Spectrum Computing

C80 - should be 22uF, was 1uF - this will affect the power output of the DC/DC converter
C74 - should be 4.7uF, was 22uF - this will affect the response time of the regulation of the +12V line
C27 - should be 1uF, was 4.7uF - that will just give a slightly longer reset pulse to the Z80.

Did the person who you got this from wear a stetson?

Mark

1024MAK wrote: ↑Sun May 03, 2020 12:37 am C80 - should be 22uF, was 1uF - this will affect the power output of the DC/DC converter
C74 - should be 4.7uF, was 22uF - this will affect the response time of the regulation of the +12V line
C27 - should be 1uF, was 4.7uF - that will just give a slightly longer reset pulse to the Z80.

Did the person who you got this from wear a stetson?

Mark

Haha, I couldn't tell you unfortunately but it could be the case.
I was hoping this would have resolved the problem but nope

Border colour is changeable, it's very random, the coloured blocks flash at ~ 2Hz
Same result with lower ram removed.
Same result with upper ram disabled.
ULA is fine
Z80 swap didn't help
I think I might end up changing the ROM or the multiplexer chip I'm thinking.
No sign of RAM test on reset.

Hi!
Sorry, I like to share, I hope it's interesting.

After reading some of the Z80 Technical Manual (Great little book happy to have found it) I realised a silly mistake. My replacement Z80 was actually a Z84C20 with is a PIO chip.
I put an actual Z80 in and got a white border with some more familiar failure modes. What's very interesting is that it sometimes (rarely) starts with blue and yellow lines (and static noise) around the border as if it's loading something (it doesn't do this every time).
Anyway I removed the lower RAM and disabled the upper RAM and got something that looks correct (although with lots of noise).
I found that with the RAM installed I could directly influence the picture by touching IC12, the result was the same when I swapped chips around so I suspected the socket.
Continuity testing shows no continuity between A5 on IC12 and the adjacent ICs, so I'll be taking the socket off and inspecting the joint.

It's so exciting to find a fault like this, looking forward to seeing if it fixes some of the issues I'm getting

Cheers

It looks like a bad joint on the topside of the board as there is continuity from the socket to the underside

HappyLittleDiodes wrote: ↑Sun May 03, 2020 9:02 pm My replacement Z80 was actually a Z84C20 with is a PIO chip.

What, and it didn't work?

dfzx wrote: ↑Mon May 04, 2020 9:18 am

HappyLittleDiodes wrote: ↑Sun May 03, 2020 9:02 pm My replacement Z80 was actually a Z84C20 with is a PIO chip.

What, and it didn't work?

I'm impressed that anything happened.
I got the screen shown in the post above and the blocks were changing.
The chip did get quite hot though...

It got me thinking, could you make a simulation so accurate that you could predict this behaviour if you put this chip into a working spectrum?

HappyLittleDiodes wrote: ↑Mon May 04, 2020 11:39 am

dfzx wrote: ↑Mon May 04, 2020 9:18 am What, and it didn't work?

I'm impressed that anything happened.
I got the screen shown in the post above and the blocks were changing.
The chip did get quite hot though...

It got me thinking, could you make a simulation so accurate that you could predict this behaviour if you put this chip into a working spectrum?

Try powering up a ZX Spectrum with the Z80 missing...
The Z80 in a ZX Spectrum is a slave

Mark

I found my broken track, hooray!

But I got impatient and made it worse.
Time to put it away until the copper tape arrives...

HappyLittleDiodes wrote: ↑Mon May 04, 2020 8:04 pm Time to put it away until the copper tape arrives...

Copper tape? You mean the desoldering stuff?

dfzx wrote: ↑Tue May 05, 2020 10:54 am

HappyLittleDiodes wrote: ↑Mon May 04, 2020 8:04 pm Time to put it away until the copper tape arrives...

Copper tape? You mean the desoldering stuff?

No it's a sticky copper tape for electronic repairs. You tidy up the damage, stick this over and solder it in.
I tried to make the link with a small wire but buggered it up

HappyLittleDiodes wrote: ↑Tue May 05, 2020 11:00 am No it's a sticky copper tape for electronic repairs. You tidy up the damage, stick this over and solder it in.
I tried to make the link with a small wire but buggered it up

Something else I never knew existed.

I've never used it before but it seems like a good solution, fingers crossed, if I make a mess again it's going to get harder to fix

Don't feel bad, I managed to lift about 4 traces too, removing all the chips on my board. Most of them from chips that were already replaced by a PO and I suspect damaged them.

Its a tricky board to remove chips from. Patience is key, as in flux and re-flowing - fresh solder.

For you feel better pleasure. Top side just had a few traces, bottom was a disaster ground. I did the full nuke on removing all the upper ram and logic chips and for some reason that section gave me all the trouble (ram). Rest was 98% fine. Good news is I have a fully socketed board once I get that last two sockets in.

Not my best work or my most proud, I have never had so much trouble desoldering components than on this board being so fragile.

So many pads came off on the 32k section from previous damage. Not needed as they were not connected to anything, but still. Next time I will try the thin wire repairs hidden by the socket, that I have seen folks do. You need a steady hand...

https://www.youtube.com/watch?v=KRSQOOY67DU

speccyplus wrote: ↑Wed May 06, 2020 2:26 pm

For you feel better pleasure. Top side just had a few traces, bottom was a disaster ground. I did the full nuke on removing all the upper ram and logic chips and for some reason that section gave me all the trouble (ram). Rest was 98% fine. Good news is I have a fully socketed board once I get that last two sockets in.

Not my best work or my most proud, I have never had so much trouble desoldering components than on this board being so fragile.

So many pads came off on the 32k section from previous damage. Not needed as they were not connected to anything, but still. Next time I will try the thin wire repairs hidden by the socket, that I have seen folks do. You need a steady hand...

https://www.youtube.com/watch?v=KRSQOOY67DU

Did you use a heat gun to help with removing the chips?
I find it makes the actual removal a lot smoother, but you still have to take the time to get as much solder out as possible!

I'll let you know how I get on using the tape

The normal answer to removing chips (this assumes the chip is being removed because it’s faulty) is to cut all of it’s pins as close to the plastic package as possible. Remove the plastic package, then each pin can be removed with long nose pliers while the solder is kept molten with the hot iron.

Of course this is no good if you want to reuse the chip... then the best method is hot air, but you have to get things just right...

Mark

Thanks Mark.
Yes I'm a bit concerned about the heat gun. No problems so far but I find it a bit unrefined when the board gets a bit wobbly with the heat

The copper tape repair went beautifully, I captured it with the microscope and the spectrum now boots correctly

HappyLittleDiodes wrote: ↑Wed May 06, 2020 2:53 pm

speccyplus wrote: ↑Wed May 06, 2020 2:26 pm

For you feel better pleasure. Top side just had a few traces, bottom was a disaster ground. I did the full nuke on removing all the upper ram and logic chips and for some reason that section gave me all the trouble (ram). Rest was 98% fine. Good news is I have a fully socketed board once I get that last two sockets in.

Not my best work or my most proud, I have never had so much trouble desoldering components than on this board being so fragile.

So many pads came off on the 32k section from previous damage. Not needed as they were not connected to anything, but still. Next time I will try the thin wire repairs hidden by the socket, that I have seen folks do. You need a steady hand...

https://www.youtube.com/watch?v=KRSQOOY67DU

Did you use a heat gun to help with removing the chips?
I find it makes the actual removal a lot smoother, but you still have to take the time to get as much solder out as possible!

I'll let you know how I get on using the tape

I have a power vacuum solder sucker, and it was horrible, it usually works quite well. I had better luck with a manual sucker, and cleaning the holes out that way.

For whatever reason extracting solder from the upper ram I ran into nothing coming out well. Other chips usually sucked them clean and I could gently push the sides of the chips to free them from within the holes.

No heat gun.