Here is my latest brain fart. Basically a stand alone 555 used as a simple switching voltage regulator. The reference is questionable, which is the BE voltage drop of Q1 (0.6V). Anything greater than that will have current flow through the BE junction, causing the transistor to turn on, causing the 555 to turn on, which in turn turns Q2 off.

It might be better if R5 were a zener diode, but I'm going to fry... try this first.

D1 is a feedback component, it will give a quick and dirty visual indication when the 555 is on. D2 shows the power supply as on as well as providing a light load for the regulator.

.

 

Bill,
Save your parts; don't try that circuit.
SD1 is upside-down, and will smoke when Q2 turns on.
There is really no point in the 555 timer, as it's not being used as a timer, just a Schmitt trigger. Since it has a Darlington follower in the output, Q2's gate will be held close to the threshold; R3/D1 make that situation worse. R2 guarantees slow MOSFET turn-on/turn-off times.

 

SD1 is easily fixed, as is R2. I can drop it down to any value, but I was aiming for easy resistors. A 7555 will go the full value, but the current drive is even worse overall, so I guess I could go with a pull up resistor between gate and source.

It is very loosely based on this concept (and I do mean loosely)...

555 design contest Post 14

 

I see. Well, the EDN article referenced in post 14 for driving LEDs is a boost converter, as I'm sure you're aware, where you're apparently trying to make a buck regulator using a P-ch MOSFET.
The circuit that you found interesting that I posted awhile back (Flyback 12v to 6v) is interesting on more than one level; for one because it's using a 555 as an SMPS controller, and the secondary side is completely electrically isolated from the primary side (that is, when the indicated ground is removed). Such a supply could be used to power one of those pesky meters that requires an isolated supply such as a battery - or just about anything that required an isolated supply.