ISSUE4A help needed

[MaxS]

Hello all
hope you all are doing well
I'm trying to fix a 48k issue 4A that is driving me crazy.
She's a yellow motherboard with blue silkscreen markings, much "loved" by previous owner..
Unfortunately she's the only issue 4 I have (so no cross checks are possible).


when I take it I've found that

- all the ICs were socketed, but I've found some traces and vias lifted (fixed by me)
- most of TRs have been replaced by equivalent ones (TR4 and TR5 blown, and replaced by me with a BC489A and BC251C respectively)
- ROM blown, replaced by me with a new one, same straps.
- some 4116 and 4532 blown (now replaced)
- all electrolytic CAPS are still the original ones.

So after some work and ALL the ICs tested OK in another working speccy she boots.. but..

- all voltages ok, +5, -5, +12v
- Only 2 or 3 times she recognized the 48k, and it failed to load a 48k game.
- usually boots as a 16k but the top of the memory is always different.
- removed all related upper ram ICs (IC24 left in place.. cause it's an issue 4 ) she boots 16k ok and loads ok.
- She works ok till the power off. tested for half an our playing Horace but she always needs to cool down before reboot.
- when warm tryin to but again I got 4 black tick vertical black bars, border color changes at every boot but never white.
- once cool she reboots ok.


I've rechecked all IC's and they're still working ok in another motherboard.
I suspect some of the traces below the MUX, CPU or ROM as they have been socketed by someone else,
or maybe the caps needs to be changed.
Do you have any suggestion?

Where can I find reliable schematics for issue 4A?
All schematics I've found have some errors (IC3 pin12 to ROM pin27/ CPU pin2 should be IC3p2 to ROMp2 as per issue 3)
Reading the forum I understood that updated schematics should exist.. but maybe I'm wrong..

thanks in advance for any suggestion..
max

[1024MAK]

If there are updated schematics, it’s news to me.

On this forum, there are these topics about ZX Spectrum schematic errors:

See viewtopic.php?f=22&t=1528

And viewtopic.php?f=22&t=1670

In most machines, I only normally recommend replacement of the electrolytic capacitors when symptoms appear. But in ZX Spectrum 16k, 48k and plus machines I recommend renewing them. Especially those associated with the DC/DC converter circuitry.

If the border is not always white at power up or reset, it’s worthwhile renewing C27 (1uF) just in case, and to eliminate it as a possible cause (it provides the reset pulse to the Z80). Also check R31.

Mark

[MaxS]

thanks Mark
R31 out of circuit read 220R instead of 220K!
C27 replaced too.
Main ICs still out of the board till I'll finish to check all address and data lines traces.
In the meantime does it make sense to check with my toy-oscilloscope (one of those cheap DIY kits I've built some time ago) if the reset works at start up?
max

[1024MAK]

Yes, even the audio frequency ‘scopes should be able to see the reset pulse.

Mark

[MaxS]

Ok..
Main ICs back in place works good in a 16k configuration.
No more problems with warm resets.
When I fit the upper memory logic chips and tms4164s boot time varies but most of the times, when it boots to start screen, is almost like a 16k machine boot time.
Half of the times it doesn't boots at all or shows a corrupted screen.
ICs are ok.
Any idea?

[1024MAK]

If the DRAM chips are okay and are fast enough, then either there is a poor connection (socket, dry joint or cracked/broken track), or IC25 or IC26 may be faulty.

Have you seen details of how to disable the upper RAM?

Mark

[MaxS]

I've completely removed from the board (issue4a)
IC23, IC25, IC26 and the upper ram 4164s.
Those same chips plus IC24, IC3, and IC4 from the very same motherboard work without any issue on a issue 2.

I've already checked for continuity address and data lines between all the ICs
Is there anything else that can produce this kind of behavior?
Otherwise I've to dismantle the board, all sockets, and check with a microscope for micro cracks.
I don't want to give up

[RMartins]

Is the CPU socketed ?
Can you replace it and check if it makes a difference ?

[MaxS]

Hi
Yes all the ICs are socketed and tested on another board.
I've also already tried another z80 in this one with no success.
I'm starting to double check all components of the upper ram circuitry as the problem happens only with the upper ram installed

[1024MAK]

Disable upper RAM.

Mark

[MaxS]

Ok Mark I'll do.
Plan is to boot 16k with upper ram installed, turn off the switch to make selectable the upper ram, and run a test program that pokes and peeks decimal 85 and then decimal 170 in all the upper ram locations to check and report if something strange occurs.
This way I should see which group of addresses or data bits are misbehaving.. Right?

[1024MAK]

Yes

The other option would be to start the machine as a 16K, then load a RAM test program.

Mark

[MaxS]

I'll search for a ram test program. Any suggestions?
Otherwise they're just a couple lines of code.

[1024MAK]

ZX Spectrum Diagnostics

Wiki

Tape-Tests

Details

Mark

[MaxS]

Appreciated

[MaxS]

little update
the "hot machine reset" problem solved replacing C27 & R31 both out of specs, I was setting up the speccy as suggested by Mark to test the upper ram when I decided to install a spare set of tms4532-15NL4 instead of the new tms4164-20 I was using as test ICs untill now.
She booted like a charm, loaded all games and passed all tests.
This same full set of TMS4164-20 works without any problem on two issue2 I have and also on a C64 I use as a test machine.
Next days I'll proceed anyway as planned installing the switch between IC23 pin5 and +5v and perform a diagnostic check of this Issue4 coupled with those 4164 because I'm too curious now to understand what happend.
I guess it's an access time problem but same ICs work flawlessly in earlier motherboard revisions.
Looking at schematics I can't see anything that can affect RAM access time speed but I'm not an expert.. any suggestions?

[1024MAK]

Out of curiosity, what are the jumpers to the right of the Mic socket set to?
That is the TI, OKI, L, H, 3, 4.

Mark

[MaxS]

TI 4

[RMartins]

I believe there is a video in youtube, by a guy named JoulesPerCoulomb, if I remember correctly, that found a timing issue on ZX that looks very similar to this. Should be easy to find.
The solution I believe was to replace or add a Capacitor and a resistor, to delay a signal line.

Maybe you want to check that out.

Edit:
https://www.youtube.com/user/JoulesperCoulomb/videos
Not sure which video, but it won't hurt to see a few.

[MaxS]

thank you
I've already seen all of his videos some time ago.. it's time to see them again!!

[RMartins]

I believe the one you want is this:

https://www.youtube.com/watch?v=FY6UdjkBdLc

Although it is for +2, he explains the problem in the intermittent RAS/CAS problem.

I would also suggest you see the previous one too, if something doesn't make sense.