Sinclair ZX Spectrum

Sinclair ZX Spectrum Issue 4S – RAM FAIL

This is the second board (out of the three) that I have bought in a lot of known defective:

Lot of 3 defective Sinclair ZX Spectrum boards

Table of Contents

Visual Inspection

The board looked innocent at first look.

One lower RAM chip was missing from the socket that was installed. On the back, there weren’t other signs of work.

Another interesting thing was that C27 capacitor, while still the original, had its terminals clipped and then resoldered:

Other than this, the board was dusty:

First power up

Upon first power up, the board draws around 540mA, video has thins flashing vertical bars. Voltages are good, but the Infrared camera shows one of the lower RAM ICs getting hotter than the others:

First power up

Infrared View:

Fixes

IC6 Replacement

I wanted to remove the IC6 that was getting hot, but my mind wasn’t there when I flipped the board. I promptly desoldered IC18. The one on the opposite side of the capacitors. Pffff. Noticed when I flipped the board back to check where I have to pull it from once I turn on the hot air station.

Left IC18 on board, but I did not solder it back yet.

I proceeded to remove IC6. The behavior did not change. I swapped ULA with the vLA82 and it behaves the same, however, at least now, on one of the diagnostics ROMs (ZX Diagnostics), I saw a “RAM FAIL” message.

Intermittent fault that isn’t there

While trying various ROMs, I noticed that the board is unstable. Tests were randomly freezing, sometimes completing, sometimes just freezing mid-runs. There was also absolutely no sound, even if the speaker does measure 41Ω and the diode is fine. Once I also heard it beep but sound was garbled. But, it then just stops working, it just shows weird patterns on the screen. Is the CPU faulty once it gets hot ? I can make the diag ROM go berserk if I push down on IC25 …

30 mins of trials and errors later -> the board works flipped up side down, with pure sound – not garbled. If I gently turn it up again -> it keeps working if I use the reset button on the diag rom board. But, if I leave it fall a bit when putting it face up -> it just stops working.

So a cold solder joint or a broken trace somewhere, I suspect chip or joint. Also, current consumption goes up (but maybe is normal, if it gets in a state where circuits stay active).

Before resoldering the ULA, since the solder joints look a bit dry, I went ahead and measured continuity between pins of ULA and pins of CPU, while the board was in a state of “not working”. After measuring 8 of them, I gave up, there was continuity, so I proceeded to reinforce the solder joints anyway – shotgun approach. But this did not fix the problem 🙁 just flipping the board up side down it would make it work.

I went to check the heat sink and the 7805 chip, but they were properly secured. When face down, something was heavy enough to start making contact. What else could ‘hang’ heavy ?! It was the Diagnostic Rom board … it was making a faulty contact 🙁

IC13 Is Cold

Having another look with the Infrared camera, something struck me as old. Only 5 warm spots on the lower RAM, but it somehow didn’t add up, I took one out and one was already out, something was missing.

I proceeded to remove IC13 too.

More ICs are faulty

I kept going on, removing all lower RAM ICs that were reported faulty. Depending on the diagnostic ROM, sometimes the all bad ICs are shown, sometimes only the “next” one. This was very tedious, and except 1, all lower RAM ICs had to be changed.

To complete the story, there was a faulty IC in the upper RAM too. Seen below in the IR camera and the reported also by the diagnostic ROMs:

Once everything was running fine with the vLA82, I decided to swap back the original ULA, and it works also. Sound is fine too:

Conclusions

I ended up installing 7 RAM ICs, 5 in the lower RAM (+1 that was already missing) and 1 in the upper RAM. Only 1 original lower RAM IC was still left on the board.

I am also used in seeing standard coil sizes, but this board had a smaller coil installed. Thankfully it works fine:

Few closing photos:

I am happy I rescued the board, the seller (I speculate the seller was the one working on the board, but I don’t know) gave up a bit too early. The ULA and ROM are functional. My reference 4S board now :).

What I learned is how to properly use the Electrical solder sucker. No more clogs now. I repeated the moves 2 x 8 x 7 = 112 times – and poor pump needed filter changes to make sure it continues to work fine.

Leave a Reply