A500 Linear PSU question

[CubeTheory]

Hi.

I've just recapped my original A500 Linear PSU using Panasonic FC caps. when i powered it up and measured the voltage with a Car brake light bulb attached to provide load on the 5v rail, it's reading 5.12v - 5.14v depending on which meter i use

This seems too high to me. It remains stable, but i did notice it drop to 5.09v for a while and remain there. When i turned it off and on it returned to 5.14v. It's only done this once in testing.

I adjusted the pot, but it turns out this is only affecting the +12v, which was also a little high, so i trimmed that to 12v exactly.

the -12v is reading 12.23v

Comparing this with the PSU I actually use, based on a Meanwell RPT-60b, with the same load applied, i get 4.85v.

Should i be concerned about this? Im not sure what the voltage limit would be for the A500. This is the original PSU that came with the machine when i bought it new back in '89, so i want to keep it working, but these days its just a backup.

I haven't tested it on the Amiga yet.

Thanks

[hooverphonique]

5.14V (5V+3%) will be fine.
I would be more suspicious of the 4.85V on the Meanwell - if that gets any lower, it might cause issues..

[CubeTheory]

Ok. I noticed an issue with one of the probes i was using, which meant the readings were lower than expected.

so i re-tested both but this time i used a 5w 120ohm power resistor to provide load.

this time I'm getting a stable 5.33v on the original supply!

This seems far to high to me.

Meanwhile, the Meanwell is reading a steady 5.04v

This has now been verified using 2 meters and 2 sets of different probes. the meters are within .02v of each other, and both have new batteries.

[hooverphonique]

The thing is, there will be voltage drop across the pcb inside the Amiga, and the input voltage needs to be high enough so the lowest voltage anywhere on the pcb is enough for the logic to function correctly.

120 ohm is way too high to provide a realistic load (5V/120ohm=42mA) - try something around 5 ohm instead.

[Mathesar]

Originally Posted by CubeTheory

Ok. I noticed an issue with one of the probes i was using, which meant the readings were lower than expected.

so i re-tested both but this time i used a 5w 120ohm power resistor to provide load.

this time I'm getting a stable 5.33v on the original supply!

This seems far to high to me.

Meanwhile, the Meanwell is reading a steady 5.04v

This has now been verified using 2 meters and 2 sets of different probes. the meters are within .02v of each other, and both have new batteries.

5.3V might be just about right so I have been told.
See my adventures with an A500 PSU here: http://eab.abime.net/showthread.php?t=92593

[CubeTheory]

Originally Posted by hooverphonique

120 ohm is way too high to provide a realistic load (5V/120ohm=42mA) - try something around 5 ohm instead.

Thanks. These are the only power resistors i have, never thought to apply ohms law!

[CubeTheory]

Originally Posted by Mathesar

5.3V might be just about right so I have been told.
See my adventures with an A500 PSU here: http://eab.abime.net/showthread.php?t=92593

Thanks for that. for some reason, i never paid attention to the trim pot further in ,the pcb is covered by a metal heat sink and its not easy to see inside.

Further investigation and returning to the car bulb to provide load, gives me 5.04v!! so a faulty probe was at play, and me being an idiot.

What worries me now is the Meanwell. This dropped to 4.8 using the same conditions, which is what i observed previously.

However, the Amiga seems to work fine with the meanwell, running an ACA500plus, trapdoor expansion, and 2 external drives (one is a gotek)

Those readings you were taking on your post, were they under load?

[Mathesar]

Originally Posted by CubeTheory

Thanks for that. for some reason, i never paid attention to the trim pot further in ,the pcb is covered by a metal heat sink and its not easy to see inside.

Further investigation and returning to the car bulb to provide load, gives me 5.04v!! so a faulty probe was at play, and me being an idiot.

What worries me now is the Meanwell. This dropped to 4.8 using the same conditions, which is what i observed previously.

However, the Amiga seems to work fine with the meanwell, running an ACA500plus, trapdoor expansion, and 2 external drives (one is a gotek)

Those readings you were taking on your post, were they under load?

Those were unloaded reading. So, they don't account for cable losses over the cable under typical loading.
However, it seems plausible that an unloaded voltage of 5.2 to 5.3 is about right.

[Daedalus]

Commodore PSUs tended to be slightly high on the 5V line, presumably to counter for losses in the long cable and across the connector. Best bet would be to measure on the board, e.g. at the floppy power connector. The Meanwell sounds fine, but again I'd check at the floppy connector, as perhaps it's losing a little and might be tweaked upwards a little to compensate.

[CubeTheory]

Originally Posted by Mathesar

Those were unloaded reading. So, they don't account for cable losses over the cable under typical loading.
However, it seems plausible that an unloaded voltage of 5.2 to 5.3 is about right.

Thanks

[CubeTheory]

Originally Posted by Daedalus

Commodore PSUs tended to be slightly high on the 5V line, presumably to counter for losses in the long cable and across the connector. Best bet would be to measure on the board, e.g. at the floppy power connector. The Meanwell sounds fine, but again I'd check at the floppy connector, as perhaps it's losing a little and might be tweaked upwards a little to compensate.

Thanks for this. The readings i was taking was from the amiga power plug, so the cable was in play.

However, I'm thinking of shortening it anyway as it's far too long for where its located.