[sdiy] 78xx regulator gives negative output half the time

[JH.]

>JH also suggested a diode to ground just after the regulator for prohibiting reverse-current and this seems indeed to work and the 
>circuit no >more latches up.
>
>However i'm still not entirely sure i understand why this is happening in the first place, JH suggested he's seen it while doing 
>circuits with unevens >loads - but i've built literally 100s of circuits using this very same PSU scheme - that definitely have 
>uneven loads - and i've never had this >happen before.

Well, it's not *just* uneven loads. Re-read the part about charge that keeps stored in caps from previous powering of your circuit. 
Only under certain circumstances you will have a charge distribution in your circuit that leads to a temporary negative voltage on a 
rail that should be positive - but *sometimes* this occurs, in *some* circuits - and when it does, even for milliseconds, your 78xx 
latches, if there is no diode.

Unless you have a circuit where the latchup occurs *every* time (which I understand you haven't), it only occurs when the circuit 
had been powered up and switched off *recently*, doesn't it? - If you leave it alone, switched off for half hour, you won't run into 
latch up. But that's nothing you can rely on in every-day use, so just use these diodes by default. Or use LM317, which seems less 
sensitive to that kind of stuff.

JH.