op-amp voltage regulator design

[okbro]

in the following op-amp based shunt voltage regulator circuit the output is supposed to be 5V, with Vz=2.5V and using equation Vout=Vz(R1/R1+R2) from standard textbooks. however in proteus is not working as expected the output is always 2.9V with R1=R2=10kohm and other values of Vz. why?

 

[AnalogKid]

First, I would move the opamp power pin to the 9 V source, before R5.

Next, check out the opamp's input common mode voltage range and output voltage swing range.

ak

 

[okbro]

First, I would move the opamp power pin to the 9 V source, before R5.

Next, check out the opamp's input common mode voltage range and output voltage swing range.

ak

i did that but the output is the same, i also change the diode reference voltage to the both the op-amp terminals but i get the same result. i am using proteus simulator.

 

[crutschow]

And as AK asked, what it that op amps minimum output voltage from its data sheet?

You can't do a correct design without first understanding the device's data sheet.
Real op amps don't act as the ideal op amp you may have studied.

 

[Tony Stewart]

in the following op-amp based shunt voltage regulator circuit the output is supposed to be 5V, with Vz=2.5V and using equation Vout=Vz(R1/R1+R2) from standard textbooks. however in proteus is not working as expected the output is always 2.9V with R1=R2=10kohm and other values of Vz. why?
View attachment 149214

Voltage regulators have the ideal property of a voltage source = 0 Ohms but real regulators have "Load Regulation" error which is always defined by design.

Load Reg. error = ΔV/Vo =1~2% V error for a step load of ΔI ( e.g. 0% to 100% and visa versa for static voltage and smaller steps for dynamic overshoot..) This means the Rout impedance ratio is the same % of the min. load.

So you whenever you design something or ask a question, state Vo (min-max) Imax and error budget.

Now looking at your design, the shunt regulator has the source impedance of the collector pullup R = 100, rather than something like << 1 Ohm such as like the Rout of an Op Amp with negative feedback and a non-inverting gain of 2 * Vz.

For more current than the Op Amp can deliver, add a transistor amplifier with care to avoid oscillation.

So may I see some design specs and your next design Rev?

 

[schmitt trigger]

I join the chorus of mentioning either the opamp's input common mode range or its minimum output voltage (Vol).
Why did you choose a TL072in the first place?

 

[Tony Stewart]

p.15 of 90 of TL072 sheets says 1.5V min for Vcm which is expected for Nch inputs. The Vz is 2.5 V so this satisfies the Vcm range.

The real problem is the output headroom for linear operation. does not guarantee operation down to drive Vbe. Adding series Rb could correct that with divider pair. A pull down on base to lower Vbe will allow the collector to rise. A series Rs will raise the Op Amp out to operate in the linear range.

 

[danadak]

Lot of base drive with no limiting -

Using LM324 output can shut off NPN and put load into regulation

 

[okbro]

i will recheck the circuit today but i think the TL072 operates with a single supply voltage ranging from 3V to 36V (or ±1.5V to ±18V for dual supply configurations). This should be sufficient for most low-voltage circuits, but considering the above comments:

About the Input Common-Mode Voltage (Vcm), i know it should be within the range defined in the datasheet. For TL072, this is typically 1.5V above the negative rail (if using single supply), and it may struggle if the input voltage is too close to the supply rails. TL072 does not have rail-to-rail output. It can typically swing 1.5V to 3V below the positive supply (depending on the load) and 1.5V to 3V above the negative supply. So, using a 9V supply (for example), the output could realistically only reach a maximum of 6V-7V, not 5V as I wanted. This is likely the cause of the output being stuck around 2.9V instead of 5V in my simulation. A;sp in the circuit, a reference voltage (Vz) of 2.5V, which is well within the TL072's capabilities. However, the output needs to reach 5V, which might not be feasible due to the op-amp's output swing limitations.

 

[danadak]

I think key issue is outout swing of 072 will not go close to ground,
so NPN does not get reduced base drive.

Using LM324 (works) :

Using TL072 (not working) :

Using TL072 with a 2.4V zener to get output into operating range circuit
regulates to 5V (working) :

Note I am not a fan of zener approach, as its leakage and beta variation of NPN
could keep the drive worst case minimum too high affecting regulation. But am a
fan of LM324 like OpAmp due to its ability to drive close to its negative rail, 20 mV.

Of course you could always provide a -2V to TL072 OpAmp - supply pin......to accomplish
complete control of NPN.

 

[alec_t]

Adding series Rb could correct that

Agreed. The op-amp output is otherwise clamped at ~0.6V.

 

[Nigel Goodwin]

I think key issue is outout swing of 072 will not go close to ground,
so NPN does not get reduced base drive.

I agree, it really needs a resistor from opamp to base, and another from base to emitter.

 

[danadak]

That affects load regulation due to reduced loop G. Sim work could establish
how impactful that would be.

First trial......ugh.

 

[okbro]

yes as danadak said it works with LM358N(LM324) but not TL072. Also changing the load resistance or changing the zener diode breakdown current changes the output voltage, which should not be the case. also with LM358N I had to use +15V power supply. the same series version works with both LM358 and TL072 with +9V.

 

[Tony Stewart]

They call it "negative rail headroom" in the datasheet.
That means it is almost rail to rail but not enough to turn off Vbe, unless a diode is added. in series or R-divider or a Darlington is used or a different Op Amp.

I suspect this soft limiting speeds up the pulse recovery time for other applications by reducing Cout.

 

[atferrari]

R1 and R2 are suposed to be a voltage divider?

 

[Nigel Goodwin]

R1 and R2 are suposed to be a voltage divider?

Yes, they are used to set the output voltage, and their midpoint voltage is compared with the reference on D1 - so equal values give 5V out.

 

[danadak]

Of course we have OpAmp possibly acting more as a comparator, slew rate
dominated in 324 358 types of parts.

TL072 :

Moving Vz and OpAmp supply directly to battery made no appreciable difference.

 

[Nigel Goodwin]

Of course we have OpAmp possibly acting more as a comparator, slew rate
dominated in 324 358 types of parts.

Acting as a comparator, plain and simple.

 

[danadak]

Usable in some situations :

T swept 0 - 50 C

Reducing input dropping R :