Power supply | need help deciphering the schematic

[Pixelmarth]

I need some help deciphering the circuit
I know that P1 controls the voltage and P2 controls the current, but I can't figure out how the opamps work.
My instructor wants me to build it, but I want to understand it first.

A universal power supply unit

So that we don't always have to rely on batteries for experiments and testing our circuits, we should get a universal power supply unit for our workshop. As true hobbyists, we will of course build it ourselves. As a blueprint for this, I offer my laboratory power supply, which I built more than 30 years ago and which still works today.

The terms voltage source and current source are often used in the same context. Strictly speaking, however, they are two different things. A voltage source provides a stable output voltage that does not change when the load at the output changes. A current source supplies a constant output current and regulates the voltage required for this. For our laboratory power supply, it is desirable if it combines both properties. It supplies a constant voltage to operate our circuits. However, if a certain maximum current is reached, it no longer allows the current to increase further. In this way, we can prevent the components from being destroyed by excessive current in the event of errors in the test circuit. In addition, our power supply unit is not affected by a short circuit.

At the time, I opted for a power supply unit that delivers two independent voltages of 1-30V at a maximum current of 1A. The circuit described is therefore completely duplicated.

A few comments on the circuit. The power transformer with an M102a core is self-wound. The choice of active components is actually quite uncritical as long as the limit values for voltage, current and power dissipation are adhered to. A Graetz bridge 3PM1 was used as the rectifier. Of course, you can also use 4 individual rectifier diodes. This is followed by a large charging capacitor. When selecting it, remember that it charges with the peak voltage of the AC voltage before rectification. At 40V effective AC voltage, this is about 56.6V! The Z-diode SZX21/1 at the input of the first operational amplifier supplies a reference voltage of approx. 0.75V. As Z diodes cannot be manufactured for such low voltages, this is actually a normal forward biased silicon diode. I have simply used the base-emitter diode of a miniplast transistor for this.

A rotary switch can be used to preset the desired maximum current in 5 stages (10mA, 30mA, 100mA, 300mA, 1A). The fine adjustment is made with the potentiometer P2. The output voltage can be set with P1. At the time, I did not use precisely calibrated scales and always set the desired values beforehand using a measuring device. Of course, you could also integrate suitable measuring instruments in the housing. The two power transistors (SU 169 and SD 337) are screwed onto aluminum heat sinks for heat dissipation. Of course, sufficient ventilation must also be ensured in the housing. The resistors on the output switch must be designed for the power losses that occur there, especially in the high current ranges.



Here is the Link to the German Website: https://technik.reicke.de/bauanl1.php

Also, if anyone could explain to me how you go about deciphering schematics that you might not quite understand at first, I'd be grateful!

[wasedadoc]

The first opamp sets the voltage at the emitter of the first transistor at twice the first zener voltage.  The second opamp controls the output transistors (a compound emitter follower type thing) to make the output voltage equal twice the voltage at the slider of P1. The transistor most further to the right senses a controllable (by P2) fraction of the voltage dropped by the output current through the switchable resistors.  When the voltage seen by the base-emitter junction of that transistor gets to about 0.6 Volts the transistor begins to conduct and its collector current steals the drive to the output trio.

[Kim Christensen]

I need some help deciphering the circuit
I know that P1 controls the voltage and P2 controls the current, but I can't figure out how the opamps work.

First off, those Op-Amps (TAA761?) are a bit unusual in that they have an open collector output instead of a push pull type.

What puzzles me is why they use a 1V zener. (SZX21/1) Especially since this means that the reference voltage out of the 1st OpAmp will be around 2V. The 3.3K resistor in series with the POT P1 means 1.5V is the max possible voltage into the non-inverting (+) input of the 2nd OpAmp.
But then, they use two equal value (12K) feedback resistors for the 2nd OpAmp which means that the output voltage will not go higher than 4V.

[Pixelmarth]

well I've also tried simulating the circuit in falstad if you'd like to have a look:
https://tinyurl.com/2behxkwa

maybe you can help me figure things out that way!

[wasedadoc]

What puzzles me is why they use a 1V zener. (SZX21/1)

I think that is an error. Meant to be a 15 Volt one same as the other two but the trailing '5' has been destroyed by the circuit line?

[pqass]

well I've also tried simulating the circuit in falstad if you'd like to have a look:
https://tinyurl.com/2behxkwa

maybe you can help me figure things out that way!

Your choice of zener is 5.6V (the default) which results in the first op amp power rail being only 10.96V. Therefore, it only allows for a max. pot wiper voltage of 8.2V.  This limits the PS output voltage to at most double that.   But you've also used the same default 5.6V zener (x2) for the second op amp power rail!  You should have created a new model with 15V breakdown and applied to both stacked zeners.

To me, the first opamp is over complicated.  It can be removed entirely for a 15V zener to ground (and removal of the 3.3k).  See this simulation.

You should also use a similar 4k7+2x15Vzener+NPN to power the right op amp. That way, the zeners don't have to dissipate much power.  Actually, the 2.4k may be too high a value since, with 2x15V zeners, limits the current to 3.5mA (of which up to 2mA will be used to power the op amp). Better to put any power dissipation through the NPN and not the zeners.

And, not sure why the need for a "triple" darlington arrangement. It just adds another V_BE drop.  Two should suffice to give >1000 gain.

[Kim Christensen]

What puzzles me is why they use a 1V zener. (SZX21/1)

I think that is an error. Meant to be a 15 Volt one same as the other two but the trailing '5' has been destroyed by the circuit line?

That would work much better. It looks like the OP translated and pasted the text from a German site. But the translation text implies that they really meant it to be a very low voltage zener which, as you say, makes no sense:

The Z-diode SZX21/1 at the input of the first operational amplifier supplies a reference voltage of approx. 0.75V. As Z diodes cannot be manufactured for such low voltages, this is actually a normal forward biased silicon diode. I have simply used the base-emitter diode of a miniplast transistor for this.

well I've also tried simulating the circuit in falstad if you'd like to have a look:
https://tinyurl.com/2behxkwa
maybe you can help me figure things out that way!

The problem with the falstad sim is that your only choice for the OpAmp is a LM741 or a LM324. Neither of those would work properly in this circuit since their outputs cannot go higher than their supply rails.
In the real world, a LM741 or LM324 won't work properly in this circuit either.
ie: Your max output voltage would be limited by the max output voltage of the 2nd OpAmp minus the 3 Vbe drops (-2.1V) of the output transistors

[Pixelmarth]

Thank you for your response,

so to be honest the circuit I'm meant to use seems very dated, so I'm asking myself it's just simpler to use a voltage regulator 
also I still cant get 1A output which bothers me as well.
Isn't there a simpler "up to date" power supply I could build on my own? I'm only half a year into my apprenticeship, so this is all new to me !

[Pixelmarth]

I've been watching IMSAI guys video on this 0-30V 2mA-3A power supply which was very informative but that schematic is quite difficult for me to understand, despite his explanation!

[MarkT]

And, not sure why the need for a "triple" darlington arrangement. It just adds another V_BE drop.  Two should suffice to give >1000 gain.

Its to reduce headroom when the load is heavy, there's only a 5k1 resistor to drive the triple, more gain means less voltage across that.  That resistor has to be large enough not to fry when the current limiting kicks in at full voltage.

BTW the first stage is a boosted shunt regulator that bootstraps the zener current from the output voltage to greatly reduce its sensitivity to raw supply voltage changes.

[pqass]

See this article to understand how the CV and CC loops work in discrete component power supplies. 

Attached is a generalized schematic for a CV/CC power supply similar to the HP E3610.  It is slightly different to that presented in the article above. Notice that the op amp power rails and +12V is referenced to the +Vout (red) post (the GND symbol).  You might say this is a negative power supply; red post=0v black post=0..-20V.   The op amp rails are usually derived from an extra (floating) transformer winding + regulator (note: not sure why op amps in the simulation need a -1V supply since output never goes to 0V. Nevermind. It's because of the Vf of the LEDs. If you used regular diodes (Vf=.7) op amp lower rail could be 0V).  This arrangement allows the sharing of the same Vref to set current and voltage independently.   

Play with the simulation here.  Notice that the voltage is constant unless the load changes (drawing > Cset) or the Cset is lowered (< current load draw), in which case, the voltage is lowered until the Cset is met.

[Wallace Gasiewicz]

Lab PS that I have seen have TWO sources of voltage. Actually two separate secondary transformers windings and rectifying circuits.  One is for the OP Amps and the Reference Volt. It is lower voltage and is regulated to within within the specs for the op amps (or the regulator IC that also supplies a ref voltage). The other is the higher voltage main for the output. This setup provides lower voltage, more consistent supply for the op amps. Most op amps cannot use a 40 volt volt supply.  This takes a lot more thinking than using one transformer.    But it also allows monitoring of the output by the op amps for stable output.     Let's not get into this now.       

 Your PS that you posted in your original post uses ONE voltage source and uses that to power the op amps and ref and output voltage.
The first zener, op amp and first transistor provide the first op amp supply voltage and also the consistent, stable ref voltage for the second op amp, the ref voltage is adjusted with the first pot.   The first "zener" is actually just a Base-Emitter junction of a regular BJT, as he states, you can measure the B-E junction of a transistor with your multimeter. You could also use a normal silicon rectifying diode. The resistors around the first op amp dictate its "gain" and therefore the output or reference voltage.
The second op amp supply voltage is regulated by the next two zeners and supplies the drive for the output transistors in response to the input of the "adjusted" reference voltage input.   

I do not know what your Prof wants. I THINK he wants you to build his PS with only ONE transformer output. This is not as good as the two voltage source model in most Lab grade PS, but I think he wants you to do it this way. I THINK he wants the students to learn how to use the individual components. This is a good way of teaching basics. Also the components are cheap and you wont be out big bucks if you fry something....Not that that EVER happens to me.....HA!

[Pixelmarth]

thank you for the article and simulation they are both very helpful in understanding how CC/CV/R_L work together!

I THINK he wants the students to learn how to use the individual components. This is a good way of teaching basics.

For now I'll just build it and see where it leads to hehe!

[wasedadoc]

The first zener, op amp and first transistor provide the first op amp supply voltage and also the consistent, stable ref voltage for the second op amp, the ref voltage is adjusted with the first pot.   The first "zener" is actually just a Base-Emitter junction of a regular BJT, as he states, you can measure the B-E junction of a transistor with your multimeter. You could also use a normal silicon rectifying diode. The resistors around the first op amp dictate its "gain" and therefore the output or reference voltage.
The second op amp supply voltage is regulated by the next two zeners and supplies the drive for the output transistors in response to the input of the "adjusted" reference voltage input.

The resistors  around each opamp set their gains as only 2. If the first "zener" is only 0.6 or 1 Volt, the voltage at the emitter of the first transistor is twice that. Not more than 2 Volts. The voltage at P1 slider can never be more than 10/13.3 of that. The PSU output voltage would never exceed twice that. ie 2*2*10/13.3 = 3.

Also, using the forward base emitter junction of a transistor is subject to variation with temperature. Plus the voltage across the 120k resistor is equal to the "zener" voltage. 0.6 or 1 Volt across 120k means not more than 9uA through the "zener".

And how well does tha opamp work with a supply voltage of only 1.2 or 2 Volts?

Doesn't need a simulation to see that something is very wrong.

Perhaps the person who used a transistor instead of a 1 Volt Zener mistakenly swapped the base and emitter. Typical silicon BJT has emitter to base breakdown voltage somewhere around 7 and will behave in a similar way to a Zener of that voltage.

[xvr]

A little bit offtopic, but just my opinion:

If you want to make your own Laboratory PSU - just buy a good off the shelf one (may be used). You can make PSU cheaper then ready made, but you can't make it better (at least not from first try).

I remember my first PSU (more than 40 years ago) - it was simple, 3 channel device with voltage regulated by variable resistors (without any scales on them). I use DMM to check output voltage. Whole PSU was a web of wires, packed in a box with a wooden top. I use it to power my hand made computer (based on i8080). Also I use this PSU for power up handheld drill (it require adjustment of output from 5V to 12V).
One day my roommate used my PSU and drill to his own needs, and didn't return voltage back. When I plug it (PSU) to my computer afterward, it (computer) expose 'blue magic smoke' and died instantly.

There is only reason to build your own PSU - to learn how it works. After this you would throw it away and buy ready made. 

PS. All above is a my own opinion and I do not insist to take it as an the ultimate truth.

[Wallace Gasiewicz]

wasadedoc and kim christensen have good points about the first "zener."   
But the voltage on the other side of the 120K resistor should be higher. Perhaps about the normal VCC of the op amp.   Perhaps the voltage thru the first transistor is brought up by the base resistor of 12K and then the op amp can take over control.
The two zeners setting the VCC of the second op amp are confusing. I would think that ONE 15 V op amp would be the correct voltage?    I looked at the original article, I do not speak German. Perhaps two smaller value zeners are used to get 15 V.     
Let us know if  it works!
Edit:
If your instructor wants you to build it, it is a good bet you will get a better grade if you do as he says!   
I do not think this is a great Lab PS but it is quite interesting
.

[Pixelmarth]

It won't be "graded" but it's meant to be a learning experience

For now I'm gonna do some tests with a 741 instead of the 761 (which I don't have).
I might just build a PS with different components altogether 

There is only reason to build your own PSU - to learn how it works. After this you would throw it away and buy ready made.

Haha, you're right, it's for learning. I want to know what the first transistor does and also why zeners are used instead of just resistors.
I've been playing around with the second opamp in falstad https://tinyurl.com/2cdho3pa slowly grasping the concept.

[Kim Christensen]

For now I'm gonna do some tests with a 741 instead of the 761 (which I don't have).
I might just build a PS with different components altogether 

Try a LM321 instead of the LM741. Or a LM358 (It's a dual so you'd make a sight mod to the circuit. Or you could be wasteful and use 2) These will work OK, but will limit the upper output voltage as I explained before.
The problem with the LM741, is that it's inputs don't work all the way to ground like the LM358, 321, & 324, do. This will mean that you won't be able to adjust the supply output below apx 3V if you use a 741.
What other OpAmps do you have available to you besides the 741?

[Pixelmarth]

T_T actually that the only one I've got, but I'll try to get my hand on the LM321 or 358 as you mentioned, there shouldn't be a problem. I bought the 741 without knowing too well what it could do 
It's my first time working with opamps so I'm learning by doing. can't hurt!
I'm looking at this: https://www.joretronik.de/Web_NT_Buch/Kap3/Kapitel3_2.html#3.2.6 Image below

[Wallace Gasiewicz]

If you are going to solder things together on a board, use sockets for the op amps so you can switch them easily.

[Kim Christensen]

Not surprising that you ended up buying some LM741. It's the most used chip in tutorials and examples. While it's been superseded by more modern chips long ago, it's still a usable IC within it's limitations. Worth plugging into a breadboard and playing around with.

The L100 you mentioned is also obsolete, but it could still be a useful learning tool. Buy a spare if you're getting them cheap.
The adjustable current limiting in the LM100 circuit you posted appears to interact with the voltage setting, which would be a bit of a pain in a lab power supply.

If you don't need adjustable current limiting, the LM317 could be used instead.  (Good for 1.5A and 40V max input)

[Pixelmarth]

Yeah I've got sockets for the opamps, will use them

[Pixelmarth]

It's the most used chip in tutorials and examples.

Definitely fell for that haha!
I've got a few LM317Ts at work actually, however I need an adjustable current.