Every Amiga 500 owner will know what this is: The A500 Power Brick.
There are two types of these bricks known as the Light and the Heavy brick. These names refer to the physical weight of the units rather than the output specification.
The ‘Light’ Brick
This one almost feels empty compared to its Heavy sibling. Despite being the lighter of the two it has almost double the output current for the +5V rail with 4.5A vs 2.5A. The output for the +12V and -12V rails are similar.
Light PSU: +5V @ 4.5A, +12V @ 1.0A, -12V @ 0.1A
Heavy PSU: +5V @ 2.5A, +12V @ 1.0A, -12V @ 0.1A
The Light unit is based on a switchmode PSU design which allows for the components to be smaller and lighter. The photos below show the internals of the Light unit. The Amiga power plug Shield connections are connected to the UK 3-Pin plug’s Earth on both the Light and Heavy units.
Switchmode PSUs convert the voltage by chopping the input current on and off very quickly and sending it through a transformer as a series of pulses. These are usually 20KHz to 100KHz depending on the design of the PSU. At the output side of the transformer the current is smoothed out (the best it can) and a regulation circuit monitors it. When the output comes under a
heavier load the regulation circuit will tell the pulsing circuit to make the ‘on’ time of the pulses longer than the ‘off’ times therefore sending more energy through the transformer. The idea is to efficiently keep the output rails at a steady voltage under a range of loads.
The ‘Heavy’ Brick
The Heavy brick uses a hybrid design for the voltage conversion. Unlike the Light brick’s switchmode design this one uses a regular mains transformer, two linear regulators for the +12V and -12V rails, and a switching regulator for the +5V rail.
The transformer is a regular mains 50Hz unit so it is much bigger than the high-frequency unit in the Light brick. This design is not as efficient as the Light brick as the regulation circuit for the +12V and -12V rails burn off excess energy as heat. The +12V regulator is mounted to the big ‘Z’ shaped heatsink, and the -12V regulator just sits in fresh air as a 0.1A load shouldn’t make it too hot.
The +5V rail is rated as 2.5A so a more efficient type of regulation is needed or the unit will get very hot. For this it uses a switching regulator circuit to chop the transformer’s output current into high frequency pulses and adjusts the ‘on’ and ‘off’ times to keep the rail steady at 5V under different loads. The 16 Pin IC does the pulsing and monitoring, and the transistor mounted on the ‘X’ shaped heatsink switches the current on and off with each pulse. There is a smoothing circuit to reduce the switching noise on the output as much as possible.
This particular unit had an annoying audible whine when in operation. Although the unit had been fully re-capped the noise was still present. I tried to reduce this by applying electronics-grade silicone on the large toroid inductor but this didn’t have much effect on the sound level. There was something unusual going on here as the switching regulator should be pulsing outside the audible range.
I took a few scope readings and discovered what was going on. The regulator was switching at around 47KHz but there were regular intervals where unwanted pulses were appearing and it was those bunches of pulses I was hearing. I clamped the scope probe to the large ‘X’ heatsink and grabbed a few waveforms.
After a bit of probing I found that the supply rail for the regulator IC was a bit noisy and had no decoupling capacitor nearby. I experimented with several values and decided on a 22uF / 50V item to mount across the IC’s GND & VCC rails.
The audible noise was now gone and the waveforms were now a lot better.
This successfully fixed the noise problem and the toroid inductor now appears to run cooler than before. The +5V rail is much smoother so the A500 will appreciate that. I closed up the brick and will be keeping an ear close to it in the future 😉