Thanks for the recommendation.
My application can peak at 1A for a few seconds while performing some tests, but then drops to basically nothing (<10mA), so I need the regulator to be able to tolerate those bursts. It's powered from a 9V battery, so the battery wouldn't even last that long if I drew 1A continuous! As for the capacitors ... how does 470uF sound? That is again another part I buy in bulk, so I just use it for both input and output! They are low ESR tantalums (<400m\$\Omega\$) and this combination has proven so stable and robust that I've used it pretty much in all of my designs so far without a single issue! I'm really leaning towards using a P channel fet.
Edit: Yes , sorry, I meant switching off the negative terminal of the battery, not just the LDO on it's own.
I know I'm off topic and I apologize for that, but wow...
So you have a product that for most of its life uses less than 10mA, but you're using a linear regulator that's eating through a few mA just by functioning on its own. And then, you're using 9v batteries which are known for having low capacity and low output current, risking a reset from your device when you're pulling near 1A.
What happens when the battery discharges down to around 7v and your product pulls 1A for a few seconds? Don't you risk going below 5v with your output voltage considering the 1.3v voltage drop on the regulator itself?
You're pissing away probably half of that 9v battery's energy.
If as much battery life matters, it would make more sense to replace your 9v battery with 2 or 3 AAA or AA batteries and use a step up/boost regulator to get your 5v
It would be more expensive, probably around 1$ in all, but AAA and AA batteries are cheap (and can be rechargeable)... and they can certainly handle the 5w output (5v 1A) for a few seconds without stress.
Here's an example, Richtek RT 4812 .. around 95% at 2.5v..3v->5v conversion and I think <1mA quiescent current :
https://www.digikey.com/product-detail/en/richtek-usa-inc/RT4812GJ8F/1028-1512-1-ND/5640521