Sneaky ZX Spectrum Fault Hiding the TRUTH

In this episode, I’m repairing a ZX Spectrum Issue 3B that came from a batch of computers previously worked on by another technician. While it initially seemed like a straightforward fix, it turned into a deep dive into the data bus. After ruling out the ULA and CPU, I discovered that the signals on data lines 5 and 6 weren’t reaching a logic high. The culprit? Two partially failed RAM chips (IC20 and IC21) that were dragging the bus down. It’s a great example of why modern chip testers can’t always be trusted—both chips passed several external tests but failed immediately when put to work in the actual machine. Key moments in this repair: * Initial Triage: Identifying previous “bodges” and a poor composite mod. * Power Rail Checks: Testing for shorts and verifying the voltage regulator. * Diagnostics: Using the Steam Card Pro to pin down IC11 as a failure point. * Advanced Troubleshooting: Using an oscilloscope to identify weak data line signals. * The Solution: Testing pull-up resistor values and replacing the faulty lower RAM.