Using an Oscilloscope to Measure X68000 Power Supply Ripple

The Sharp X68000 is old enough that a working power supply should be treated as a component under investigation, not an automatic sign of good health. A computer may boot, draw a picture and run games while excessive ripple on its DC rails slowly stresses memory, custom chips, floppy-drive electronics and expansion boards. Measuring that ripple gives a much clearer picture than checking the output with a multimeter alone.

An oscilloscope shows the changing voltage that sits on top of the nominal supply. That makes it useful for finding tired electrolytic capacitors, high-ESR parts, switching spikes, poor regulation and load-related faults. The work is manageable on a careful electronics bench, but the supply contains hazardous mains circuitry. In Australia, that means respecting a 230 V, 50 Hz input and keeping primary-side probing firmly in the territory of a suitably qualified technician.

Why Ripple Matters In An X68000

A multimeter normally reports an average or RMS value. If the X68000’s 5 V rail reads 5.02 V, that measurement does not reveal whether the line is clean or carrying 100 mV of sawtooth ripple and sharp switching transients. Digital logic can tolerate some variation, yet repeated noise may cause intermittent crashes, corrupted floppy writes, audio faults or failures that appear only after the machine warms up.

The X68000 power supply generally provides a logic-oriented 5 V rail and additional rails such as 12 V, with the exact arrangement depending on the model and supply design. The 5 V output deserves particular attention because the main board, memory and many expansion circuits depend on it. The 12 V line is significant for drives, analogue sections and motors. Measure only the rails actually present in the unit, using the relevant service information rather than assuming every X68000 has an identical connector or voltage layout.

Ripple is usually discussed as peak-to-peak voltage. For example, a trace moving between 4.97 V and 5.03 V has 60 mV peak-to-peak variation. There is no single magic limit that applies to every original Sharp supply, so compare readings with service documentation and the behaviour of a known-good unit where possible. A stable rail with modest, regular ripple is a different problem from a rail carrying large bursts, high-frequency spikes or a slowly increasing waveform under load.

A healthy-looking startup is not enough. A floppy motor starting, a hard disk accessing data or a Nereid-X and other expansion hardware drawing current can expose weakness that remains hidden at the desktop screen. Ripple testing should therefore include both a quiet idle condition and realistic activity.

Preparing The Bench And Scope

Use an oscilloscope with a properly earthed mains lead, a 10× probe and sufficient bandwidth for switching-supply work. A 20 MHz instrument is useful for general ripple checks, while a faster scope can reveal narrower spikes. The scope’s bandwidth-limit function, often set to 20 MHz, is valuable because it reduces radio-frequency pickup and makes the low-frequency ripple easier to judge.

Before connecting to the computer, compensate the probe using the scope’s calibration output. Check that the probe is set to 10× on both the probe body and the instrument. Select DC coupling when you want to see the rail’s actual voltage, then switch to AC coupling when examining small ripple around that voltage. AC coupling is convenient for viewing a 40 mV waveform on a 5 V rail, but record the DC voltage separately because AC coupling hides regulation errors.

The probe ground lead is the main source of misleading measurements. Its long wire acts as an antenna and can create impressive-looking ringing that is mostly an artefact. Use a probe ground spring or the shortest possible connection to a nearby secondary-side ground point. Keep the tip and ground connection close together, and avoid draping the probe lead across the switching transformer or mains wiring.

An Australian bench often includes a Jaycar or Altronics probe accessory, a second-hand scope and a power board crowded with chargers. Keep the X68000 supply away from loose metal tools and make room for the probe cable. A neat setup is more than tidiness: it lowers the chance of touching a primary-side heatsink or accidentally shorting adjacent pins.

Finding Safe Test Points

Disconnect the console from mains before opening it, and allow time for capacitors to discharge. Do not assume that a capacitor is safe merely because the unit has been unplugged. The primary reservoir capacitor can retain a dangerous charge. Ripple measurements for routine diagnosis should be made on the isolated, low-voltage secondary side, not across the primary switching components.

A safe test point is usually a secondary-side output capacitor negative terminal for ground and its positive terminal for the corresponding rail. A connector pin can also work, provided its pinout is confirmed. The metal chassis may be connected to protective earth, signal ground or neither, depending on the design, so use the circuit’s secondary ground rather than treating the chassis as an automatic reference.

Never connect a grounded oscilloscope probe to a point on the primary side unless you fully understand the topology and use appropriate professional equipment and procedures. The probe ground is connected to protective earth through the oscilloscope. Clipping it to a live primary point can create a direct short circuit, damage the scope and expose you to lethal voltage. An isolation transformer does not make careless probing safe; it changes the supply arrangement, not the fundamental hazards.

For a first test, reconnect the secondary-side harness and place the probe on the 5 V output capacitor or a verified 5 V connector pin. Set the scope to DC coupling, 1 V/division and a timebase around 5 ms/division. Once the rail is centred and stable, use AC coupling and reduce the vertical scale to 20 mV or 50 mV/division. Adjust the timebase between 10 µs/division and 10 ms/division to inspect both switching spikes and slower capacitor ripple.

Taking Useful Measurements Under Load

Record the rail at power-on, at idle and during activity. Watch the startup waveform for overshoot, delayed settling or a rail that repeatedly rises and falls. At idle, note the peak-to-peak ripple and its frequency. A regular waveform at twice the mains frequency may point towards rectifier and bulk-capacitor issues, while a high-frequency pattern may relate to the switching stage, output capacitors or probe technique.

Next, exercise the machine. Boot from floppy, access the drive repeatedly and run software that produces sound or graphics. If the computer has a hard disk or expansion board, include those in the test. Let the unit warm for at least 15 to 20 minutes, because ageing capacitors can show increased ESR and higher ripple once their temperature changes. A Brisbane summer bench and a cool Melbourne winter room can produce noticeably different thermal conditions, so note the ambient temperature with the measurement.

Compare the scope trace with the rail’s DC reading. If the 5 V output falls significantly when the floppy motor starts, the issue may involve regulation, wiring resistance, a connector, the drive load or failing capacitors. If the DC value remains correct but the ripple grows sharply, suspect output filtering, ESR or a feedback problem. If only the trace is noisy with the probe’s long ground lead, repeat the test with a spring ground before replacing anything.

Save photographs or scope screenshots with the vertical scale, timebase, coupling mode and load condition visible. A note such as “5 V, 42 mV peak-to-peak, warm, floppy access” is much more useful than “looks okay”. The same method can be applied after service, allowing a before-and-after comparison. The practical PSU recapping guide is a useful companion when ripple and capacitor age point in the same direction.

Interpreting Results And Choosing Repairs

Ripple alone does not identify a failed component. Old electrolytics are a common cause, but poor solder joints, cracked connectors, dried adhesive, failed rectifiers, regulation faults and excessive load can produce similar symptoms. Inspect the board for bulging or leaking capacitors, darkened areas and disturbed solder joints, then test systematically rather than replacing parts at random.

An ESR meter can support the diagnosis with the supply unplugged and discharged, although in-circuit readings may be affected by parallel components. Capacitance measurements are also useful but should not be treated as a complete health check. A capacitor can measure close to its marked value while having poor high-frequency performance or elevated ESR. If recapping is appropriate, use reputable parts with suitable voltage, temperature and ripple-current ratings, and observe polarity carefully.

Check connectors and wiring as well. A corroded contact can create voltage drop that appears only when the drive starts. Measure directly at the power-supply output and again at the main-board connector. The difference between those readings can reveal resistance in a plug, crimp or cable. This is especially relevant for machines that have spent decades in storage or have been imported through the Australian second-hand market, where an X68000 may arrive via eBay Australia or Gumtree with unknown repairs and replacement plugs.

The following guide helps separate the most common patterns. It is a diagnostic aid rather than a substitute for the circuit diagram or a qualified repairer.

Oscilloscope observation	Likely direction	Useful next check
Small, regular ripple with a steady DC rail	Normal operation or acceptable ageing	Compare with a known-good supply and check under load
Large low-frequency ripple on 5 V or 12 V	Worn bulk or output filtering	Inspect electrolytics, rectifier path and ESR
Sharp spikes that disappear with a short ground spring	Measurement artefact	Repeat using a spring ground and nearby reference
Ripple increases after warm-up	Temperature-sensitive capacitor or regulator fault	Test warm, inspect capacitors and compare rail voltage
DC voltage droops when a drive starts	Load, connector resistance or weak regulation	Measure at both the PSU and main-board ends
Bursting, pulsing or unstable switching waveform	Control-loop, startup or overload problem	Stop extended operation and obtain service-level diagnosis

If a replacement supply is being considered, confirm its output voltages, current capacity, connector pinout, polarity and earth arrangement. A modern substitute can be electrically sound while still creating mechanical, cooling or noise problems. Retaining the original enclosure may require insulation and strain-relief work. Never fit a generic adapter simply because its label says 5 V; the X68000 needs the correct complete set of rails and a suitable return path.

The broader X68K.NET resources include preservation notes, hardware projects and repair information that can help place a measurement in the context of a particular model. Keep a record of the original waveform before modification, especially if the machine is historically significant or if a replacement board such as Nereid-X is installed.

With the console unplugged, recheck every connector and inspect for solder bridges before powering it again. After a repair, repeat the same warm and loaded measurements. A good result is a stable DC rail, controlled peak-to-peak ripple, no unexplained overshoot and no deterioration during drive activity. If the waveform remains abnormal on the primary side or the fault requires live mains diagnosis, hand the unit to an Australian technician who is equipped for mains-powered switch-mode repairs.

Use the scope as a preservation tool rather than just a fault-finding instrument. Document the model, rail values, load, temperature and waveform before and after service, then add those observations to your X68000 maintenance records. That careful evidence helps keep another machine running safely for years, whether it is being restored in Perth, used in Sydney or sitting on a hobby bench in regional Queensland.