Troubleshooting X68000 SCSI boot failure with terminator checks

A Sharp X68000 that suddenly stops during startup can appear to have a serious system-board fault, yet the cause is often much simpler: an incorrectly terminated SCSI bus. A missing resistor pack, a second termination at the wrong end, or a device that is not receiving termination power can prevent the computer from finding its hard disk.

SCSI boot failures are especially common after installing a replacement drive, reconnecting an external enclosure, or moving an X68000 between storage locations. The machine may display a boot error, hang after its ROM startup screen, repeatedly access the disk, or return to a command prompt without loading Human68k.

Terminator checks should be performed methodically rather than by swapping random parts. The X68000 uses a narrow 50-pin SCSI bus, and its electrical rules remain important even when the storage device is a modern SCSI2SD, ZuluSCSI, or another solid-state replacement. A fast emulator can still fail if the bus wiring is electrically incomplete.

For Australian owners, replacement parts may arrive through eBay Australia, specialist retrocomputing sellers, or overseas postage from Japan. That makes careful diagnosis worthwhile: a resistor network is inexpensive, while shipping a complete drive or controller across Sydney, Melbourne, Brisbane, or Perth can quickly become costly.

Recognise the failure pattern

A terminator-related problem usually affects device detection rather than the X68000’s basic power-up sequence. The computer may reach a SCSI configuration screen but show no hard disk, report a SCSI error, or pause for an unusually long time while scanning IDs. If the machine boots from floppy disk but cannot access the hard disk, the SCSI path becomes a strong suspect.

A complete freeze can also result from a bus conflict. Two devices may share the same SCSI ID, or both ends of the cable may be terminated when only the physical ends should be. In other cases, the device appears intermittently: it boots when the case is moved, fails after warming up, or works with a short cable but not with an external enclosure attached.

Separate SCSI symptoms from unrelated display faults. A machine that boots normally but shows an unstable picture may have a video or monitor issue, while a machine that never reaches disk access points towards storage, power, or SCSI signalling. The differences between X68000 graphics modes are useful when diagnosing a display complaint, but they do not explain a disk that vanishes from the SCSI bus.

Power down and inspect the bus

Switch off the X68000, disconnect the mains lead, and allow the power supply to discharge before opening the case. Australian domestic power is nominally 230 volts, and the computer’s internal supply contains hazardous voltages even after shutdown. Do not work inside the power supply unless you have the appropriate training and test equipment.

Photograph the existing cable arrangement before changing anything. Trace the 50-pin ribbon cable from the X68000 SCSI connector to every device, including any internal-to-external adapter. Look for crushed ribbon cable, reversed connectors, missing keying, bent pins, loose sockets, and insulation damaged by a sharp metal edge.

Check for oxidation on older connectors. An X68000 stored in a humid coastal area such as Sydney or Brisbane may develop contact problems that are less obvious in a dry indoor environment. Do not scrape gold-plated contacts aggressively; use suitable electronics contact cleaner sparingly and allow it to evaporate fully before reconnecting power.

Confirm that the storage device receives the correct supply voltage. A drive that spins up does not necessarily have clean logic power. Measure the supply at the drive connector if you are equipped to do so, and inspect any Molex adapter or custom power lead for reversed wiring. A weak or repaired power supply can create symptoms that resemble a SCSI termination fault.

Confirm IDs and cable order

Every SCSI device on the bus needs a unique ID. The host adapter in the X68000 commonly uses ID 7, so the hard disk or SCSI emulator should normally use another value such as ID 0. The exact preferred boot ID can depend on the software and configuration, but duplicate IDs must be eliminated first.

Disconnect all optional devices for the initial test. Remove external hard disks, scanners, optical drives, ZIP units, and SCSI peripherals, then test the X68000 with only its intended boot device. If the computer starts correctly, reconnect one peripheral at a time until the failure returns.

The cable should form a single electrical path rather than a collection of long branches. A short internal cable is preferable for diagnosis. External SCSI boxes, pass-through connectors, and unused cable branches increase reflections and make termination placement harder to verify.

Inspect the connector orientation at both ends. Pin 1 is usually marked by a coloured stripe on the ribbon cable, although the stripe alone is not proof that the connector is correctly aligned. A reversed cable can prevent detection and may stress a device if power connections have also been modified.

Check physical terminator placement

A narrow SCSI bus requires termination at its two physical ends, not at every device. In a basic X68000 setup, the host controller represents one end and the internal storage device represents the other. The terminator should therefore remain fitted at the storage end if no external equipment extends the bus.

If an external device is connected after the internal drive, the last external device becomes the far end. The internal drive must then have its terminator removed or disabled, while the external unit retains termination. An unused terminator on an intermediate device can cause reflections, unreliable detection, or total bus failure.

Many older drives use removable resistor packs, often small single-inline networks installed near the SCSI connector. Others provide a jumper or switch marked TERM, TERM PWR, or similar. SCSI2SD-style boards may use configuration settings instead of a physical resistor pack, so consult the board documentation rather than assuming the circuit is active.

Do not confuse a missing resistor pack with a missing jumper. Some drives need the resistor network installed and a separate jumper enabled for termination power. Others have built-in active termination that cannot be removed. Record the original position of every jumper before changing it, and make one change at a time.

Verify termination power

Termination resistors need a stable supply, commonly called TERMPWR, on the SCSI bus. The host or a selected device may provide it, and a protective diode is often used to prevent one device from feeding power backwards into another. If no device supplies termination power, the bus can behave as if its terminator is absent.

With the system disconnected from mains, inspect the termination-power jumper and any fuse or diode associated with it. If you use a multimeter, check the relevant voltage only when you are confident about the connector pinout and probe placement. A short between adjacent 50-pin contacts can damage the host adapter or storage device.

Do not assume that a modern solid-state board automatically solves this issue. Some adapters draw little current but still require TERMPWR for their terminator network. Others can supply termination power but should not do so when another device is already supplying it, particularly on a bus with external equipment.

An unstable power supply deserves attention at this stage. Recapped or replacement supplies may have incorrect output, poor grounding, or excessive ripple. The Nereid expansion board is a useful example of the wider hardware work carried out around the platform; its Nereid-X documentation also illustrates why expansion hardware should be installed with careful attention to power and signal connections.

Test the bus with a controlled setup

Start with the simplest working arrangement: X68000 host, one short ribbon cable, one storage device, one unique ID, and termination at the two physical ends. If the internal device is the only peripheral, remove external cables entirely. This test establishes whether the core SCSI path works before you investigate enclosures or adapters.

Boot from a known-good floppy or other reliable startup medium if available. Use the system tools to scan for the device and observe whether its ID appears consistently. A device that appears on one boot and disappears on the next suggests a connection, power, termination-power, or timing problem rather than a damaged filesystem alone.

If the device is detected but the operating system fails to load, inspect the disk image, partition structure, and boot files separately. A SCSI controller can communicate correctly while the media contains corrupted Human68k system files. Avoid repeatedly writing to an ageing hard disk until you have made a sector-level backup.

Reconnect external equipment only after the single-device test succeeds. Add the next device, assign a different ID, and move termination to the new physical end. Test again. This approach may seem slow, but it prevents several simultaneous changes from hiding the actual fault.

On a bench in an Australian workshop, avoid relying on an untested mains extension board or crowded power board while troubleshooting. Use a sound outlet, keep the case supported, and maintain proper earth connections. If the unit has a damaged plug, burnt connector, or questionable imported power conversion, repair that condition before extended testing.

Replace parts without creating new faults

Original X68000 hard disks can be noisy, slow, and difficult to source. Modern replacements are practical, but compatibility varies between SCSI emulators, CF adapters, SD adapters, and period controllers. Confirm that the selected device supports the X68000’s narrow SCSI signalling and expected command behaviour.

A passive resistor terminator may work perfectly on a short, slow bus, while an active terminator can be more tolerant in a longer arrangement. The important point is correct placement and compatibility, not simply choosing the newest-looking component. Never install multiple terminators because a seller describes a board as “SCSI ready”.

Keep a written record of device IDs, jumper settings, cable direction, termination status, and test results. This is particularly useful when ordering from Japanese sellers or Australian marketplace listings where product photographs may show a different revision. It also prevents the common mistake of returning a known-good jumper to the wrong position.

If the fault remains after termination is verified, investigate the X68000 SCSI controller, cable continuity, oscillator or timing circuitry, and the storage device itself. A logic analyser or oscilloscope can help an experienced technician, but visual inspection and controlled substitution should come first. Replacing an irreplaceable controller chip without evidence can turn a recoverable machine into a much harder repair.

Compare likely causes before ordering parts

The following checks help distinguish a termination problem from other common causes of an X68000 SCSI boot failure. They are arranged as practical clues rather than absolute rules, because old hardware can present overlapping symptoms.

Symptom or finding	Likely cause	First check
No device appears after a new installation	Incorrect ID, reversed cable, absent power, or disabled terminator	Confirm ID, pin 1, supply voltage, and termination setting
Device works alone but fails with an external unit	Incorrect end termination or duplicate ID	Remove internal termination and terminate only the external end
System hangs during SCSI scan	Bus conflict, poor cable, or missing termination power	Disconnect peripherals and test one device with a short cable
Device is detected but will not boot	Damaged filesystem, incompatible image, or media failure	Boot from floppy and inspect the disk independently
Device disappears after warming up	Power-supply instability, marginal connector, or failing drive	Check voltages and repeat a cold-versus-warm test
Several devices respond unpredictably	Duplicate IDs or excessive cable branches	Assign unique IDs and simplify the bus layout
Terminator checks pass but nothing is detected	Controller fault, cable continuity issue, or incompatible adapter	Test with known-good cable and storage hardware

A terminator check is most useful when combined with isolation. Testing resistance with the power disconnected can provide clues, but readings vary according to the devices still attached and the circuit design. A resistance result alone should not be treated as proof that the bus is correctly configured.

For valuable machines, preserve the original parts even when a modern replacement works. Keep resistor packs, jumpers, cables, and the original drive in labelled bags. Australian retrocomputing communities often trade parts through local forums, swap meets, and marketplace listings, and an original component may later help diagnose another X68000.

Work through the checks in order: simplify the bus, confirm IDs, inspect the cable, verify termination at the physical ends, check termination power, and then examine the disk contents. That process gives the X68000 the best chance of returning to reliable SCSI booting without unnecessary part swaps. Document the final configuration and store a backup of the working disk image so the repair remains repeatable.