White screen of death

[nama]

So after a year or so being in storage, today I powered up my ZX81 again, and would you believe it, the white screen with no 'K' cursor.
I then plugged in my ZXpand and I saw this.

IMG_6140.JPG (76.41 KiB) Viewed 4340 times

Certainly looks like the 'ZXpand commander', minus any text. I could move up and down, select something, and I think I even loaded 'forty niner'. Although the screen was like 'forty niner' but made from lines. No graphic per-say...looked funky

So, the machine looks to be running, but the image is all screwey. My first thought was that the character ROM had failed, but I' don't believe the ZX81 has a stand alone character ROM...so maybe the 'standard' ROM has failed. Does this contain the character set?
But then again, if it was the characters that had died I would have thought that 'forty niner', which uses high resolution routines, would have worked properly, but it doesn't.

So my uneducated guess is the ULA is acting up.

Unfortunately everything is non-socketed, so it would be great to have some second opinions before tackling any desoldering.

Are there any definitive ULA test I can be performed with either a scope or a logic probe?

Much thanks

Phil

[mrtinb]

Try looking in this thread, when my ZX81 had died with white screen, and I used a digital probe:

viewtopic.php?f=7&t=1851

[1024MAK]

The ROM contains the BASIC interpreter, the system routines and the character data. So it is unlikely that the ROM has failed. As you have used an expansion that has it's own RAM, it's not a RAM chip problem.

It sounds like the Z80 CPU is working, so that leaves the ULA. The pin on the ULA that outputs the video signal uses three levels. If you have a 'scope you could test it. Keep in mind that it is outputting a TV signal.

Have you tried a different display / TV?

Are you using UHF/RF, or composite video to feed your display? If using composite video, which circuit are you using?

Mark

[sirmorris]

It could be the ULA failing to enable the ROM.

Then again I wouldn't be so quick to dismiss a ROM fault. If you tie ROMCS high on the edge connector you will get all black characters as the ULA gets FFs when it tries to access the internal ROM.

You're seeing this effect with ZXpand attached - the ZXpand's overlay (a full but hacked copy of the ROM) is able to allow the machine to run, but it can't deliver the character data.

Without ZXpand you're seeing a white screen, which _could_ be a Z80 that is executing RST 38s - opcode FFh.

[1024MAK]

@Charlie - Good point about the ZXpand overlay ROM.

@Phil - is the ROM plastic chip with Sinclair printed on it (with no window), or an EPROM (has a window, which may, or may not a sticker or paint over the window)?

The former is a mask ROM, these are usually fairly reliable (but that does not mean they don't fail). While the later will eventually fail.

Also do you have a normal 16k byte RAM pack that you can try?

Mark

[nama]

Hi guys,
Thanks for all the great insights.

To answer some of your question about the ROM, it looks like a ceramic Motorola masked ROM (basically like a ceramic EPROM without the window). Has ZCM38818C written on it.

I also tried putting a 16K RAM pack in, but the results were the same...white screen of death.

I looked at pin 16 (TVTAPE) in the ULA, and I believe I can see 3 levels present.

IMG_6141.JPG (23.31 KiB) Viewed 4278 times

finally I probed each pin on the ULA with a logic probe, with the ZXpand installed, and then again without.
Here are the results.

Code: Select all

Pin  ULApin  ULA  My ULA (ZXPAND)   My ULA (NO ZXPAND)
     name                                 
1    A7'     P    P                P     
2    A8'     P    P                P     
3    A2'     P    P                P     
4    A1'     P    P                P      
5    A0'     P    P                P      
6    /RD     P    P                P      
7    /IORQ   PH   P (PH?)          L       
8    /WR     PH   P (PH?)          P      
9    /MREQ   P    P                PL      
10   /M1     P    P                P      
11   A14     P    P                P     
12   /RAMCS  P    P                P      
13   /ROMCS  P    H                NOTHING      
14   CLK     P    PL               PL      
15   /NMI    PH   P                P      
16   TVTAPE  PH   H                H     
17   /HALT   P    P                H      
18   A15     P    P                P     
19   D7      P    P                HIGH LOW BUT NO PULSE      
20   TAPEIN  L    L                L      
21   D6      P    P                P      
22   USA/UK  OC   H                H     
23   D5      P    P                P     
24   D4      P    P                P      
25   KBD4    H    H                H      
26   D3      P    P                P     
27   KBD3    H    H                H      
28   D2      P    P                P     
29   KBD2    H    H                H      
30   D1      P    P                P      
31   KBD1    H    H                H      
32   D0      P    P                P      
33   KBD0    H    H                H      
34   GND 0V  L    L                L      
35   OSC     H    H                H      
36   A3'     P    P                P      
37   A4'     P    P                P      
38   A5'     P    P                P      
39   A6'     P    P                P     
40   +5v     H    H                H      

Seems there are a few anomalous things happening. one seems to be around the ROM chip Select pin 13.
And also what D7, /HALT and /IORQ are doing when no ZXpand is installed.

Hoping this may give someone some insights.

Thanks again

Phil

[1024MAK]

I haven't looked at the signal details you have posted yet, but it is looking very much like the ROM has failed.

Mark

[sirmorris]

I agree with Mark. ROM swap - that's the first thing to try.

What do you mean by 'nothing' on ROMCS? Is that hi-z or inactive? If hi-z that's strange because there's a pull-up on the line. Could the probe be interacting with it? Does the ULA actively control ROMCS or is it asserted/low all the time, leaving !RD to control the output? If it was low all the time I'd have assumed the designers would have saved themselves a pin on the ULA...

The disparity on some of the other signals. IRQ/IORQ etc will be caused by the fact that the ZXpanded test was running code, but the naked run wasn't.

[nama]

Hi guys,
Well, when I probed D7 with the logic probe it literally showed no high, low or pulsing lights. I looked at the resistor and that looked within spec... Or at least the same as all the other for D0 to D7.
Strange.

I'll whip out the ROM and try another.

Cheers

Phil

[nama]

ROM removed socketed and replaced with another from a spare machine. Unfortunately there was no change

What's next... The ULA?