sometimes . . . the voltage regulators do not like diodes . . . at certain places (the suspects D9 CLED2)

if i remember right - then the optimal value for R4 R11 is 120Ω ← me.self-chk. 1.25V / 120Ω = 10mA ??? the regulator's ADJ pin uses 400µA to 5mA
i don't remember the reason behind the 120Ω anymore . . .

. . . is likely there
** Adjustment Pin Current Change -- ΔI.adj : Typ. / Max. -- 0.2 / 5 µA
Conditions : 10 mA ≤ I.OUT ≤ I.MAX , 3V ≤ (VIN − VOUT) ≤ 40V ← allows "good" regulation with no useful load attached (so it's not a must if you have over 10mA @ OUTP at any time) . . .

combined with (somewhat a "fuzzy math.") ::
Adjustment Pin Current -- I.adj : Typ./Max 50/100 µA and ΔI.adj : Typ. / Max. -- 0.2 / 5 µA →
→ R.adj = 1.25 / .1mA ≈ 12kΩ ← somewhat can be considered to be in parallel with your R4 or R11 (the smaller the R1 on the datasheet** the more precise the regulation)
→ ΔR.adj = 1.25 / 5µA ≈ 250kΩ . . . ×0.1% = 250000Ω / 1000 = 250Ω ← a maximum valuer for R1 to get 0.1% precision for OUTP voltage

 

Why did you think it would limit to 5A?
As you found out, it doesn't.
To limit the current you need a current-limit circuit.
No, a LM311 is a comparator, not an op amp.
A single-supply LM339/393 op amp should work.

@crutschow
I believe the LM339/LM393 are single-supply comparators, not opamps. Perhaps you meant LM324/LM358?

 

What do RV1 and RV4 do?

RV4 is used to adjust the output voltage. RV1 is trimmer to make sure 15V (not less or over) when RV4 pot is turned full clockwise.

 

@crutschow
I believe the LM339/LM393 are single-supply comparators, not opamps. Perhaps you meant LM324/LM358?

Yup.
I get those confused if I don't carefully think about it.

 

What do RV1 and RV4 do?

RV4 is used to adjust the output voltage. RV1 is trimmer to make sure 15V (not less or over) when RV4 pot is turned full clockwise.

 

I would appreciate if you point me to a direction (book, link) on how to build a power supply with my specification from scratch,

Don't know of anything the covers how to do such a specific task.
It's mostly a matter of knowing basic analog design and going from there.
My motto is to follow the KISS school of design or, to paraphrase Einstein, "Make it as simple as possible, but not simpler".
You can usually make a complicated circuit to do what you want.
The real design work is to simplify that circuit as much as possible, while still having it do the desired function.

As an example, below is my take on a design that does what you want.

It uses the same basic LM317 transistor boost circuit to generate 5A output.

To get 0V minimum output from an LM317 and to allow current limit into a 0V short circuit, the ADJ bottom resistor (pot U5) must be connected to a minus 1.25V reference voltage.
I generate that with an LM337 negative regulator.
Pot U5 thus allows adjustment of the output voltage from 0V to over 15V.

I also used that -1.25V as a reference for the current limit function of op amp U3.
U3 starts to turn off U1 when the current through U4 from the shunt resistor R3 voltage exceeds the current through R5 from U2's voltage.
The limit is adjustable by U4 from essentially zero to 5A.

U2's Adj pin is connected to the bottom of the load connection instead of ground so that the voltage drop across the shunt resistor R3 doesn't degrade the voltage regulation.
This does have the effect of slightly changing the reference voltage for the current limit as the load current changes, but that has no significant effect on that limit function, other than to change the value of limit resistor R5 from that calculated if the Ref voltage didn't change.

The LTspice simulation shows the operation for three positions of the voltage adj pot U5 to give nominally 0V, 10V and 15V out, while the output load resistance varies from 10 ohms to 1m ohm (short), with the current limit set to the 5A max.

 

sometimes . . . the voltage regulators do not like diodes . . . at certain places (the suspects D9 CLED2)

if i remember right - then the optimal value for R4 R11 is 120Ω ← me.self-chk. 1.25V / 120Ω = 10mA ??? the regulator's ADJ pin uses 400µA to 5mA
i don't remember the reason behind the 120Ω anymore . . .

. . . is likely there
** Adjustment Pin Current Change -- ΔI.adj : Typ. / Max. -- 0.2 / 5 µA
Conditions : 10 mA ≤ I.OUT ≤ I.MAX , 3V ≤ (VIN − VOUT) ≤ 40V ← allows "good" regulation with no useful load attached (so it's not a must if you have over 10mA @ OUTP at any time) . . .

combined with (somewhat a "fuzzy math.") ::
Adjustment Pin Current -- I.adj : Typ./Max 50/100 µA and ΔI.adj : Typ. / Max. -- 0.2 / 5 µA →
→ R.adj = 1.25 / .1mA ≈ 12kΩ ← somewhat can be considered to be in parallel with your R4 or R11 (the smaller the R1 on the datasheet** the more precise the regulation)
→ ΔR.adj = 1.25 / 5µA ≈ 250kΩ . . . ×0.1% = 250000Ω / 1000 = 250Ω ← a maximum valuer for R1 to get 0.1% precision for OUTP voltage

I built the circuit from scratch, removed the 742 op-amp for -1.2v reference in order to narrow down where the problem was and as you suggested I changed value of R11 to 120 Ohm and checked the diodes and LED. I tested the circuit by setting the Vout at 12v then connected a 12v/1.5A fan as a load the voltage dropped to 11.64 like before. If I connect a higher load the voltage drop gets even worse. There is something basic wrong with this circuit or I am making the same mistake again and again because the basic circuit is advertised as high end bench power supplier on the web.

 

Don't know of anything the covers how to do such a specific task.
It's mostly a matter of knowing basic analog design and going from there.
My motto is to follow the KISS school of design or, to paraphrase Einstein, "Make it as simple as possible, but not simpler".
You can usually make a complicated circuit to do what you want.
The real design work is to simplify that circuit as much as possible, while still having it do the desired function.

As an example, below is my take on a design that does what you want.

It uses the same basic LM317 transistor boost circuit to generate 5A output.

To get 0V minimum output from an LM317 and to allow current limit into a 0V short circuit, the ADJ bottom resistor (pot U5) must be connected to a minus 1.25V reference voltage.
I generate that with an LM337 negative regulator.
Pot U5 thus allows adjustment of the output voltage from 0V to over 15V.

I also used that -1.25V as a reference for the current limit function of op amp U3.
U3 starts to turn off U1 when the current through U4 from the shunt resistor R3 voltage exceeds the current through R5 from U2's voltage.
The limit is adjustable by U4 from essentially zero to 5A.

U2's Adj pin is connected to the bottom of the load connection instead of ground so that the voltage drop across the shunt resistor R3 doesn't degrade the voltage regulation.
This does have the effect of slightly changing the reference voltage for the current limit as the load current changes, but that has no significant effect on that limit function, other than to change the value of limit resistor R5 from that calculated if the Ref voltage didn't change.

The LTspice simulation shows the operation for three positions of the voltage adj pot U5 to give nominally 0V, 10V and 15V out, while the output load resistance varies from 10 ohms to 1m ohm (short), with the current limit set to the 5A max.

View attachment 190350

Thanks for your explanation @crutschow.

 

#13 : LM301A Slew Rate of 10V/μs , Output Short Circuit Duration Continuous ((if it makes any difference = i don't have any good ideas yet ...))
LM317T has also a thermal protection
? LM317 Ti (Fig.23)
? ON-Semi (Fig.22)
- - - -
You didn't tell your transformers data -- so i made a modest speculation . . . that it doeasnot output much over 5A
-
Note : the R15 was added initially with a higher resistance value of 330mΩ to give a bit emitter (neg.-)feedback . . . but then i studied that i have come up with a transformer model near my estimated target (5A 12V) requirements (and that the R15 appares to be one of the bottle-necks in the circuit below) ← ← what that means is that if you actually have more powerful transformer - - then there yet won't be issue with the R15 --or-- Q1.E.R.ser value /// In the simulation below i just wanted to grasp some touch of the combination of LM317 and Power-PNP ← and that's all

 

what i then thought is that your transformer is cryptically connected to supply power to both
to the LOAD & to the Op Amps
/!\ the high load current may start have impact on the Op Amp-s' supply output /!\

 

what i then thought is that your transformer is cryptically connected to supply power to both
to the LOAD & to the Op Amps

Thank you @ci139 . I guess that was it. I simulated the circuit with a separate supply for the op-amp and voila it is stable even with a high load. I have to make the actual circuit though but I am hopeful that it will work. I have to change the -1.2v reference with @crutschow's idea too, because it looks better. I am gonna be away for a while I will post if the actual circuit works. Thanx again.