No cursor on ZX81 display

[jab73]

Hello, I recently dug out my old UK zx81 (issue 1 motherboard). After installing a composite mod (ginger-electronic) and DC-DC converter I get a crisp white display . . . but no cursor. I had recently acquired a couple of working Timex 1000 computers, so I tried a bit of component swapping to see if I could fix the zx81:

ULA swapped . . . no change
Z80 swapped . . . no change
ROM swapped . . . no change
Inserted working Timex Sinclair 16K RAM module . . . no change (I read this would bypass the 1K on-board memory)
5V from the DC-DC converter looks good.

Would anyone have suggestions of steps I could take next? Thanks.

[1024MAK]

Did you try the ‘suspect’ chips from the ZX81 in the working TS1000? Did it still work? If yes, it proves the swapped chip is good.

Can you post a photo of your ZX81 board please.

A white screen normally means that the ULA is ‘okay’ (at least, it’s partially working) and hence the oscillator is also okay, but the Z80 is unable to execute instructions. So that could be a faulty Z80 (occasionally these fail, but they are normally reliable), a faulty ROM (again, occasionally these fail, but they are normally reliable), a RAM failure (the most common problem), a problem with the reset circuitry for the Z80 (C5 or less likely R15), a problem with the clock circuitry for the Z80 (TR2 and associated components).

Of course, there could also be a dry or broken solder joint, faulty socket contact or other problem with the board.

And yes, any 16K or larger external RAM pack deselects the internal RAM chip(s). Then normally a ZX81/TS1000 will run using the external RAM pack. However, the internal RAM chip(s) are still in circuit, so if it/one of them has suffered a failure that affected the address or data buses, that could still result in a non-working machine.

Apart from a multimeter, do you have any other test gear? Or have access to any?

My next step would be to measure pin 26 (RESET) on the Z80 to see if the voltage has gone high. Then short out the two legs of C5 (or short Z80 pin 26 to any suitable 0V/GND point). To see if that helps.

If you have a ‘scope, a logic probe or a logic analyser, then test pin 6 on the Z80. Then test the following pins:

Pin Function
16 INT (should be pulsing)
17 NMI (should be pulsing)
18 HALT (should be pulsing)
19 MREQ (should be pulsing rapidly)
21 RD (should be pulsing rapidly)
22 WR (should be pulsing)
24 WAIT (should be pulsing)
25 BUSRQ (should be high)
26 RESET (should be high, except for a very short time after power is applied)
27 M1 (should be pulsing)
28 RFSH (should be pulsing)

Then test address lines A14 and A15. These give an idea of which memory the CPU is accessing and if it is trying to generate a TV display.
You can also test the remaining address lines A0 to A13 and the data lines D0 to D7. A lack of activity is suspicious.

Mark

[jab73]

Hello Mark, thank you for your reply, I have attached an image of the board.

(the three wires routed to the modulator "can" are from the composite mod board)
I have not swapped the zx81 chips into the working Timex, I will do this and see what happens. I also do not have much beyond a multi-meter but I will look into your suggested actions. Once again, thanks you for your response.

[1024MAK]

The Z80 is fitted the wrong way round!

On ZX81 issue one and issue three boards (an issue two board has never been seen as far as I know), all the chips are orientated the same way round, that is, with the edge-connector/expansion port furthest from you, they have pin 1 on the left.

Note that the factory that made them did not always fit the sockets with the ‘D’ mark the correct way round! So this is a bit of a trap to fall into…

Mark

[jab73]

Mark . . . . you were 100% correct. I removed, rotated and re-installed the Z80 and got a crisp display with the cursor. Have now ordered a new keyboard to complete the effort. Thanks for the advice!

[1024MAK]

Hey, good to know it now works (well, apart from the keyboard).

Thank you for replying.

Have fun

Mark