Could this work as a simple regulated psu?

[little_carlos]

Hello guys, today i was experimenting with the lm358, and i tried using it as a buffer for a potentiometer, so it drives a transistor to regulate voltage, its not very accurate, but i used it with some loads like a motor, some lds, a fan, and it seems to be stable, at least for small loads, so the gold question here, could it really work as a regulated psu? if it does, is there a way to make it better? more accurate? thanks for your time, here is the schematic

[Yansi]

Lower the value of the base resitor to about a tenth, connect the inverting input of the opamp on output (that means the emitter of the NPN) an it will work exactly as a voltage regulator. In its simpliest form with all problems associated. (Lack of compensation, current limiting, potentiometer fault protection, etc.)

*the pot input should be connected to a voltage reference too.

[T3sl4co1l]

A true regulator exhibits relative independence from both input and output changes.

I'd call this more of a controller, or something like that.  At least when paired with the motor example.  A controller you might not care about what it's relative to (i.e., some fraction of the input voltage), as long as it's adjustable (the user control itself being negative feedback).

Tim

[Wim_L]

No, it wouldn't work very well as a regulated PSU. The output would be the voltage set by the potentiometer RV1 (minus transistor BE voltage drop), but this voltage would vary with the unregulated input voltage. So the output would vary too.

[JacquesBBB]

What you have is a power potentiometer. To get a regulated output, you need to add a voltage reference. The minimal could be a zener diode.

[Yansi]

Where had the schematic gone? The image isn't showing up on my side. I just wanted to download it and make notes in it to help him make that regulated 

[little_carlos]

Where had the schematic gone? The image isn't showing up on my side. I just wanted to download it and make notes in it to help him make that regulated 

can you see it now?

[Yansi]

Might be a problem in my Opera browser. Firefox and IE both displays it normally. 

[Zero999]

You didn't say what the supply voltage was.

Just use a zener reference and take the feedback from the emitter.

You could also use a lower voltage zener and make the gain higher.

[Yansi]

That way.

C2 for reference noise reduction (10-100uF)
R2 for stability (isolates opamps output from capacitive load), somthing round 30-100 ohms
R1 should supply enough current to the zener to work in suitable area of its VA characteristic. Recommend using 5V6 zener with a diode in series to compensate for temperature drift
R3, C4 compensation cell. Depends on actual circuit. I'd expect R3 in kohms, C4 tens to hundreds puffs. Maybe someone more experienced will help with this.
Q1 should not use TIP31, bcs of very low hFE. Use darlington instead. Like TIP122.

Note that this circuit configuration provides output voltage from zero to Vref(zener) maximum. For higher output voltage, the amplifier should be connected as noninverting amplifier and the placement of compensation capacitor C4 probably changed.


More advanced stuff: For the regulator to be more idiotensicher-bombenfest, overcurrent protection should be added (small signal NPN, current shunt and one other resistor at least) and protection diodes should be connected. One anti-pralellel over output, second  from output to the input and third between emitter and base of the darlington. In that configuration, R2 should have around 1k. The power darlington then must be with high beta.

Even more stuff to make it better: If the wiper of the VR1 loses contant with the resistance track, the opamps imput will be flying, picking noise, or just shoots to the sky by the bias current. That makes the output of the regulator go at maximum voltage.  An easy fix can be done to that: Add about 100k resistor between the noninverting input and ground. Adding a small cap accros it as a bonus will help to smooth out short dropouts of the wiper when turning the (more or less) crusty pot.

[Zero999]

The problem with this kind of circuit is the drop-out voltage tends to be high as you've got the losses in the LM358's output stage, which is a Darlington pair on the high side, plus that of the emitter follower.

To reduce the drop-out voltage, you could add a pull-up resistor from the LM358's output to +Vin, to help it supply the current to the Q1's base when the output voltage is high. The resistance value can't be too low, otherwise it will limit the minimum output voltage. The output voltage should be able to go down to 0V, as long as the LM358's output can go within V_BE of the negative rail. If a Darlington pair is used for Q1 then the LM358's output can  sink be up to 10mA before this will happen.

[TimFox]

Technically, your original circuit is a "regulator", since the potentiometer will change the output voltage.
The correct term for a circuit that maintains the output relatively constant against changes in the input voltage and output current is a "stabilizer", where a stabilized power supply uses negative feedback to control the output against a reference.