Long time listener, first time poster. Dead - ex spectrum

[sirmorris]

So I bought a 16k speccy and of course it's DOA.

Voltages check out on the lower RAM, but screen display is...

crapola.jpg

(1.63 MiB) Downloaded 368 times

RAM?

I have a DivIDE. I thought I saw a hint that you could install some kind of test ROM on it to ... test the board. Is that possible?

Ta.

[1024MAK]

First time poster with 2569 posts so far

Yes, that Spectrum looks like it has a "lower" DRAM fault. When you say that the voltages check out, I presume you mean that at three (+5V, -5V and +12V) are okay.

Does this machine have 48k bytes of RAM fitted? Are the "upper" DRAM chips fitted in sockets? If yes, unplug them.

Does it always power up with a white border? Try it five or six times. If yes, the white border indicates that the CPU is running the first part of the code in the ROM chip, so it's likely that both CPU and ROM are at least partially functional.

The DivIDE has both on-board ROM and on-board RAM. Not tried it myself, but yes, I believe you can use it to run a ROM based RAM test program.

Brendan has improved and expanded on some existing ideas and code for testing ZX Spectrum memory. See here. But does not target the DivIDE.

However, I recommend you investigate this DivIDE web site here

Mark

[Paul]

Hi Charlie
is your ROM of the speccy in a socket?
Then you can insert a test ROM directly there.
If you are interested I can mail you several test ROMs.
I have used one of my w27C256 EEPRoms with eprom pin layout without any modification in my ROM Socket to find out which RAMs were faulty.
I tried to desolder the broken RAMs but the solder seems to be VERY difficult to remove.
So I decided against replacing single RAMs and cut out all and replaced them with a lower RAM module bought at SMR.
If the spectrum fails its mostly because of wrong voltages at the DRAMs. So all are replaced by SRAM now (lower and upper RAM, just to be shure).
This will keep my spectrum alive for the next 35 years to come.
Kind regards
Paul

[sirmorris]

+12, -5 and 5v all ok on the lower RAM chips. I suppose that doesn't mean it'll always be good. The coil is chirping like a little cricket.

No upper RAM is fitted. It usually powers up to a white border. I've seen green a couple of times though.

Thanks for the tip Paul - I'll get the test ROM burned and just replace the inboard PROM.

[sirmorris]

So a couple of hours spent building software from sources (Cmake, sjasm, testrom) and the net result is that I've discovered the computer behaves exactly the same without a ROM in it at all

I suppose that narrows things down somewhat. ULA innit. Or processor perhaps. Bum, it's soldered in. And dead-bugged.

[dr beep]

You could try an external ROM that disables inner ROM to test if it is the ROM.
Not too hard

[1024MAK]

So a couple of hours spent building software from sources (Cmake, sjasm, testrom) and the net result is that I've discovered the computer behaves exactly the same without a ROM in it at all

I suppose that narrows things down somewhat. ULA innit. Or processor perhaps. Bum, it's soldered in. And dead-bugged.

In a otherwise working ZX Spectrum, with no ROM, the CPU will try to read the instruction at 0x0000, but with no ROM, it will only get whatever value is on the data bus. Hopefully this will be 0xFF due to the bus floating high. This is RST38H. It will then put the PC on the stack and jump to 0x0038. At which point the sequence will start all over again. Each time the stack will grow. At some point, the stack will fill up the area of the lower RAM used for the screen. You should then get a screen like this:-

Displayed image from a working ZX Spectrum when no ROM fitted:


Of course, if the CPU is not working correctly thus won't happen... Faulty lower RAM will cause a different display.

More tips in these posts:
https://www.worldofspectrum.org/forums/ ... ent_903340

http://stardot.org.uk/forums/viewtopic.php?t=10896

http://stardot.org.uk/forums/viewtopic. ... 27#p121372

Mark

[sirmorris]

Thanks for the tips. I have a thing about replacing processors so I did that anyway.

Result:

bitbetter.jpg

(3.41 MiB) Downloaded 313 times

Test rom gives 8 short beeps then a derr dip (much like a beeb startup). I get 8 red stripes after the memory test, indicating that all lower ram is dead. Ah well. It's a good job I left the solder sucker on.

[1024MAK]

Looking a lot better. Just looks like the RAM now. Although I think it is only one or two DRAM chips that are duff given that the ROM code got that far...

Mark

[sirmorris]

Well I had a tube of lovely ceramic AMD chips so I whipped all the old ones out and bingo!

So luckily it was just the processor and RAM. Easily fixed.

Phew!