Old A500 PSU warning and question about exact voltages

[Mathesar]

Hi,

My A500 has a 2.5A "heavy" powerbrick. Recently the Amiga randomly rebooted for no reason and was generally unstable. Suspecting a power problem I monitored the +5V using a scope and sure enough: The +5V was unstable, wandering to voltages up to 6,5V or so . Sometimes the voltage would completely drop to 0V causing the reset.

Sooo, I opened up to the power supply to find the problem. For those that don't know this particular design; the +5V is generated using a ST L296 switching stepdown DC/DC converter (http://www.st.com/en/power-management/l296.html). The +12V is generated in a similair way but than using an ST L4960 converter. Both voltages are finetuned using a small potentiometer (variable resistor) in the UPPER LEG ( again) of the voltage divider. This is a serious design flaw as this will cause the voltage on the output to go through the roof if the potentiometer fails. And potentiometers will fail at some point. It turned out that the +5V potentiometer in my unit was indeed failing causing the wandering 5V. Luckily the power supply contains a protection system called a crowbar that will short the +5V if the voltage gets above 6,4V or so. This caused the reset in my system and it probably saved my A500 from a certain death.

But now for the weird part; the potentiometer for the +5V can only tune the +5V higher than +5V. In other words, the power supply was tuned out of the factory for a nominal +5.6V or so. This is because the L296 cannot go lower than +5.1V and the potentiometer was tuned to somewhere in the middle. I just removed the potentiometer and replaced it with a short giving me a (stable!) +5.1V. While I was at it I also replaced the +12V potentiometer with a fixed resistor. I needed an 1500 Ohm resistor to give me slightly more than +12V.

So a little warning when your Miggy is unstable, it might be the potentiometers in your PSU. It can kill your Miggy if you are unlucky. Also, did anyone ever measure the actual +5V voltage on their A500 PSU's? It seems too high. Maybe Commodore tuned it this way to compensate for the voltage droop over the cable?

[Daedalus]

Excellent observations, and glad no permanent damage was done!

I've found that most Amiga PSUs give around 5.2-5.3V, and like you I presume that's to compensate for losses on a relatively long cable. It also helps overcome the contact resistance of those 30-year-old power connectors, which is one reason why some people have problems when they use the original PSU cable with a modern PSU - if the PSU outputs an accurate and stable 5.0V, it might be ~4.7V by the time it reaches the chips, even lower when it reaches peripherals like accelerators and RAM boards. That's too marginal and they start seeing crashes on floppy access and other strange symptoms.

[Mathesar]

Originally Posted by Daedalus

Excellent observations, and glad no permanent damage was done!

I've found that most Amiga PSUs give around 5.2-5.3V, and like you I presume that's to compensate for losses on a relatively long cable.

Thanks for the info.
I replaced the short with a 120 Ohm resistor to give me 5.25V. Right in the middle of the range.

[CubeTheory]

Thanks for linking to this. I have one question. Are there any special considerations for the choice of resistor? Or will a standard 0.25w one do? I’ve followed this and based on the existing trim pots, I adjusted to the appropriate level, measured and then selected the appropriate level (which, to be honest came to the same values that you picked ). I’m using standard fixed resistors and they seem fine so far.