Fender '65 Deluxe Reverb Reissue — Full Recap + Robinette Main Filter Mod

Symptom

Came in with a farting, blubbery tone on the low notes — classic sign of an amp that can't hold its supply voltage when you dig in. Then it got worse: after sitting on the shelf a week waiting for parts, it powered up into a loud hum at full volume with nothing plugged in. The noise only appeared with the amp driving a load, which meant it was coming through the output stage — the amp was amplifying something, and that something was its own power supply ripple.

Diagnosis

Ruled out the cheap stuff first. Output tubes tested okay, but on a high-hour rental a fresh matched pair of 6V6s went in as insurance — tired outputs are a classic cause of the farting symptom. Rectifier tested fine, replaced anyway on the same logic. Preamp tubes all good, with one bit of housekeeping: the weakest 12AX7 moved to the tremolo oscillator socket, where it never touches the guitar signal. With tubes eliminated, the filter caps went on the meter. The stock reissue reservoir is two mismatched caps in series — a 47µF/500V and a 220µF/100V — and the readings told the whole story: the 47µF measured just 2µF, effectively open, while the 220µF had drifted down to 190µF. That's not a coincidence. In a mismatched series pair the smaller cap takes the bigger share of the voltage, so the 47µF spent twenty-five years doing most of the work — and it's the one that died. The week on the shelf finished it off: an open reservoir turns the B+ into raw pulsing DC, and the amp cranks its own hum out the speaker, no input required. With the diagnosis made, I didn't power a sick amp for further probing — no reason to stress the output tubes or the transformer. Straight to surgery.

Fix

Full recap with the FCK-23A kit from amprepairparts.com — main filters, three decoupling filters, all seven cathode bypass caps, footswitch filter, and both bias supply caps. The headline is the main filter mod, based on work by Rob Robinette, which fixes the exact failure mode this amp just demonstrated: the mismatched series pair goes out, replaced by two matched 70µF/350V caps in series with a 220K 1W balancing resistor across each, so the voltage divides exactly in half — no more one cap doing all the work. 700V of combined rating, and the resistors double as bleeders after power-off. One cap at a time, polarity checked against the board markings before every joint, flux and wick on these fragile Fender traces, and a bead of RTV under the dog-house caps so speaker vibration can't fatigue the leads. Brought up slow on the variac to form the new electrolytics, then full voltage — dead quiet. Plate voltage at 410V, bias set to 21/22mA per side on the 6V6 pair. Burn-in, bias rechecked, stable. Back to the rental fleet.

What I Took Away

Noise that only appears under load is coming through the output stage — the amp is amplifying it, which points at supply ripple, not mechanical buzz or grounding. Failing electrolytics don't fail on a schedule, they fail on a curve; this one went from 47µF to 2µF and fell off the end of it sitting on a shelf. The failure pattern matched the design flaw exactly: in a mismatched series pair the smaller cap carries most of the voltage, and that's the cap that died while its low-voltage partner only drifted. Never replace one cap from a failed set — the rest are the same age from the same factory with the same miles. And when the diagnosis is already made, don't keep powering a sick amp to admire the fault: every second of that hum was stressing the output tubes and transformer. The full job is documented on video — this was Filament video 001.