Amiga 600 Audio Output Repair

The audio stages on Commodore's A600 schematic, sheet 5: right channel above, left below. Each passes through two sections of U15, a 22 µF coupling capacitor and a 1 kΩ resistor to its RCA jack.

On the Amiga 600 the sound from Paula (U3) passes through U15, a quad op-amp listed as an LF347 or TL084, to the RCA jacks CN3 (right) and CN4 (left). Paula's analogue side has its own supply through R309 and its own ground, joined to the rest of the board through the ferrite bead E301, and U15 is biased from a reference voltage, AVREF. Repairers have traced A600 sound faults to leakage damage in the op-amp circuit, failed op-amps, a broken E301, a corroded via in Paula's supply and corroded jacks.

Symptoms

A600 audio faults reported by repairers

Symptom	What was found	Source
One channel quieter and distorted; after a recap the sound became loud and glitchy and the screen corrupted while sound played	The LF347 had failed. AVREF read 4.2 V; it should be about 2.3-2.5 V. A new LF347 cured the sound and the corruption, and AVREF read 2.4 V.	Pure Amiga
Sound totally distorted on one channel	The amplifier feedback loop on the right channel destroyed by capacitor leakage	robthenerd
Sound from one side only	The op-amp	RETRO is the New Black, via The Oasis BBS
Volume wandering and the right mouse button acting as if held down	The ferrite bead E301 had snapped in two, leaving the analogue ground with no reference to digital ground	LinuxJedi
Kickstart colours stop at a white screen; Caps Lock stops responding; DiagROM tests crash	Corrosion round C306 had broken a via near R309, cutting the supply to Paula	Pure Amiga
Corroded audio jacks	Found corroded inside when removed to reach two capacitors; replaced	LinuxJedi

How the circuit works

Paula's audio supply on sheet 6: VCC through R309 (1 Ω), AVREF from R303 and R304 with C305 and C306, and the ferrite bead E301 joining Paula's analogue ground to the audio ground.

Commodore's schematic shows these stages (right channel designators first, left in brackets):

Stage	Parts
First op-amp section	Paula's output into U15 pin 9 (pin 13); pin 10 (pin 12) at AVREF; feedback R331 (R321), 360 Ω, with C331 (C321), 0.1 µF; output pin 8 (pin 14). The right channel's input also takes the NOISE and AUDIN lines through R329 and R339 (10 kΩ).
Filter section	R332 and R333 (R322 and R323), 10 kΩ, with C332 (C322), 6800 pF, and C333 (C323), 3900 pF, round U15 pins 5-7 (pins 1-3). The JFET Q331 (Q321), an MPF102, switches the filter; its gate is driven through R336 (R326), 470 kΩ, from Q341.
Output	From pin 7 (pin 1) through C334 (C324), 22 µF, R334 (R324), 1 kΩ, and E331 (E321) to the jack CN3 (CN4). R335 (R325), 390 Ω, loads the output.
Mono mix	R337 and R327 (10 kΩ) mix both channels to the MONO line, which feeds the encoder U12 for the RF modulator.

On sheet 6, Paula's supply comes from VCC through R309 (1 Ω), decoupled by C307 (470 µF) and C3 (0.33 µF). R303 and R304 (1 kΩ each) divide that supply to make AVREF, smoothed by C305 (0.33 µF) and C306 (10 µF). E301 joins Paula's analogue ground to the audio ground; the schematic notes that the grounds interconnect near the audio jacks. Pure Amiga describes U15 as running from +12 V and −12 V, with outputs on pins 1 and 2 and on pins 6 and 7.

Paula (U3) on a Rev. 2D board with E301, R309 (marked 1R0), C305 and C306 above it.

Audio parts (Commodore parts list)

Ref	Part	Commodore part no.
U15	LF347 or TL084ACD, surface mount	391103-01
Q321, Q331	JFET MPF102 or PN4302, SOT-23	391145-01
C324, C334	22 µF 25 V aluminium electrolytic, surface mount (case D)	391097-08
C306	10 µF 25 V aluminium electrolytic, surface mount (case C)	391097-07
C307	470 µF 16 V aluminium electrolytic	390101-03
R309	1 Ω, 1 W, 2512	391093-01
E301	Ferrite bead, large (1812)	391092-01
CN3, CN4	RCA jacks, right and left audio	252122-02, 252122-01

Tools and parts

• Multimeter; an oscilloscope helps to follow a signal through the stages.
• Amiga Test Kit or DiagROM to play test sounds. Pure Amiga used Amiga Test Kit.
• A replacement LF347 (or TL084) for U15.
• A 1812-size ferrite bead for E301. LinuxJedi fitted an 0805 bead as a temporary fix until the right size arrived.
• Fine wire and solder mask for via repairs, as Pure Amiga used.
• Replacement RCA jacks if the originals are corroded.

Procedure

First checks

1. Take the main board out; see the Amiga 600 Maintenance Guide. Look for leakage round C324 and C334 by the jacks, round C306, and round U15.
2. Recap the board if it has not been done; see the Amiga 600 Capacitor Replacement Guide. On LinuxJedi's board the audio jacks had to come out to reach two of the capacitors, and they were corroded inside, so he replaced them.
3. With the power off, measure the resistance between digital ground and analogue ground. LinuxJedi says it should be almost nothing; on his board it was nearly infinite because E301 had broken.
4. With the power on, measure AVREF across C306. Pure Amiga expects about 2.3-2.5 V.

One channel quiet, distorted or silent

U15, the audio op-amp, on a Rev. 2D board, with the filter JFETs Q321 and Q331 below it.

1. Play a test sound and follow it through the table above, comparing the bad channel with the good one at the same point.
2. Check the feedback parts of the first stage, R331 and C331 for the right channel or R321 and C321 for the left, and the tracks between them. robthenerd's machine had lost the right channel's feedback loop to leakage.
3. If AVREF is high, or the channel stays bad with the passive parts intact, replace U15. Pure Amiga removed the LF347 first and found that the screen corruption stopped even though there was then no sound; a new LF347 restored the sound and brought AVREF back to 2.4 V.
4. Before fitting a new Paula for an audio fault, make sure nothing is feeding a high voltage into it. A reader of the Pure Amiga article lost a replacement Paula because something on the board was connecting +12 V to Paula's right audio output. Removing the resistors and capacitors with a path to the op-amp one at a time and cleaning everything made the problem go away, though no single failed part was found.

Volume wandering or right mouse button stuck

1. Check E301 as above. LinuxJedi's bead was snapped in two and fell apart when desoldered.
2. Fit a new bead. With E301 replaced, LinuxJedi's board passed Amiga Test Kit with the mouse buttons working normally.

Machine stops at a white screen

1. Check continuity from R309 to Paula's supply pin. On Pure Amiga's board the via near R309 had corroded through, next to a badly leaked C306.
2. Clean the via, bridge the break with wire on either side, and cover the repair with solder mask.
3. Replace C306 and the other leaked capacitors.

Reassemble

1. Refit any replaced jacks, clean the board, and refit it in the case.

Testing

1. Measure AVREF again: Pure Amiga's repaired board read 2.4 V.
2. Play sound on each channel in turn with Amiga Test Kit and check both jacks.
3. Watch the screen while sound plays. On Pure Amiga's board the failed op-amp corrupted the display only when sound was playing.
4. Check that the right mouse button works in Amiga Test Kit; LinuxJedi's broken E301 showed up there first.

Board revisions

Commodore's schematic set is schematic 315987 Rev. C. LinuxJedi's board with the broken E301 was a Rev. 1.3. The board photographs show a Rev. 2D board, where E301, R309, C305, C306 and U15 carry the same designators. The Amiga 500 Audio Output Repair covers the same op-amp and filter design on the A500.

Related pages

• Amiga 600
• Amiga 600 Troubleshooting Guide
• Amiga 600 Capacitor Replacement Guide
• Amiga 600 Reset Circuit Repair
• Amiga 600 Video Output Repair
• Amiga 500 Audio Output Repair