A long time ago I built a variable power supply, mostly from scavenged parts. Voltage regulation was adjustable up to about 30V through discreet transistors (a design from a book) and current was limited to about 0.6A.

I’m in the process of upgrading it to supply from a few volts up to about 40V with adjustable current limiting from 0.1A to a max. of about 4A.

The primary will supply (as measured with no load) 34V at about 6A continuous (200W).

The intermediate is a boost regulator. I’ve removed the onboard pot and substituted it with a potentiometer with series/parallel resistors to enable the Vintermediate output to be adjustable from a Min. of 37V (must be > Vprimary) to a max of 50V. The pot. will be ganged so that when Voutput is lowered, Vintermediate is also lowered to reduce dissipation in the last stage.

The last stage is a TL783 adjustable regulator. It will have bypass transistors to increase the output current.

Most of the designing/assembly is done. The final hurdle is implementing the adjustable current limiting. To date I’ve figured that a current sense resistor driving a non-inverting amplifier will give me the signal that I want to drive a transistor which bypasses the R2 part of the TL783’s adjust pin. If the transistor can be turned on in a controlled way, Voutput will be lowered until constant current output is reached.

I’m new to amplifiers so at first glance LM741 seems a good choice.

Q1. Where to source V for V+ and V- ? The primary will also feed a 12V(?) regulator to drive fans for cooling – could this 12V source be used for V+ and V- ?

See block diagram for clarification.

 

Welcome to AAC!

LM741 is over 50 years old and is primarily used as a model op amp in text books. You can use it only if you have a good reason for doing so.

LM358 is a popular dual op amp.

For supply voltages, you need a dual voltage power supply, one that can supply both positive and negative voltages. Check the datasheet of the op amp for minimum and maximum supply voltages.

If your circuit only needs positive voltages to operate, you can eliminate the negative supply voltage and use GND instead, what we call single supply operation. There are some design issues to be tackled.

 

Thank you. More research to do!

 

@MrChips I'm sold on the LM358 - thank you.
Found this link (https://web.mit.edu/6.101/www/reference/op_amps_everyone.pdf) in another thread which is GOLD.
Before I can solve the configuration/R values for the op amp, I need to solve for; how to operate the ADJ bypass transistor in analogue mode (not saturated - also detailed in the link above). Still some research to do.

 

You can get devices (INA181 for instance) which are designed specially for the job and will replace the op-amp and all four resistors.
You should research the phase response of your circuit. Otherwise you will find that you have built an oscillator.

 

I keep "re-inventing the wheel". I think, "use a differential amplifier and feed the output to the non-inverting input on a dual (like the the LM358)". Purpose made devices exist but at >$10 each, I'm happy to add some resistors to a LM358 (~$2 each)
What is the maximum voltage you can apply to the input pins on a LM358?
I should mention, the completed device is a one-off (not for commercial manufacture).

 

I’m in the process of upgrading it to supply from a few volts up to about 40V with adjustable current limiting from 0.1A to a max. of about 4A.

the point is -- how you prevent unwanted oscillations . . . up to voltage regulator or current limiter distortion/malfunction . . .

741 has 2Mohm input impedance but has no RRIO nor significant slew rate - likely drifts with T° the Vio limits it's sensitivity
▲
the output zero balance is usually dependent on Vs but may be dependent on inputs common mode voltage level also = if your Lab supply voltage setting is high your current limit has one systematic error & if it's low it has another (!maybe!)

however the BJT has a miller capacitance & storage time . . . with the phase delay from Op Amp it may destabilize the output dependent from inductance and capacitance of the final load and wiring

the → Rf : ( R₊ ( = R₋ – 5mΩ ) ) = R₋ : Rꜰ ← sets your dif gain with respect to GND -- IF the regulated Lab sup. OUTP is below 6V then you likey want to have your Op Amp Vcc above that and your Vee below the GND (the least you want to have anyway to be able to drive your BJT) --or-- you may further compromize your feedback stability by adding the E follower to your 741 OUTP and using it as zero (GND) going output . . .

 

LM741 is over 50 years old and is primarily used as a model op amp in text books. You can use it only if you have a good reason for doing so.

LM358 is a popular dual op amp.

The LM358 is 54 years old. And it has higher input stage noise and much worse crossover distortion. Both of these limitations come from design choices. To reduce the static operating current to what was then an almost record low value, the input stage static current is low (higher noise) and there is almost zero static current through the output stage (crossover distortion).

OTOH the input stage has PNP transistors, which is why its common mode voltage range can extend *below* its negative rail. When the chip is powered by a single 5 V rail, both its input and output are TTL compatible. That was a very big deal.

Its big advantage in power supply applications is that input common mode voltage range. It can form an adjustable regulator that goes down to almost exactly 0.00 V without requiring a negative supply rail.

ak