Linear regulator- Topologies-pass transistor-sensing current

[atferrari]

Three basic questions about linear regulators (DC power supply):

1 - The op amp controlling the pass transistor gets at the inputs, the reference voltage (user control) and a sample of the output voltage:

a) In a closed loop, it controls, in a continuous way, the pass transistor driving voltage. From all what I've read so far, a quite common topology.

b) As a comparator in an open loop , it controls, in an on/off way, the pass transistor driving voltage. Quite uncommon, at least for me. I run across such a topology in an article from ELEKTOR (Spanish version - No. 260 and/or 261, if i recall properly).

Question 1: The first seems more appropriate for a smooth more ripple free control of the output. Am I right? What are the advantages of each topology?
-------------------------------------------------------------------------------------------

2 - Question 2: is it imperative for the positive rail pass transistor to be an NPN type (and PNP for the negative rail)?

Before posting I made a quick search and I couldn't find examples of the opposite but I think I saw some somewhere.

If both types may be used, which should I prefer and why?
--------------------------------------------------------------------------------------------------------------------------------------------------

3 - Sensing the current may be done at four places:

a) between the rectifier and the pass transistor.
b) between the pass transistor and the load
c) between the "common" (output) and the "common" of the regulator circuit itself. ELEKTOR issue above, again.
d) in the ground "rail" which I've already discarded for a dual supply.

I've implemented in the past, a) and b).

Raw voltage in a) could demands high voltage components but voltage variations are much more limited, while b) gives chances to be in more comfortable voltages.

Intrinsecally by design, in c) the load is always riding some mV above the regulator's common due to the sensing resistor and affects the actual voltage feedback to the opamp in charge..

Question 3: which is best? Besides the drawbacks I've mentioned, is there anything else to account for?
----------------------------------------------------------------------------------------------------------------

Sorry for the long posting. I would like to reinvent my wheel at least once!

Gracias for enlightening replies.

Agustín Tomás

 

[bmcculla]

1) I think the off-on (b) topology works on similar priciples to PWM; by switching the transistor it has much lower resistance so it dissipates less power. The down side is more ripple.

2) The NPN for positive regulation has the source at a regulated voltage which would make the transistor turn-on independant of the input voltage. If you used a PNP variations in input voltage would cause changes the transistor drain current making for a more complicated control problem.

3) My first instinct is that b) would be best because you can eliminate the effect of the resistor with your feedback control. The only problem I can think of is that this adds an extra RC constant to your control loop which might make regulation difficult at high frequencies.

Good luck with your new wheel
Brent

 

[audioguru]

Hola Agustin Thomas, que tal estas. My wifey is from Spain but that's about all the Spanish I speak.
1a) All linear IC voltage regulators have a high-gain opamp driving the series-pass transistor with a voltage reference and a sample of the output voltage as negative feedback. Most have a fixed output voltage but a few, such as the LM317 have a user voltage adjustment. The benefit is very low ripple.
1b) Comparator voltage regulators have high ripple because they sense that the output voltage is low and turn-on the pass transistor to saturation. When the very-high-value output capacitor is charged to a high enough voltage, the comparator switches-off the pass transistor. The benefit is efficiency, the pass transistor is switching and doesn't consume much power.

2) Low dropout positive voltage regulators use a PNP pass transistor. Unlike an NPN emitter-follower pass transistor that doesn't have gain, the PNP transistor has gain and phase-shift causing the regulator to require a fairly high-value output capacitor for stability (to avoid oscillation). I have attached a schematic of an LM2931 low-dropout positive regulator.

3) The current sensing can be anywhere. I have seen voltage regulators that sense the load current in the common return to the power source, and the sense resistor is outside the voltage negative feedback loop so the voltage regulation isn't affected by it.
Adios.