Crystal oscillator issues on Amstrad 6128 Plus

[llopis]

I have a 6128 Plus that just shows a black screen on startup. No borders or anything. I have both a working 6128 Plus and this non-working one, so it's really handy to compare.


Voltages are fine everywhere, so I moved to check clock signals next and... they're really bad (there are some ripples but they aren't the right amplitude or shape).


So next I look at the signal created by the crystal oscillator and it's slightly wrong.


This is the related schematic (I guess I can't insert attachment here, so see below).


Both PCBs are fitted with the individual crystal oscillator instead of the optional circuit in the dashed box. In either case, the output from the oscillator circuit are two pins going into the GA (1 and 2). They're supposed to be 40MHz signals and I assume they're negatives of each other.


When I check the good one, I get the following signal pretty much on both pins (see below).


But when I check the bad 6128, pin 2 of the GA is the same, but pin 1 has a much weaker signal (correct frequency, but at around 0.3V and less variation too).


Any ideas what might be going on? I naively replaced the crystal, but I got the exact same result. Part of my problem is that I don't understand exactly what's going on in pin 1. It looks like this crystal generates a signal only in pin 2? Where's the one in pin 1 coming from, inside the GA? If so, does that mean that I'm out of luck?


Any suggestions appreciated since I'm out of options over here.

[gerald]

Any ideas what might be going on? I naively replaced the crystal, but I got the exact same result. Part of my problem is that I don't understand exactly what's going on in pin 1. It looks like this crystal generates a signal only in pin 2? Where's the one in pin 1 coming from, inside the GA? If so, does that mean that I'm out of luck?


Any suggestions appreciated since I'm out of options over here.

A bit of backgroud first.
The 6128 Plus PCB has two option
1. use an external oscillator
2. use the ASIC internal oscillator with an external quartz.
To my knowledge, only option 1 has been used.Pin 2 (OSCI) is the clock input of the ASIC while pin 1 (OSCO) is the ouput of the ASIC oscillator circuit.
The external oscillator only feed OSCI, and OSCO is of no use.

Seeing a weaker signal on OSCO on the failling Plus may hint that the ASIC is damaged and internal voltage is not as expected (internal short or broken supply bonding wires)
Can you see anything on the /RAS (pin91) going to the RAM or CLK1 (pin16) going to the PSG ?

[llopis]

A bit of backgroud first.
The 6128 Plus PCB has two option
1. use an external oscillator
2. use the ASIC internal oscillator with an external quartz.
To my knowledge, only option 1 has been used.Pin 2 (OSCI) is the clock input of the ASIC while pin 1 (OSCO) is the ouput of the ASIC oscillator circuit.
The external oscillator only feed OSCI, and OSCO is of no use.

Seeing a weaker signal on OSCO on the failling Plus may hint that the ASIC is damaged and internal voltage is not as expected (internal short or broken supply bonding wires)
Can you see anything on the /RAS (pin91) going to the RAM or CLK1 (pin16) going to the PSG ?

That makes a lot of sense, Gerald.


I looked at those signals, and I'm seeing more or less normal signals but... maxing out at 2V instead of 4.8V like the working Amstrad.
Here's CLK1 for example, and the good version for comparison.


I checked and the GA gets full 5V (well, 4.8V) everywhere that it's supposed to. It doesn't get warm with usage and current draw is less than a working Amstrad (0.3A instead of 0.5A), but I suppose that doesn't remove the possibility of a short inside the GA.


In any case, it sounds like my initial feeling that this one is not salvageable is correct, right? This is why I hate those ASIC chips. You can't even replace them! :-(

[GUNHED]

Could the system be overclocked? For example 5 MHz instead of 4 MHz (Z80 frequency, not crystal). Did somebody ever try?

I had to idea to do it, but the crystal on the Plus PCB is kind of "hard" to replace, it's way more easy on CPC6128.

[Audronic]

@llopis


""I checked and the GA gets full 5V (well, 4.8V) everywhere that it's supposed to""

Do you get 4.8 Volts on the good working Plus ?.

Ray

[llopis]

Do you get 4.8 Volts on the good working Plus ?.

Yes. The working Plus has is more like 4.6V on all the +5V inputs to the GA. I imagine that's because it's actually working and causes a bit more of a voltage drop.


There was nothing tricky about replacing the crystal. It was just 4 solder points and out it goes.

[robcfg]

Have you tried to check the RAM chips?


The clock signals do look good, but they have lower amplitude on the broken one.


I don't know if there's a transistor acting as amplifier that has gone bad.


I guess that if you insert a cartridge, there's no sound from either a game or pressing the DEL key on Basic, right?