Amiga Technical Resource

A guide to repairing the Amiga A4000 main board audio circuit



This section will cover how to diagnose and repair common faults with the audio circuit, typically caused by leaking surface mount electrolytic capacitors.


PLEASE NOTE that this information is specifically for the A4000D, but theory and repair notes can also be applied to other models.
The exact devices used and component pin functions can change between other models. It should be noted that the A600, A1200, A4000T and CD32 use these same capacitors which have been known to leak electrolyte and cause problems.

These repairs require reasonable electronics knowledge and SMD soldering ability, get help from someone experienced if you are not familiar with these skills.



You can jump directly to the relevant section:



Disclaimer and caution

While the information on this page has been checked and is correct to the best of my knowledge, there is still the possibility of unintentional errors.
Please report any errors directly to this address so they can be corrected.

As with any repair work to delicate electronic equipment, you risk causing further damage to your system or yourself. I cannot be held responsible for any equipment damage or personal injury.

It is strongly advised that you possess some good common sense and preferably have previous experience in working with electronics before undertaking any repair on your system.
As always, you should take all anti-static precautions when working with semiconductor devices.



Theory of operation
Though it may appear complicated at first, it's worthwhile understanding how the audio stage works. Below is the simplified schematic section of the audio circuit.
Click here for an enlarged version of the simplified schematic below.
Audio schematic


Click on the above schematic for an enlarged version with more detail.
This detailed schematic has idle state typical DC voltage measurements in red, typical AC audio measurements in yellow, component pin numbers in blue and additional notes in green.

The large rectangle on the left, U400 (Paula), is the Amiga custom device which generates the analogue audio waveform.
It contains 4 separate 8 bit digital to analogue converters. The output from two of these converters are combined internally to form the left audio channel (output on pin 34), the other two converters form the right channel (output on pin 33).
Paula performs other functions such as handling floppy disk serial data and reading the analogue inputs from the mouse/game ports.

The low level audio is sent to U402, a LF347 quad op-amp on pins 13 (left) and 9 (right).
The audio is amplified to approximately 3 to 8V AC peak to peak on the output pins 14 (left) and 8 (right). Of course the final audio level will depend on the audio data which is being played.

From here the audio is sent through a two pole 4kHz low pass active filter, formed by 10K resistors, 6.8nF capacitors and the other amplifier sections of U402. The filter is software controllable and is used to remove the high frequency alias distortion (steps in the output waveform) which is introduced in the digital to analogue conversion process. Filter on/off is set by a 0/5V logic change on CIA U350 pin 3 (_LED), which is sent to RS232 line driver U304 (a 1488) pin 2 and level shifted to -12V/+12V to become the _CUTOFF signal. This signal then controls the bias on the base input of PNP transistor Q400. -12V on the base via resistor R408 causes Q400 to conduct, thus applying +12V to the gate inputs of N-channel FETs Q430 and Q440. The FETs then conduct, effectively bridging the 10k resistors in the low pass filter circuit, which will disable the low pass filter operation.
When Q400 is turned off (low pass filter enabled), the gates of the two FETs are biased with -12V via resistor R406 to make them turn off.

The second stages of U402 which form the low pass filter do not have any gain, thus the AC audio level here should still be around 3 to 8V peak to peak.

The final output stage is formed by U403, a LM833 dual op-amp.
Just before the output stage, there is an audio combiner formed by networks of 10k resistors. This is to mix CD-ROM audio from connector CN404 into the motherboard audio. To overcome the loss introduced by the combiner, U403 is configured to have some gain, as set by the 4.7k feedback resistors R403C and R403G.

Final audio output on U403 pins 1 (left) and 7 (right) are AC coupled via two 22µF electrolytic capacitors before being sent to the RCA output jacks, CN402 and CN403.

The RCA jacks contain switches which join the independent left and right audio channels together if a single (mono) RCA cable is plugged into the computer.


It is worth noting that this circuit is actually the same design repeated twice. Once for the left channel and again for the right channel. This can make fault finding easier, because the two circuits can be compared against each other for reference.

Also it is very important to note that the entire circuit up to the RCA output jacks, is DC coupled.
This means that a bias voltage problem at the beginning of the audio stage will effect all DC voltage readings for all the following components for that channel.




Leaking SMD electrolytic capacitors

SMD electrolytic capacitor
 A common fault exists with "classic" Amiga motherboards based on surface mount device (SMD) technology.
These are the A600, A1200, CD32, A4000D and A4000T.
The problem is that some of the SMD electrolytic capacitors begin leaking their electrolyte, a mixture of corrosive chemicals, over the motherboard. Aside from causing the capacitor failing to work correctly, the electrolyte attacks solder joints and copper tracks on the circuit boards.

The picture below is a good example of what to look for. Notice how the right hand leg on the 4.7µF capacitor (centre) has become stained and discoloured. The solder joints should normally appear smooth and shiny.

Capacitor corrosion


Also note how the corrosion has spread to the right hand side of the brown capacitor at the top of the picture, and to the pins of op-amp U402.

Common faults include:
The tracing and repairing of these faults are covered in the sections below.




Crackly, distorted or no audio

Most of the time, this problem is directly caused by capacitor corrosion damaging PCB tracks or vias and often the op-amps themselves.

There are two methods for fixing this:
  1. Replace all the suspect components and use a multi meter to test circuit continuity, repairing PCB tracks where necessary.
  2. Using an oscilloscope to find out exactly where the problems are and methodically repairing them.

Personally I'd go with the 2nd method, assuming you have an oscilloscope and know how to use it. This is the repair method I'll be using here.


To begin, we'll confirm the fault and find where the main problem is.

Download this audio test utility and run it on the computer.
The program is very easy to use and generates a 1kHz sine wave tone in both audio channels at the same level.
Audio test program


Connect two RCA cables into the computer's audio output sockets and connect them to a stereo amplifier, monitor, or something you can listen to the output audio on. Be sure to connect both left and right cables to the computer, as there is a circuit which combines both audio channels into one output if only one connector is used.

On a working unit, you should hear a clean, undistorted tone on both left and right channels.
Looking on the RCA jacks CN402 and CN403 with an oscilloscope, a working machine will look like this:
Waveform on RCA output jacks


Note that enabling the 4kHz low pass filter (software controlled) will produce a slightly cleaner waveform as it mostly eliminates the aliasing distortion. The waveform above is with the filter enabled. It should appear the same in both left and right channels.

If the audio output stage is faulty, the waveform will be lower in level and distorted in one or both channels.

It's a good idea to print a copy of this schematic, in colour if possible. It shows typical voltages and levels to expect in the circuit.
Because of component tolerances, there will be a small variation if comparing between two machines, but in the same computer, a good rule of thumb is that the left audio channel measurements should be nearly identical to the same position in the circuit in the right audio channel.



 Fault finding using an oscilloscope
The first thing to measure are the circuit supply voltages on U402 and U403. The table below gives the typical DC voltages on U402 while there is no audio being generated.
U402 in circuit DC voltages
DC measurements on op-amp U402, referenced from ground
Pin 1 +2.3V
Pin 2 +2.3V
Pin 3 +2.3V
Pin 4 +12V (Vcc)
Pin 5 +2.3V
Pin 6 +2.3V
Pin 7 +2.3V
Pin 8 +2.3V
Pin 9 +2.3V
Pin 10 +2.3V
Pin 11 -12V (Vee)
Pin 12 +2.3V
Pin 13 +2.3V
Pin 14 +2.3V

Pay particular attention to pins 4 and 11 (power supply to the device) and pins 10 and 12 (connected to the audio voltage reference, VREF, +2.5V).
If the supply is missing from pins 4 and 11, then investigate this first. The +12V and -12V to the audio circuit come from the computer's power supply, so if these voltages are missing or appear low, check the output from the power supply first.

If VREF on pins 10 and 12 is less than 2V or greater than 3V, there is either a problem with the VREF supply (R401/R402), or more likely, something else interfering with VREF.

Next check the voltage on the first stage outputs (pins 14 and 8). If these are bad, this will effect the rest of the circuit "downstream" of this.
A problem here is probably caused by an open circuit feedback loop (test for 750 ohms between pins 13-14 and between 8-9) or a faulty op-amp. Always measure resistance with the computer powered off.

Now check the voltage on the second stage outputs (pins 1 and 7). If these are bad, voltages on U403 will be effected.
A problem here is possibly in the audio filter section, located between audio stages 1 and 2. With the computer off, measure approx 20k ohms between pins 3-14 and between 5-8.


In many cases, repairs to the audio stage will require lifting U402 and possibly U403 to measure and repair tracks and vias underneath them. It's very common for tracks under U402 to be corroded open circuit. If either U402 or U403 appear to have been corroded, then it's a good idea to replace them if you think they're faulty or not.
If you can't find new LF347 or LM833 op-amps, the more commonly available LM324 and LM358 op-amps will do the job just as well. Be sure to get the SOIC surface mount package and not the DIL leaded package. These devices are available in both types.
Refer to the SMD soldering guide for tips on how to remove and replace a SOIC.
It's likely you'll also have to repair a few open circuit vias. This is easily done using a strand of wire as shown here.

At the same time, you should replace ALL of the 22µF electrolytic capacitors on the board (there are 5 in total), also the 47µF and 4.7µF capacitors in the audio stage at least. There nothing wrong with replacing all SMD electrolytics on the board, it's possibly not necessary and of course more work.
Refer to the SMD soldering guide for tips on how to remove and replace a surface mount electrolytic capacitor.

Once you've replaced ICs, capacitors and repaired tracks, you should start getting measurements like these on the U402 audio signal outputs pins 1 and 7:
U402 output pins, filter off


The image above image is with the audio low pass filter disabled, with audio test program generating 1kHz audio tone.
The below image is the same measurement (on U402 pins 1 and 7) with the audio filter enabled. Note the effect the audio filter has.
U402 output pins, filter on



Also check the in circuit voltage of U403 while the computer is on, but not generating any audio.
The table below shows typical voltages:
U403 in circuit DC voltages
DC measurements on op-amp U403, referenced from ground
Pin 1 -1V
Pin 2 0V
Pin 3 0V
Pin 4 -12V (Vee)
Pin 5 0V
Pin 6 0V
Pin 7 -1V
Pin 8 +12V (Vcc)

First check the device power supply on pins 4 and 8 are OK. As mentioned earlier, the +12V and -12V to the audio circuit come from the computer's power supply, so if these voltages are missing or appear low, check the output from the power supply first.
The AC output of the device swings both negative and positive around a centre voltage of approximately -1V. Ideally the 22µF output capacitors (C443, C433) should be bi-polar (non-polarised) types for this reason. Many argue that this is the primary reason for these capacitors failing in the first place.
I'm not convinced of this, as there are other 22µF capacitors in the machine which are correctly DC biased at all times and they still leak as the ones in the audio stage do.

After you've checked and repaired the circuit around U403, power up the computer and run the audio test program again. The waveform on the output pins of U403 (1 and 7) should look like this with the low pass filter enabled:
>U403 output pins, filter on




Still can't get it right?
If you've been right through the above guide, but it still doesn't seem to be right, here's a few things to check and try:

And if you're still stuck, Email me.



Repair Service
I also offer a comprehensive repair service down to SMD component level on bare motherboards for customers from both in New Zealand and internationally.
Repair times are typically 1-2 weeks (not including time in transit). Items are returned with a full product test sheet.
Contact me via Email for an estimate and current repair timeframes.








Back