ZX Spectrum Issue 3 repair

[1024MAK]

Yep, pin 1 is indicated by the dot and is the input. The middle pin is 0V, and the right hand pin is the +5V output.

With a 7805, the left hand pin is the input, the middle pin is 0V, and the right hand pin is the +5V output.

So, your picture shows the Traco positioned correctly

See also this picture from the www.mutant-caterpillar.co.uk site:


Mark

[Tezz]

Thanks Mark for the confirmation. I've installed the DC converter now and powered the board on. I have the meter set to 20 on DC voltage and the readings don't appear correct. 5v shows 3.42, 12v shows 8.3, -5v shows -0.9 and, 0v shows 0.04 Perhaps TR4/TR5 need replacing. The only visual thing I can see on the board is that the positive terminal of capacitor C46 has lifted a little and is touching the trace next to it. I thought solder had run through and I'd wick it off but it seems to be the through hole thats lifted and sat on the trace. Not sure which way to tackle that

[PokeMon]

Well - did you check the battery of your multimeter ? There should be a symbol somewhere in the display. Getting 3.4V instead of 5V is suspicious.

[Tezz]

Well - did you check the battery of your multimeter ? There should be a symbol somewhere in the display. Getting 3.4V instead of 5V is suspicious.

hmm that's a good point no actually I haven't tried replacing the battery, it showed the correct voltages on the issue 3b board that's now functional when I was working on that a few days ago but I guess it could be getting low as I've been using it quite often recently.

EDIT: well that's embarrassing, I just opened the multimeter and the battery says use by 2014 With no low battery symbol appearing I didn't think about that at all. I'm off to the supermarket now to buy a new one

Thanks, Karl.

[Tezz]

I've replaced the multimeter battery now with a new Duracel one and retested the board. Unfortunately it gave exactly the same readings as before. Well I did see -5v at the start of the test but then when I tested again it then gave a low reading. I noticed now it was powered on for a minute that the Z80 was extremely hot however just like the '81 that I worked on recently. That was also an original Zilog type and was the cause of the fault with that board. I'll have to remove the Z80 in the morning and install a socket to try one of the new CMOS Z80s that I bought a week or two ago or maybe just risk an NEC Z80 first just incase some other circuit problem killed the CPU.

[1024MAK]

Remove the Z80 CPU, then power up again and test the voltages of the +5V, the +12V and the -5V. You don't have to keep checking the 0V, as this is a common "ground".

If the +5V is okay, then power down and fit a socket. Power up, check the voltages again. If okay, power down. Wait for a few minutes for the capacitors to discharge. Then fit a replacement CPU. Hopefully when you now power up, both the supply voltages will be okay, and the CPU will be attempting to do something.

Mark

[Tezz]

I've removed the Z80 now and retested. The voltages are looking more like it : 5v = 4.88, 12v = 11.81, -5v = -4.87. The NEC ROM chip is also getting extremely hot however to the point that you can't touch it after less than a minute of the board being powered on. Could that be an indication that there's still a circuit issue to resolve or has it perhaps failed also? Would C46 positive terminal touching the trace next to it cause this? I still need to fix that somehow without damaging the track. The DC converter has a slight audible buzz/hum, not sure if that's expected. I'm using a serviced and recapped original psu. Should I be concerned the 0v shows shows 0.04? My other boards do show 0.00 on the meter.

[1024MAK]

ROM chips should never run hot. It has been damaged and needs replacing. If you don't have a spare Sinclair ROM chip, a EPROM will need to be programmed and a modification done to the board so that the EPROM works correctly. I'll post up these details if you need them.

I don't know what effect C46 positive terminal touching the trace next to it will cause, as I don't have an issue 3 board to hand to look at. I may be able to help if you can post up a photo...

The DC converter is working overtime supplying lots of power that the ROM is converting into heat... so I'm not surprised that it has a slight audible buzz/hum.

Should I be concerned the 0v shows shows 0.04? My other boards do show 0.00 on the meter.

No. 0.04V = 40mV. Hopefully once it is working and no more chips are trying to be cookers (well, apart from the ULA...), this should drop to a lower figure. But even 40mV is okay. It's caused by the current flowing through the PCB tracks. Thin PCB tracks have a slight resistance, current flowing through a resistance will result in a proportional voltage across that resistance. In electronics it's often called "volt drop". The same problem occurs with long cables.

Mark

[Tezz]

Thanks Mark, I'll remove the ROM too and retest. Luckily I can buy another NEC manufactured ROM chip. I'll look to buy a socket also.

I've taken a photo of C46 to show you but It's also helped me to see more clearly now that actually part of the through hole ring on the top of the board has come away and is sitting on the track. Maybe if I apply a little heat with a fine tip on the iron and carefully push it away from the track with a micro screwdriver. I'd need to be very careful not to touch the trace.

Hopefully nothing else is at fault, otherwise I'll be left with a bare PCB soon

Capacitor_C46.png (442.64 KiB) Viewed 3445 times

[Tezz]

I've removed the ROM chip and retested, slightly higher readings now : 5v 4.91, 12v 11.93, -5v 4.83, 0v 0.03. The DC converter still has the same buzz and I notice that the upper RAM chips all feel hot too. I should check the DC converter pins maybe to see what's happening there?