LM 317

Hello, I am trying to learn more about circuitry and I am on the topic of the LM317 adjustable voltage regulator. This is different than the fixed voltage regulator that instead of a ground pin it has a Adjust pin that takes part of the output and through a resistor feedback network, it goes into the adjust pin to regulate the voltage. Like the first circuit shown.
What I want to know is actually how this works, i will also have a functional block diagram pic below.
Is this correct to say about the first circuit shown?
When we attach a load to the output terminals, the voltage will naturally drop. This drop will also occur on the resistor divider and show up on the adjust pin as a drop in voltage and will cause an op-amp inside of the Lm317 to either turn on or off a darlington pair of transistors in the block diagram.

But is that correct? A load causes a voltage drop on the output and resistor feedback is fed through the adjust pin and the lm317 sources a higher voltage?
Thanks again.


download.png

download.png

This post was wrong as @ReverseEMF explaines below. I have adjusted it, and hopefully it is now better.

The LM317 keeps the voltage over R2 constant, so the current through R2 and R1 is constant. As a result, the voltage over R1 is also constant. That is used to set the output at a certain voltage.

The LM317 uses a trick to minimize the error by the current from the Adjust pin. It is very low and it is a constant current.
The current going out of the Adjust pin is typical 50µA with a variation of 0.5µA. That current is so low that it can be neglected. Even when it is not neglected, it has minimal effect because it is a constant current, regardless of the load.

It is not the Adjust pin that regulates the output voltage, but the voltage difference between the Output pin and the Adjust pin (the 1.25V) regulates the output current.

. This is different than the fixed voltage regulator that instead of a ground pin it has a Adjust pin that takes part of the output and through a resistor feedback network, it goes into the adjust pin to regulate the voltage.

Actually, it's very similar to a regular voltage regulator... The voltage between pins 1 & 2 is held constant, but instead of grounding pin 1, the voltage comes from a voltage divider. You could "fake out" a regular voltage regulator in the same way...

When we attach a load to the output terminals, the voltage will naturally drop.

Not necessarily... In fact the purpose of the voltage regulator is to hold the output voltage constant when the load changes or when the input voltage changes.

This drop will also occur on the resistor divider and show up on the adjust pin as a drop in voltage

Yes, the voltage divider does divide the voltage.

and will cause an op-amp inside of the Lm317 to either turn on or off a darlington pair of transistors in the block diagram.

Not on or off... Partially-on as voltage-follower current-amplifier.

The op-amp is operating as modified [u]op-amp buffer[/u]. Putting the Darlington transistors inside the feedback loop makes the op-amp automatically compensate for the voltage drop across the transistors.

An op-amp operating linearly with negative feedback has (approximately-theoretically) the same voltage on the + and - inputs. In a buffer (unity gain) application, the output voltage is fed-back to the - input, so the output always follows the + input. If you know what the + input is (relative to ground) you know the output is the same.

Koepel:
... It is typical 50µA with a variation of 0.5µA. ... The only way to get a constant output voltage is a constant current output of the Adjust pin.
... and the Adjust pin (the 1.25V) regulates the output current.

The current through R1 is the constant current from the LM317 plus the current through R2. But that does not influence the voltage regulaton.
... the LM317 increases the output current to get 1.25V over R2, that will also restore the total current through R1 as it was before.

These statements are all incorrect. Yes, there is a current that flows from the Adj pin, but it has nothing to do with regulating the output voltage, and in fact, it's a source of error.

[quote author=From the ST datasheet]Equation 1: VO= VREF(1 + R2/ R1) + IADJ R2 These devices minimize the term IADJ (100 μA max.) and keep it constant with line and load changes. Usually, the error terms: IADJ × R2can be neglected.
[/quote]

But, since it's necessary to maintain a minimum load current of 3.5mA Typ [10mA Max], this IADJ error current becomes negligible -- and can usually be ignored.

And, yes, there is a constant current that flows through the resistors, but it does not come from the Adj pin, but, instead is a secondary effect of the regulation of the 1.25 reference voltage across R2. The LM317 maintains the reference voltage [nominally 1.25V] across R2, which causes a constant current to flow into R1, and it's the sum of the R1 voltage and R2 voltage, that produces the Output voltage. The lack of appreciable contribution from IADJ falls out of the math. When 3.5mA is flowing through R2, 50uA has a mere 1.4% affect on the output voltage. And, because the LM317 keeps the IADJ current "constant with line and load changes" this current can't contribute to regulation -- it only slightly skews the output voltage. And, it's because changes in current flowing through R2 are detected and compensated for, that that is where the actual regulation is occurring.

DVDdoug:
Not necessarily... In fact the purpose of the voltage regulator is to hold the output voltage constant when the load changes or when the input voltage changes.

I think what the OP meant was: when a load is applied to the output of the LM317, the instantaneous effect is the output voltage begins to fall.

And yes, this is true. And the result is, the LM317's internal Op-Amp, detects this deviation, and turns the Darlington transistor further on, to compensate. This is known as the Transient Response.

ReverseEMF:
These statements are all incorrect.

:-[ Thank you for the correction and explanation. I rewrote my post.

Koepel:
It is not the Adjust pin that regulates the output voltage, but the voltage difference between the Output pin and the Adjust pin (the 1.25V) regulates the output current.

Yes. The Adj pin is the sense point. And changes in the divider voltage, sensed at the Adj pin, are amplified, and [negative] feedback to the Output pin, to keep the Reference voltage constant -- which in turn, keeps the total output voltage constant [within the LM317s capability to do so -- which is characterized in the datasheet as the Line and Load Regulation OR "Ripple Regulation" & "Output Voltage Regulation"]

The "adjustable" voltage regulator is exactly the same as a regular "fixed" regulator. You can make a 5V 7805 give you 12V by using the same kind of voltage divider. The calculation method is identical.

You can't adjust the 5V regulator down to 3.3V with the voltage divider method so the LM317 has a wider range of output voltages. I guess that is why it is called "adjustable".

What inside block diagram keeps a voltage difference of 1.25 volts? Is it the op-amp? And what is the purpose of That?
Does when the output voltage drops for a split second, it is sensed on the adjust pin and the input to the op-amp is lower than the 1.25 reference on the other pin of the op amp and this turns on the transistor?

1.25V comes from the diode symbol labelled 1.25V.

It is not a matter of on and off. It is a linear device. If the output draws more current and the voltage starts to drop, the circuit elements represented by the opamp symbol turn on the output transistor more.

tjones9163:
What inside block diagram keeps a voltage difference of 1.25 volts? Is it the op-amp? And what is the purpose of That?
Does when the output voltage drops for a split second, it is sensed on the adjust pin and the input to the op-amp is lower than the 1.25 reference on the other pin of the op amp and this turns on the transistor?

There is an internal "Voltage Reference" [the Zener diode marked "1.25V]. And, yes, you pretty much have it right. The OpAmp compares the voltages at it's two inputs. One input [the non-inverting input marked with a +] is connected to the 1.25V reference, and the other input [the inverting input marked with a -] is connected to the upper side of R2 [the output of the LM317]. The OpAmp performs a differencing operation [thus "Operation" in the device's name], and when the difference is other than zero, it changes the voltage on it's output in an attempt to bring these voltages back to a difference of zero. If the inverting input voltage is higher than the non-inverting input, the OpAmp output voltage will move towards the negative rail [i.e. go more negative]. And, if the inverting input is lower than the non-inverting input, the output will go more positive. Thus, because of the feed back arrangement, this will cause a correction in the output voltage, resulting in a stabilizing of the voltage -- otherwise known as "regulation".

ReverseEMF:
There is an internal "Voltage Reference" [the Zener diode marked "1.25V]. And, yes, you pretty much have it right. The OpAmp compares the voltages at it's two inputs. One input [the non-inverting input marked with a +] is connected to the 1.25V reference, and the other input [the inverting input marked with a -] is connected to the upper side of R2 [the output of the LM317]. The OpAmp performs a differencing operation [thus "Operation" in the device's name], and when the difference is other than zero, it changes the voltage on it's output in an attempt to bring these voltages back to a difference of zero. If the inverting input voltage is higher than the non-inverting input, the OpAmp output voltage will move towards the negative rail [i.e. go more negative]. And, if the inverting input is lower than the non-inverting input, the output will go more positive. Thus, because of the feed back arrangement, this will cause a correction in the output voltage, resulting in a stabilizing of the voltage -- otherwise known as "regulation".

Awesome. Thank you guys so much. Im getting to an exciting point in my learning about electronics where things are starting to click and it is becoming easier to read circuits, thanks to your guys help.