Avalon U5 — Total Positive Rail Failure (Transformer Winding Open)

Symptom

Unit presented with suspected power supply fault. No audio. Voltage measurements on intake confirmed severely abnormal supply rails — positive side completely dead, negative side partially degraded.

Diagnosis

Filter cap survey taken first. C1 at expected −54V (negative rail alive). C2 at only −27V — exactly half expected, indicating half-wave rectification on that leg. C3 and C4, which should hold the +54V raw positive rail, reading essentially zero. Positive supply completely dead.
All four rectifier diodes (D1–D4) tested in-circuit — all showed 0.5V forward, OL reverse. Diodes were good. Replaced as a precaution regardless, powered back up — no change. C3 and C4 still zero. Fault definitively upstream of the rectifier.
Q3 (positive regulator control transistor) measured at all three pins — base, collector, and emitter all sitting in negative territory. Not a faulty transistor — no positive supply to power it. ZD1 (6.2V positive reference zener) initially appeared suspect at 0.74V in-circuit, but recognized as a correct diode-check behavior — a standard DMM can't trigger a 6.2V zener into avalanche. ZD1 cleared. All anomalous downstream readings were symptoms of the missing positive raw rail, not independent faults.
AC measurements moved to the T5 toroidal transformer secondary terminals directly — bypassing connectors and PCB pads. Confirmed: one secondary winding of the T5 open circuit. No AC output on the positive winding. Single root cause for all observed symptoms.

Fix

Removed and replaced the Avalon T5 toroidal transformer with correct replacement unit. Reconnected all secondary wiring with polarity confirmed on all winding taps. Solder joints at all wire termination pads inspected and reflowed.
Post-repair verification: all four filter caps restored to correct voltages (±54V raw rails). All six regulated rails confirmed within spec — +32V, +15V, +6.2V, −32V, −15V, −6.2V. Q3 bias voltages normalized. ZD1 reference measuring +6.2V correctly. Unit returned to service.

What I Took Away

Toroidal transformer winding failures are silent — no burn smell, no visible damage, nothing wrong visually. The transformer looked completely normal while delivering no output on one winding. Only AC measurement at the transformer terminals themselves will confirm or clear the transformer.
Key lesson: follow the AC before chasing the DC. In any power supply fault, confirm AC is present at the rectifier inputs before replacing DC-side components. The four new rectifier diodes were unnecessary — had the transformer been measured first, it would have been identified sooner.
Every anomalous reading in this unit — wrong Q3 voltages, ZD1 confusion, collapsed regulated rails — was a downstream symptom of one dead transformer winding. No regulator components needed replacement.