Switching Power Regulators

Richard Prosser says:

[This circuit] is used to drive a 5V(~15mA) relay off a supply ranging from about 20 to 70V.
I can't guarantee all the component values but they are about right. The coil is the relay coil and the current is sensed by R7. If R7 became the load and D1 became a zener at about 5.6V then it should work as a buck converter. Transistor types are BF422/BF423 or BSR19A/BSR20A for the smt version. Similarly the IN4148s are actually BAX12 or BAV99. D2 stabilises the feedback so that the hysterisis is stable with supply voltage. Input voltage is limited mainly by transistor & diode ratings.

Wagner Lipnharski says

This is somehow a crazy idea. Imagine currents, at the secondary 28Vac, primary 117Vac, ratio is about 4.17, take the 0.17 as losses, it means that 10A at the secondary would be 2.5A at the 117V primary. Now, it would be much easier to control the 117Vac @ 2.5A than the 28Vac @ 10A. Suppose you connect a 400V x 8A triac (Q4008L4-ND Digikey $2.00) in series with the primary to the 117Vac. Suppose you connect this triac to keep conducting all the time, using a resistor between Anode and Gate, then you connect the output side of an opto-coupler to switch off (when active) the triac. The LED side of the opto-coupler would be fed by a voltage comparator, like an LM358, the "dirty one". Use an easy 5V zener to feed a trimpot, center point fo to the LM358's (-) input, while the (+) goes to a voltage resistive divider (3V) connected to the Cap + side. Whenever the cap Voltage is higher than the trimpot preset, the LM358 will drive the opto-coupler LED, that will cut off the triac at the transformer's primary. Without supply, the cap will loose Voltage until the resistive goes below the 3V (trimpot preset) and the triac will be free to conduct again, recharging the cap and so on.
                                      VCAP
110Vac-------.   .-------.    .-------o--------o---------o----o Output
Neutral      |   |       |    |       |        |         |
              3 C      ~ |    |+      |        R2 22k    R4 2k2
              3 C     .----------.    | large  |         |
              3 C     |          |   +++ CAP   o--.  .---o
              3 C     |  BRIDGE  |   ---       |  |  |   |
              3 C     '----------'    |    3k3 R3 |  | .---'
              3 C      ~ |    | -     |        |  |  |   A  5VZener
             |   |       |    |       |       Gnd |  |   |
             |   '-------'    '-------o           |  |  Gnd
             o-----.                  |           |  |
             |     |                 GND   +VCAP  |  '------.
             |     |                        |     |         |
            ---    R1 20k ?               .---.   |         R5 10k
    Triac   V/A    |                     /   +|---' V*      |
             |\    |             .------<     |             |
             | '---o----.  Opto  |       \   -|----------->TRIMPOT
             |          |        |        '---'            22k Ohms
             |          V   <-_  V        LM358             |
             |         ---      ---         |               R6 10k
Phase        |          |        |          |               |
117Vac-------o----------'        |          |               |
                                Gnd        Gnd             Gnd
Optocoupler should be a Zero Crossing Triac Output, p/n MOC3040IS-ND Digikey $1.60. The output of the optocoupler should conduct in both sides of the sinewave.
Another opto could be the PS3601LNEC-ND a dual inverted photo diodes ($1.15 at Digikey).
Trimpot Voltage swing from 1.13V to 3.63Vdc, with the circuit above it will regulate VCAP from 8.66 to 27.84 Vdc.
Change R5=22k, TRIMPOT=10k, R6=33k, Trimpot will regulate from 19.6 to 25.52 Vdc.
The zero-crossing opto would give you smooth noise and triac will work cool (I guess), so you probably would need a small heatsink (if any)... pretty different from the linear regulator with transistors, etc...
Problem is; your Voltage regulation would be as fast as 16ms. Changes can be made to change phase at the triac, still regulation at 16ms.
It also requires a fixed resistor at output as a small load to help the regulation.
I don't know how this circuit will behave at power on... perhaps needs some fix to avoid 358 latching up, because voltages at the LM358 would be crazy, so its output. Paul? anything you see? probably a 9V zener in series to the LM358 Vcc to make it works only after Vcap goes above 18V?