear what you think of him.

Bryce.

[eto]

Are you laughing at this post too?

How is my post related to that?

It was just to demonstrate that a short on a data-pin will have exactly that effect. And that was my own hardware and my own risk. I would not recommend anyone to do that.

[SerErris]

To be honest, the learning out of this thread for someone new to digital electronics and CPC in general is pretty much zero.

Probing things without knowledge on why and what and the background of what you actually do is pretty much confusing and not helping any bit to better understand what is going on.

Setting a Color in the GA with wires is the strangest thing I heard and I am not sure if I would be able to execute this, even knowing how all those thing work. It would be just out of my manipulation capabilities to get this timed correctly and working.

I would still say:
If anyone has no knowhow on any of those topics, you should not attempt to fix it. If you have good soldering skills (which we also do not know), you can attempt to replace the RAM.

However if you do not have ever desoldered a multi pin chip and do not have proper tools to do that, I would still recommend to give it to someone, that has the skills required.

Desoldering the RAM chips has the inherent danger of damaging the mainboard and creating even more damage to the CPC, that is actually beneath the RAM issue probably fine. (e.g. no critical parts damaged (like GateArray)).

[Bryce]

To be honest, the learning out of this thread for someone new to digital electronics and CPC in general is pretty much zero.

Probing things without knowledge on why and what and the background of what you actually do is pretty much confusing and not helping any bit to better understand what is going on.

Setting a Color in the GA with wires is the strangest thing I heard and I am not sure if I would be able to execute this, even knowing how all those thing work. It would be just out of my manipulation capabilities to get this timed correctly and working.

I would still say:
If anyone has no knowhow on any of those topics, you should not attempt to fix it. If you have good soldering skills (which we also do not know), you can attempt to replace the RAM.

However if you do not have ever desoldered a multi pin chip and do not have proper tools to do that, I would still recommend to give it to someone, that has the skills required.

Desoldering the RAM chips has the inherent danger of damaging the mainboard and creating even more damage to the CPC, that is actually beneath the RAM issue probably fine. (e.g. no critical parts damaged (like GateArray)).

If you don't need to save the chip you are removing, then this is probably the safest way to do it: https://www.cpcwiki.eu/index.php/IC_Repair
You don't need a desoldering gun to do it either.

Bryce.

[McArti0]

Setting a Color in the GA with wires is the strangest thing I heard and I am not sure if I would be able to execute this, even knowing how all those thing work. It would be just out of my manipulation capabilities to get this timed correctly and working.

I've never heard of it either. But I do it with one hand probably because I wanted to do it.


Turn on captions!!!

;
;                              /--|>|-------- XXX------ M1
; READY/WAIT ---<
;                              \--|>|-------- XXX------ IORQ
;                                    |          crocodile
;                                diode
;                                    |
; A15 ------->    -----|>|-------GND
;
; always you can check Address and data when Z80 stopped.
;
; clip IORQ crocodile, clip M1 crocodile and reset (or swith on CPC).
;
ORG #0000     
LD BC,#7F89 ;recatch M1 (1)
OUT (C),C                        ;recatch IORQ (2)
.0005
JP #0591        ;recatch M1 (3)

ORG #0591
DI
LD BC,#F782
.0595
OUT (C),C                      ;recatch IORQ (4)
.0597
LD BC,#F400    ;recatch M1 (5)
.059A
OUT (C),C ;store by short A15 ground and recatch IORQ (6)
.059C
LD BC,#F600 ;recatch M1 (7)
.059F
OUT (C),C ;second chance to store by short A15 ground and recatch IORQ ( 8 )
.05A1
LD BC,#EF7F ;recatch M1 (9)
.05A4
OUT (C),C ;ground one data line 0 to 4, and store by short A15 ground

; pen 0 set to 5 different colors

; set ROMDIS to +5V and remove crocodiles. If You see vertical bands RAMs work.

[mmldrm]

I'm back! You can decide if that is good or bad news. 

Before trying the latest test from @McArti0, I finally took the CPC to a local computer repairs guy. (He has repaired C64s before but not CPCs. Not many options around here for repair shops.)

Yesterday he replaced the lower 64Kb of RAM and... still the same white screen and black border.

I will collect the CPC today but unfortunately no more updates for a while, because I'm away for the next couple of weeks.

So, I guess I need to decide what to do next. Dandanator? More experiments? Or the recycle bin.

(Thanks again for all time spent helping me here...)