the device on the little board I thought I remembered was, in fact, an LM2574. When I looked at it earlier, I didn't take time to put on the "up close" glasses, so I saw what I thought I should see.

On the reverse (solder) side of the board, however, in quite visible silkscreen, it's clearly marked LM2574.

As you may know, I'm not happy with > 5 mV of random noise or of ripple on the Vcc when there's no perturbation, say, from logic device switching. A raw 5-volt supply with more than 0.1% of ripple into a resistive load at 90% of full rated power is certainly of questionable value in cases where that same supply is also used to power gain stages that have gains, of, say, 50-100. Once you add the huge amount of noise created by clocked logic, you'll have 50-100 mV of noise. If all you're running is digital logic, well, that might work out, but if you also have a converter or an amplifier or two, you'll be sorry.

RE