Need some help fixing CPC6128 not booting

[Bryce]

Ah ok, I thought you'd removed the RAM ICs. In that case I'd try swapping the logic first.

Bryce.

[ikonsgr]

Well,i remove the 1st bank indeed, but i have installed dip sockets and put them back again! 
Tomorrow i will go buy a nice 74HC373 (better cmos for lower consumption, after all even in the schematics of the service manual they propose 74LS or 74HC anyway  ),i already have some 20pin dip sockets and a 74HC244 and.... let the desoldering begin (once again...)! 

[RobertM]

I would be careful substituting different spec chips at this time as you may introduce more variables into the equation while you still in diagnosis stage.

While chip manufacture has come a long way since the CPC's were made, the manufacturers still target the same specification as the old originals.

No only does CMOS have different Vil and Vih (Without looking it up) the fanout (how many chips it can drive) is different from my memory.

If your socketing then you can always change to CMOS later, once the fault is found.

CMOS is most probably fine but if it isn't then your in for a long journey to find out.

Also I did notice a jumper on the main board that had CMOS written beside it but I havent looked at the circuit diagram to see what it is for.

[Bryce]

If you have the money then -
DSO (ebay $66)

By pure coincidence I had the chance to try one of these out at the weekend. I would definitely NOT recommend getting one of these, not even for just checking if a signal is present. It's a toy, not a meter. It's meant to be a 1Mhz device, but the results on signals above 100Khz were completely wrong and pretty random too, so it would be useless for even retro electronics. Connected to a 1Mhz square wave displayed a straight line suggesting that no signal was present. In reality the clock signal was there and running fine. It only displayed a halfway realistic result when I turned the clock down to about 70Khz!!! And even then the compensation was completely off.
There is a more expensive 12Mhz (some sites claim this to be 36 or even 72Mhz) version called the DSO203 which might be better suited, but at €135 it's too expensive. For €110 you can already get a "real" 70Mhz USB oscilloscope which will serve you much better.

The only thing worse than having no test equipment, is having test equipment that you can't trust.

Bryce.

[ikonsgr]

[UPDATE]
Finally the ram chips arrived and so i installed all the first bank with brand new chips (the packet was 10 items), remove the second bank completely and also change the 74ls244 and 74ls373 with new ones. So EVERYTHING in the path of the Dout signal of all ram chips is replaced!
Unfortunately the result is the same: completely nothing on the screen. Although now the ram chips seem to get warm like in the working cpc (the old ones barely got warm)
So now my question is: If you remove the 2nd bank of a working cpc 6128, can it still works, or at least give some video signal (a window or something) other than nothing at all? Because it would be really frustrate to order another pack of ram chips to fill in the 2nd bank, and still get the same problem. And in that case 99% chances mean that the problem is a faulty rom chip.... 

[Bryce]

Are you sure the Gate Array works? Sounds like it's gone.

Bryce.

[ikonsgr]

Yeap, because i have a second cpc which works perfect, and i'm using the 40010 from there! 

[gerald]

So now my question is: If you remove the 2nd bank of a working cpc 6128, can it still works, or at least give some video signal (a window or something) other than nothing at all? Because it would be really frustrate to order another pack of ram chips to fill in the 2nd bank, and still get the same problem. And in that case 99% chances mean that the problem is a faulty rom chip.... 

A 6128 will work without the second bank. SW trying to use it will obviously fail, but basic and FW does not use it at all.

The screen initialisation is done in the very beginning of the ROM code, even before using any RAM. This should give you a proper screen with border.
If you are sure the CRTC/Z80/GA are OK, the ROM may be your problem. However, it can also be an address/data/ctrl bus problem or ROMDIS forced high (no access to ROM), or WAITn forced low (Z80 waiting forever) .....

[ikonsgr]

And what faulty chips on motherboard might cause any of these you mention (address/data/ctrl bus problem or ROMDIS forced high (no access to ROM), or WAITn forced low (Z80 waiting forever)?
Up until now i have replaced almost any major chip (Z80,6845,40010,HAL16L8,74LS244,74LS373,1st bank ram) on motherboard, so is there a chance that some of the remaining chips (for example a faulty 74lsxxx) which i haven't replaced, could cause a so serious problem?

[Bryce]

Have you checked the resistance of each Address and Data line to Vpp? Maybe a pull-up resistor has failed?

Bryce.

[gerald]

Just trying to list what can conflict with rom fetch :

ROMDIS is pulled down by a resistor and connected to pin 22 (CE\) of the ROM. It is driven high by external ROM management or internally (on 6128) by the AMSDOS rom management. Check that the voltage on pin 22 is 0v. If not, the amsdos ROM management may be the problem.

WAITn It is driven by the GA through a resistor with a 3/4 ratio (3 cycle low, 1 cycle high). You should measure about 1V with a multimeter on this pin. If you find 0V, the resistor may be broken.

Regarding the data bus, faultly PPI (8255) and the printer buffer can be (unlikely) candidates. AMSDOS ROM also.

However, checking that the ROM is OK is a priority.

[Bryce]

Just trying to list what can conflict with rom fetch :

ROMDIS is pulled down by a resistor and connected to pin 22 (CE\) of the ROM. It is driven high by external ROM management or internally (on 6128) by the AMSDOS rom management. Check that the voltage on pin 22 is 0v. If not, the amsdos ROM management may be the problem.

WAITn It is driven by the GA through a resistor with a 3/4 ratio (3 cycle low, 1 cycle high). You should measure about 1V with a multimeter on this pin. If you find 0V, the resistor may be broken.

Regarding the data bus, faultly PPI (8255) and the printer buffer can be (unlikely) candidates. AMSDOS ROM also.

However, checking that the ROM is OK is a priority.

That's exactly the direction that I was thinking towards. Some signal is being held high / low somewhere, but that means it could be almost anything on the PCB

Bryce.

[gerald]

That's exactly the direction that I was thinking towards. Some signal is being held high / low somewhere, but that means it could be almost anything on the PCB

Bryce.

Yes, and that's why the ROM need to be checked first after the romdis/waitn.
If the ROM is OK, then oscilloscope/LA becomes a mandatory hunting tools !

[ikonsgr]

Thanks for the info guys. I'll check it tomorrow and let you know!

[ikonsgr]

Ok, 'ive checked the 2 signals you said, romdis on pin 22 of the amsdos/basic rom and "wait" on pin 24 of the cpu.
On faulty amstrad i get ~1,5mV and 1.2Volts respectively, while on the working one i get ~0.5mV and 1,15Volts.
So you think ROM is ok?
I also try the working amstrad without the pal16l8 and it gives a grey box. The faulty one doesn't give anything on the screen, but it has only the 1st bank installed (2nd bank is completely removed)
Finally i checked almost all resistors on the board and they seem to match the values with resistors on the working cpc.
You think it's worth buying another pack of ram chips for the 2nd bank, or most probable the problem is elsewhere?


P.s i just noticed another strange thing: when i get my hand close to the ic 203 (a 74hc240 on the top right corner of the board above the 765 controller i think it's part of the floppy disk circuit) i get sound distortion from loudspeaker! Of course this is not happening on the working cpc. Does this mean this chip is faulty, and if it is, can it cause such a serious problem?

[gerald]

Ok, 'ive checked the 2 signals you said, romdis on pin 22 of the amsdos/basic rom and "wait" on pin 24 of the cpu.
On faulty amstrad i get ~1,5mV and 1.2Volts respectively, while on the working one i get ~0.5mV and 1,15Volts.
So you think ROM is ok?

These two test result are expected. This mean that
1. The ROM is not disabled
2. The Z80 is not halted

For the ROM itself, you cannot say that it is working or not. But, if there are no problem on the memory bus (address/data/control), the ROM is likely failing.

I also try the working amstrad without the pal16l8 and it gives a grey box. The faulty one doesn't give anything on the screen, but it has only the 1st bank installed (2nd bank is completely removed)

Removing the PAL will have the same effect as a faultly RAM. Since you do not have anything on screen, you should forget about the RAM for now 

P.s i just noticed another strange thing: when i get my hand close to the ic 203 (a 74hc240 on the top right corner of the board above the 765 controller i think it's part of the floppy disk circuit) i get sound distortion from loudspeaker! Of course this is not happening on the working cpc. Does this mean this chip is faulty, and if it is, can it cause such a serious problem?

This chip is buffering signals from the floppy drive and is also part of the FDC address decoding. If the decoding goes wrong, the FDC may drive the bus.
Can you check that the CS pin stays high (IC201 pin 4), as well as the AMSDOS ROM is not selected (IC204 pin 22). Both pins voltage should be close to 5V.

[ikonsgr]

Yeap, both pins got ~4.5volts just like on the working amstrad too.
Btw, these "parados" roms you can find everywhere on ebay,is a substitute for the ic 204 (the fdd rom)?
And if i do have a rom problem, which of the 2 roms can cause this "nothing on screen" symptom?

[TFM]

If you got no sign on message then either BASIC or the Firmware has a problem... Or the GA.

[gerald]

Yeap, both pins got ~4.5volts just like on the working amstrad too.
Btw, these "parados" roms you can find everywhere on ebay,is a substitute for the ic 204 (the fdd rom)?

Parados is a AMSDOS replacement ROM, it can replace IC204.

And if i do have a rom problem, which of the 2 roms can cause this "nothing on screen" symptom?

The FW/BASIC ROM is IC103, and is the one responsible for setting up the hardware at power up.
Ideally, you should remove it from the board and verify it with an eprom programmer (comparing it to a valid rom lower rom and basic rom)

[ikonsgr]

That would be indeed ideally, IF i had an eprom programmer! 
For now,i'm waiting for the logic tester and i was thinking, if i will make a "logic" map of every pin (either low,high or pulsed) of each rom on the working cpc, and then make the same thing with the faulty one and compare them, will i be able to know for "sure" from the result,if rom chips are ok (in case of complete concurrence between the roms of the working and the faulty cpc), or NOT ok (in opposite case)?

Finally,if the basic/fw rom chip is indeed faulty, is there any way to find a replacement?
And if not, are  this the needed eeprom to be programmed (i think there is a list of all roms somewhere in cpc wiki...  )?

[gerald]

For now,i'm waiting for the logic tester and i was thinking, if i will make a "logic" map of every pin (either low,high or pulsed) of each rom on the working cpc, and then make the same thing with the faulty one and compare them, will i be able to know for "sure" from the result,if rom chips are ok (in case of complete concurrence between the roms of the working and the faulty cpc), or NOT ok (in opposite case)?

Unless you get some static data bit, you cannot conclude as you don't know what the Z80 is doing.
You could check the rom by testing it with a working 6128. But you risk to damage that working CPC as you need to unsolder its ROM (and add a socket).

Finally,if the basic/fw rom chip is indeed faulty, is there any way to find a replacement?

You need a 27C256 (32kx8) eeprom. The lower 16k need to be programmed with the 6128 FW, the upper with the basic 1.1

And if not, are  this the needed eeprom to be programmed (i think there is a list of all roms somewhere in cpc wiki...  )?

No, this is a microcontroller.

[ikonsgr]

Unless you get some static data bit, you cannot conclude as you don't know what the Z80 is doing.
You could check the rom by testing it with a working 6128. But you risk to damage that working CPC as you need to unsolder its ROM (and add a socket).

And that's exactly why i'm NOT going to do this, because the other cpc is the only working cpc i have! 

You need a 27C256 (32kx8) eeprom. The lower 16k need to be programmed with the 6128 FW, the upper with the basic 1.1

Then, you mean this?

[gerald]

And that's exactly why i'm NOT going to do this, because the other cpc is the only working cpc i have! 
Then, you mean this?

Yes !
But you need to get them programmed.

[ikonsgr]

I found another for only 2$/piece, i suppose i can use these instead?
And i also found these which they don't seem to have the window for UV erase.
I have a friend of mine which has a couple of programmers but i don't know if it can program the specific one.
Is this a rather "rare" type of eeprom which needs expensive programmer or not?

[gerald]

The 2$ one are OK, the second link point to SMD device that you cannot use on the CPC without an adaptor.
27C256 are really standard eprom. Any eprom programmer should be able to handle them.