Mega ST4 - no video output at all

[asapreta]

Well, seeing as the machine isn't able to get anywhere suggests the lowest bank of RAM, so you only need to look at the lowest 16 chips.

Myself I'd use an oscilloscope to look at the data lines so see if any look "strange". Even without the CPU the Shifter will be trying to do memory access.

Thanks for the tip of the 16 chip of the lower bank.
And yes, I will use the scope to see if any of them gives a hint, with different readings.

The weekend is close, hope I can find some spare time to focus on this! :)

[rubber_jonnie]

I'd agree with @stephen_usher and I'd also use a scope to look for stuck address lines.

Of course the RAM issue might be an MMU socket issue, so I'd be tempted to just tweak the socket pins a little to be sure of good contact.

It doesn't need much to tweak the pins so be super careful.

A diag cart is a cheap and useful if you have other ST's to work on since you can pick up useful diag info from the serial port with a PC connected and terminal software.

[Badwolf]

I'd agree with @stephen_usher and I'd also use a scope to look for stuck address lines.

Of course the RAM issue might be an MMU socket issue, so I'd be tempted to just tweak the socket pins a little to be sure of good contact.

It doesn't need much to tweak the pins so be super careful.

A diag cart is a cheap and useful if you have other ST's to work on since you can pick up useful diag info from the serial port with a PC connected and terminal software.

On my STFM, before I sat down and systematically worked on the PLCC socket pins as discussed above, tapping the MMU whilst periodically hitting reboot would show that sometimes it worked and sometimes it didn't. Bit of pressure this way or that might work too.

All you have to do is get it working once with a bit of a tap or pressure here and there to come pretty close to your smoking gun.

Diag cart is must have anyway, though.

BW

[stephen_usher]

Another non-destructive test, which only sometimes works if it's a RAM chip which has failed in a way that it's not stuck low, is to piggy-back a known good RAM chip on top of one of the ones on the motherboard. If you've got multiple 41256 chips then piggybacking multiple chips may (or may not) work to get different behaviour and possibly a working system as a test.

Non-destructive tests are always the first things to try.

Having said that, a black background suggests that the Shifter isn't able to read any RAM, which would point towards the MMU/GLUE and connections to/from that, possibly CAS/RAS lines etc. So, checking those signals both at the GLUE/MMU sockets and on the RAM with an oscilloscope would be a good thing to test.

[asapreta]

I'd agree with @stephen_usher and I'd also use a scope to look for stuck address lines.

Of course the RAM issue might be an MMU socket issue, so I'd be tempted to just tweak the socket pins a little to be sure of good contact.

It doesn't need much to tweak the pins so be super careful.

A diag cart is a cheap and useful if you have other ST's to work on since you can pick up useful diag info from the serial port with a PC connected and terminal software.

Thanks for all support and tips.
I will put the diagcart in my wishlist.

As you, I am a proud owner of 50-60 retro computers from different companies, like ZX Spectrum, Commodores, Amstrads, Amigas, BBC, Atari 8 and 16 bit, so the wishlist is kinda big...

But I will use the scope to get these stuck address lines, hope I can find them!

[rubber_jonnie]

I'd agree with @stephen_usher and I'd also use a scope to look for stuck address lines.

Of course the RAM issue might be an MMU socket issue, so I'd be tempted to just tweak the socket pins a little to be sure of good contact.

It doesn't need much to tweak the pins so be super careful.

A diag cart is a cheap and useful if you have other ST's to work on since you can pick up useful diag info from the serial port with a PC connected and terminal software.

Thanks for all support and tips.
I will put the diagcart in my wishlist.

As you, I am a proud owner of 50-60 retro computers from different companies, like ZX Spectrum, Commodores, Amstrads, Amigas, BBC, Atari 8 and 16 bit, so the wishlist is kinda big...

But I will use the scope to get these stuck address lines, hope I can find them!

Yes, and my wishlist just isn't getting any smaller, much to the disappointment of my wife!

[asapreta]

I'd agree with @stephen_usher and I'd also use a scope to look for stuck address lines.

Of course the RAM issue might be an MMU socket issue, so I'd be tempted to just tweak the socket pins a little to be sure of good contact.

It doesn't need much to tweak the pins so be super careful.

A diag cart is a cheap and useful if you have other ST's to work on since you can pick up useful diag info from the serial port with a PC connected and terminal software.

On my STFM, before I sat down and systematically worked on the PLCC socket pins as discussed above, tapping the MMU whilst periodically hitting reboot would show that sometimes it worked and sometimes it didn't. Bit of pressure this way or that might work too.

All you have to do is get it working once with a bit of a tap or pressure here and there to come pretty close to your smoking gun.

Diag cart is must have anyway, though.

BW

Hi Badwolf, thanks for you time answering my post!
I have checked the MMU's PLCC socket but will do it again.
Yesterday I was checking traces but I wasn't using the best schematics.
I notice the service manual has better ones.
I want to be 100% sure every pin on its socket has connection before start digging the bad RAM.

The only problem I have to purchase these "must have utilities" hardware is living in Brazil.
For each £ I need seven of my money (R$ - Brazilian Reais).

Annotation 2024-10-18 103539.png

And the taxes doubles the price.
I do buy stuff, but need some planning ;)

[asapreta]

Another non-destructive test, which only sometimes works if it's a RAM chip which has failed in a way that it's not stuck low, is to piggy-back a known good RAM chip on top of one of the ones on the motherboard. If you've got multiple 41256 chips then piggybacking multiple chips may (or may not) work to get different behaviour and possibly a working system as a test.

Non-destructive tests are always the first things to try.

Having said that, a black background suggests that the Shifter isn't able to read any RAM, which would point towards the MMU/GLUE and connections to/from that, possibly CAS/RAS lines etc. So, checking those signals both at the GLUE/MMU sockets and on the RAM with an oscilloscope would be a good thing to test.

Hi stephen_usher, I was to ask about this, are the 41256 compatible with the memory on the Mega ST4?
Mine has the TC511000AJ10, which is 1Mbit x 1, right?
I do have some 41256 here, and will buy more, but as you said, piggybacking can help. I saved my A500 doing this.

About the CAS lines, pin16 I checked them yesterday and got nothing.
The RAS lines, pin 3, has activity.

With no activity in CAS I saw it was connected to the 74LS244 (U59).
Probably this IC is faling as the traces to CAS lines from it to MMU are ok and MMU IC is working, tested on my 520 ST.

Does this sound right?

[asapreta]

Thanks for all support and tips.
I will put the diagcart in my wishlist.

As you, I am a proud owner of 50-60 retro computers from different companies, like ZX Spectrum, Commodores, Amstrads, Amigas, BBC, Atari 8 and 16 bit, so the wishlist is kinda big...

But I will use the scope to get these stuck address lines, hope I can find them!

Yes, and my wishlist just isn't getting any smaller, much to the disappointment of my wife!

Luck you still have a wife!

[rubber_jonnie]

Yes, and my wishlist just isn't getting any smaller, much to the disappointment of my wife!

Luck you still have a wife!

And luckily she accepts my hobbies...