ZX Spectrum Issue 3 repair

[1024MAK]

Output was the same, all black noisy output. No border showing. Just totally black with noise.

That's not good. The voltages are okay. They will be a little lower than you would expect because they go via resistors.

If the ULA, the ROM and the Z80 CPU are okay, and there are no shorts on the address, control and data busses, even with no working RAM, the CPU should start executing the ROM code. One of the first things the ROM code does is set the border to white.

if there is no lower RAM, the ULA will make the best it can of whatever levels are detected on the data bus at the time it tries to fetch screen data. You will not get a static picture with flashing blocks.

Did you say you have tried swapping the ULA?

Mark

[Tezz]

Did you say you have tried swapping the ULA?
Mark

I tried swapping the ULA from the issue 3 board earlier (which is still the one inserted now) before I tried the IC23 jumper. I don't know whether either ULA is functional though as both boards are faulty. I've removed the temporary jumper now, it's probably best to solder a jumper properly in the morning.

[1024MAK]

The wire to disable the upper RAM only needs to be present and making contact during the first 5 seconds during and after power up. The purpose is to disable the RAM during the time that the code in the ROM checks to see how much RAM is present. After that, the code in the ROM will ignore the upper RAM unless a user program accesses it (e.g. with a POKE, PEEK or USR instruction).

If there is a problem and the Z80 CPU is executing rubbish, it does not really matter if the upper ROM is disabled or not, as control has been lost anyway...

A temporary soldered connection is however better, although if you have a on/off (SPST) switch, I suggest wiring that in, it may come in useful later.

Can you post a photo of your black screen please. Also, this is via the TV out (RF/UHF/aerial) socket, yes?
Can you confirm that your TV does actually work with retro computers.

Mark

[Tezz]

The wire to disable the upper RAM only needs to be present and making contact during the first 5 seconds during and after power up. The purpose is to disable the RAM during the time that the code in the ROM checks to see how much RAM is present. After that, the code in the ROM will ignore the upper RAM unless a user program accesses it (e.g. with a POKE, PEEK or USR instruction).

thanks, understood

Can you post a photo of your black screen please. Also, this is via the TV out (RF/UHF/aerial) socket, yes?
Can you confirm that your TV does actually work with retro computers.

ok, I'll post a photo in the morning. I am using RF out and the TV definately works with other Spectrums and 81s. I tried also retuning on a spare channel just in case the preset chn34 of the modulator is not quite set the same as others.

[Tezz]

Here's the photo attached of the output. The upper RAM is enabled again without the jumper in place although as mentioned the output was unchanged.

It's a CRT TV so I've taken the photo close to the screen to try to show the noise against the black background.

issue_3b_output.png (532.52 KiB) Viewed 3195 times

[1024MAK]

Hmm...

Have you tried a temporary composite video mod?

Mark

[Tezz]

Have you tried a temporary composite video mod?

Do you mean it looks like the RF output might be suspect? I haven't touched the modulator so far, I was intending to do either the capacitor or transistor composite mod once it was functional. I have both parts here, should I try doing that next with the capacitor?

[1024MAK]

With the picture you posted, it is very hard to determine if this is a video fault or a system fault.

If you do the capacitor composite mod, it may help with the fault finding, as it removes the modulator and tuning issues... (assuming the ULA is generating a picture and there is no fault with the rest of the video circuitry).

You say you have a lower RAM module (the type that is a PCB with modern SMD chips). It is worthwhile fitting this. Then see if it makes any difference. Also if there is lower RAM, the CPU, ULA and ROM fitted, if the keyboard is fitted (or use a wire - see below), you can press some keys and see if the loud speaker gives key clicks.

If you get a logic probe, and / or a logic analyser (an inexpensive one that can do > 20MHz), it will be possible to get some idea if the Z80 CPU is running code.

The keyboard is a matrix of switches. To save wear, for testing purposes, you can use a insulated thin solid core wire (0.6 to 0.9mm diameter is fine). Use the wire (with the ends stripped to about 5mm) to connect one of the contacts in the left hand keyboard connector to one contact in the right hand connector, then remove. This simulates a key press.

Mark

[Tezz]

Thanks Mark, I'll do the composite mod tomorrow. Ideally I would like to have a known working ULA installed too to take that out of the equation. Preferably I'd like to buy a 6C001E-7 to keep around for testing or maybe the Nebula ULA replacement. If the composite mod doesn't help get me any further I'll replace the capacitors and install the lower RAM module next and see how that goes. If there is still a problem I'll need to take the time to desolder the CPU and ROM to install sockets. I have a couple of new old stock CMOS Z80s now that I can use if needs be. I'll see how it goes tomorrow thanks for your continued help and advice.

[Tezz]

I've now completed the capacitor composite mod which went well without any problems. Connecting up to the TV however I get no output at all with the TV just awaiting input like nothing is connected (I have used this TV with both RF and composite outputting Spectrums and 81s) I've swapped back to the original ULA that came with this board. I was going to recap the board and install the lower RAM module next but I was thinking perhaps I should test the CPU with a logic probe next instead. It'd be good to have a known tested and working ULA, ROM and, CPU for diagnostics. Trying to diagnose a faulty board with untested ICs isn't ideal. This board has one of those narrow type sockets for the ULA by the way so maybe I should just desolder all the main ICs and install fresh sockets to then continue testing and swapping ICs.

EDIT: I might have a functional 6C001E-7 ULA that I can try tomorrow, if so and the board still doesn't work according to some further info I found online, I should check TR1 and TR2 next then move on to check for a short between C65 and R53