What is a voltage reference, and why does the LM317 require one?

[olsenn]

This is directly from the LM317 datasheet from Texas Instruments:
       
"In operation, the LM117 develops a nominal 1.25V reference
voltage, VREF, between the output and adjustment terminal.
The reference voltage is impressed across program resistor
R1 and, since the voltage is constant, a constant current I1
then flows through the output set resistor R2, giving an output
voltage of..."

Hence Ohm's Law! That's honestly all there is too it. Below is a quote from YOU, and if I was the person asking what the reference terminology means, I wouldn't understand either.

"Without a voltage reference for most linear regulators , they would be essentially useless .
A voltage reference is a feedback loop .
It has to know what voltages you want , how else is it "adjustable" ?"

[Balaur]

Secondly, he is actually SMARTER than the bunch of you... when he didn't know the answer he asked, when you didn't know the answer, you guessed... and were WRONG. The LM317 doesn't compensate and it doesn't have any idea of what voltage it is spitting out; it operates on Ohm's Law, and that's it! It has NO VOLTAGE REFERENCE. If you want 5V, you have to hook it up to a multimeter and tune the resistors until you get 5V and hope it never changes! A voltage reference would be used for something like an analog-to-digital converted which needs to compare TWO voltage DIFFERENCES ([Vin - Vref] is five times larger than [Vref - 0], and I know Vref is 1.25V, thus Vin = 5*1.25= 6.25V). The only time you would need to consider the 1.25V in the LM317 as a "Reference" would be to pass it to the Vref of an AtoD converter (if it doesn't already have a built in reference) along with the output, so that it can determine and display what voltage that is being output instead of you having to connect the damn thing to a multimeter to see how much more you need to adjust the damn pot.

This is directly from the LM317 datasheet from Texas Instruments:
       
"In operation, the LM117 develops a nominal 1.25V reference
voltage, VREF, between the output and adjustment terminal.
The reference voltage is impressed across program resistor
R1 and, since the voltage is constant, a constant current I1
then flows through the output set resistor R2, giving an output
voltage of..."

Hence Ohm's Law! That's honestly all there is too it. Below is a quote from YOU, and if I was the person asking what the reference terminology means, I wouldn't understand either.

"Without a voltage reference for most linear regulators , they would be essentially useless .
A voltage reference is a feedback loop .
It has to know what voltages you want , how else is it "adjustable" ?"

This is highly amusing.

Please forgive my crass ignorance and impertinence, but I would like to ask you kind sir, to get a better look at the theory behind all this.

The 1.25V reference voltage means exactly that: a bandgap temperature-independent voltage reference, indifferent (within limits) to whatever supply voltage the device receives.

Please have a look at various documents available online:
ECE 237
Wiki

Cheers,
Dan

[alm]

Honestly, I am ashames at the people on this forum! Musashibaka asked a good question for the right reason; because he didn't know the answer and wanted to. There is no need to be assholes and either purposely try to confuse him or insult him for asking electronics questions on an electronics forum.

In general I think questions like 'what's a voltage reference' are better looked up in a text book (or Wikipedia) than asked on a forum. Asking such questions on a forum will result in some incorrect answers and some self-proclaimed experts arguing about some irrelevant detail.

The LM317 doesn't compensate and it doesn't have any idea of what voltage it is spitting out; it operates on Ohm's Law, and that's it! It has NO VOLTAGE REFERENCE. If you want 5V, you have to hook it up to a multimeter and tune the resistors until you get 5V and hope it never changes! A voltage reference would be used for something like an analog-to-digital converted which needs to compare TWO voltage DIFFERENCES ([Vin - Vref] is five times larger than [Vref - 0], and I know Vref is 1.25V, thus Vin = 5*1.25= 6.25V). The only time you would need to consider the 1.25V in the LM317 as a "Reference" would be to pass it to the Vref of an AtoD converter (if it doesn't already have a built in reference) along with the output, so that it can determine and display what voltage that is being output instead of you having to connect the damn thing to a multimeter to see how much more you need to adjust the damn pot.

You seem to have a strange concept of voltage reference. A voltage reference is commonly considered a component that provides a constant voltage, to be used as reference for something. How would your "just Ohm's law" regulator if Vref (which of course doesn't refer to reference voltage) was a variable voltage, eg. just 20% of the input voltage? Would there be any regulation at all? How is this internal thing which provides 1.25V but is not a voltage reference connected to the internal error amplifier? If it just operates on Ohm's law, what happens if you connect the adjust terminal to ground without any resistors between the output terminal and ground? I assume it just stops working?

[westfw]

The goal of a voltage regulator is to output some constant voltage output, regardless of load.
I order to do this, the regulator has to have some "reference" to a know voltage that it can compare against; the 'reference voltage.'  The regulator circuit could say "I want my output voltage to equal this reference"  In the 317 case, the circuit will adjust the "resistance" of a power transistor element in ways that will make the difference between the output voltage and the "adjust" input match an internal 1.25V reference.  By putting a resistor divider between the output and the "adjust", you can achieve any output voltage greater than 1.25V  (if you wanted 5V, you'd use a voltage divider that divided by 4.)  1.25V is presumably a compromise between what it is easy to fabricate, and what a reasonable "low" output voltage would be...

[Wartex]

There is a point where the only sensible advice to give is to suggest to pick up another hobby, for example, knitting.

Could someone please take this arrogant dick down a notch, he's fucking personally insulting everyone in 30% of his posts. I would have been crucified by mods for shit like this, yet he gets a free pass.

[SgtRock]

Dear Wartex:

--While I have not always agreed with you in the past, I do have to say that, usually you saved your wrath for a very few people, with whom you have had disagreements. This guy hates everybody. All he does is chime in, give advice (often not very well thought out advice) and then, when everyone does not say "Jawohl Mein Herr", he starts with the gibbering, slobbering and cursing, and oft times using the most lurid disgusting profanity in personalized attacks, though lately he has improved somewhat, causing some to guess that perhaps he had been issued the yellow card.

"There once was a worker, so bored.
That when his posts were ignored.
He'd strive on inanely,
And most times profanely.
But seldom, if ever he scored."
Gator Dundee 1948 -

Best Regards
Clear Ether

[Mechatrommer]

There is a point where the only sensible advice to give is to suggest to pick up another hobby, for example, knitting.

Could someone please take this arrogant dick down a notch, he's fucking personally insulting everyone in 30% of his posts. I would have been crucified by mods for shit like this, yet he gets a free pass.

i'll take it positively. he just wanted to save some space in this world by avoiding useless arguments such as why V=IR not I=RV, this is not a voltage reference, that is etc etc and what not. someone suggested... read it up in text book, i agree, and internet is vast, thanks to google for managerial effort to collect them. without him i believe the OP will continue to ask why and why indefenitely until why V=IR until this thread will take up space similar to one textbook. BaW did give some advice on what is the Vref (and some others), if the OP still unable to understand and dont make any effort to further investigate it, then there's a problem. you want to help explaining what a votage reference is? then go ahead. we will be glad you are willing to spoonfeed this chap. the OP and the viewer may choose to skip which posts they dont like. but instead of knitting (which i did long time ago and becaming good at) i will suggest for lawn bowl.

[musashibaka]

I really appreciate all of the helpful responses I have received.
I would like to address some of the assumptions made by
a few negative posters.

1. I actually do know how to use wikipedia and google.

2. Anyone who has tried either of the above would know that the
quality of information found on topics vary, and sometimes the right
description required to get an intuitive understanding varies per individual.

Not only that, if everything in the world could be explained well by a
google search, this forums sole purpose would be for entertainment.

3. I have read wikipedia page on voltage reference and the lm317.
I have also searched through the three electronics books
Practical Electronics for Inventors, Fundamentals of Analog Circuits, and Beginning
Electronics to name a few. So I sought further advice.

4. I am not an electrical engineer, but I would gather that throughout an engineers
career one must ask questions in order to learn. Excusing rude behavior
under the auspices of abating "spoon feeding" is a weak minded response to
human curiosity.

To those individuals who chose to be rude, or because of their own insecurity
chose to respond with a weak attempt at belittle me, I offer to you the gift of wisdom:

"Any intelligent fool can make things bigger and more complex...
It takes a touch of genius - and a lot of courage to move in the opposite direction."
-- Albert Einstein

[The_Penguin]

I really appreciate all of helpful responses I received.
I would like to address some of the assumptions made by
some negative posters.

Well said, sir!

[Mechatrommer]

know what? my wife will never wrong, always defend herself no matter what.

I actually do know how to use wikipedia and google.

do you? talk somemore. well said.

http://en.wikipedia.org/wiki/Voltage_reference
http://www.ko4bb.com/e102/e102-4.php
http://www.edaboard.com/thread215325.html

http://lmgtfy.com/?q=why+we+need+voltage+reference
http://lmgtfy.com/?q=why+use+voltage+reference
http://lmgtfy.com/?q=what+is+voltage+reference
http://lmgtfy.com/?q=what+the+fuck+is+a+voltage+reference%3F
http://lmgtfy.com/?q=why+a+voltage+reference+fucked+up%3F

took me some time to meditate whether to reduce my post count or not. decided to leave as is for entertainment.
valid question? many valid answers but still not answering. what is valid? or just a simply junk question?

[Hero999]

Check out the following links:
http://www.ti.com/lit/an/snva558/snva558.pdf
http://www.ti.com/lit/an/snva020a/snva020a.pdf

[MarkS]

I really appreciate all of the helpful responses I have received.
I would like to address some of the assumptions made by
a few negative posters.

I do not want to be considered one of the "negative posters", but break down "voltage reference". It is a voltage that can be referred to. What it is is basic common sense. You should be able to extrapolate what it is, and thusly, WHY it is from the two words that make up the description. That is why people have been negative. That and the fact that you seem to skip over the answer and ask why again while quoting the answer.

Now, there are questions that deserve asking, like how is it constructed and why is it only 1.25 volts when the output is going to be greater.

[ciccio]

This thread is very interesting::
1) a "newbie" post a simple question, in clear and plain English. From it's later post we understand that he is not illiterate, and owns and can read engineering books.
2) 30 replies in two days: it's a lot. 
3) more than 1/4 of the replies could be avoided: they did not gave an answer to the original poster, but were (I'll try to be "diplomatic") a little on the "unpolite" side, revealing a selfish and "arrogant" (sorry, I did not find a less crude synonym)  position.
3) another 1/4 of the replies is used to question some other's response.
3) another 1/4 were not giving a correct answer (I mean an answer that is both technically correct and that can be understood by the original poster).
4) the lat 1/4 tries to give, with various success, a response to the original question.
I really appreciated some of the posts in this last group, and could write here my appreciation to the posters, but I want to avoid other conflicts, so I end here.
Fortunately I knew before what a Voltage Reference was, or my ignorance could only increase, being obliged to mediate all these informations.
Best regards

[MarkS]

..snip..

I think that most people here have the best intentions. He asked a question and received an answer. He then proceeded to ask the question again, while quoting the answer he just received, as though the answer in front of him wasn't there.

While I hate it when someone is outright told to do a search first, I did a quick search on Google and this was the first reply: Voltage Reference. That leads to a rather useful article here: Understanding Voltage-Reference Topologies and Specifications

I don't think BoredAtWork, for instance, was being rude. This isn't the hobby for everyone. I was able to pick this up from a very early age and am entirely self taught. However, many people will look at what I would consider mundane and their eyes will glaze over. The topic in question is a very simple EE topic. If the OP doesn't understand after the answers that have been given, which were entirely on point, then maybe this isn't the hobby for the OP.

[SgtRock]

Greetings EEVBees:

--I hope some of you will consider being more tolerant of beginners. They are important to this blog, that is why it seeks them. Now, when someone asks; "just exactly what is a voltage reference, and what does it do", I rather think he is looking for an explanation of some length, and not just, "its a voltage, reference, you fool, how else would you do it." Or "take up knitting".

--I can remember taking courses (Calculus comes to mind) where I just could not get a particular concept. I would search and search and question and question, till finally someone would push the bingo button. So when somebody asks a question and you give a short answer, and he still does not get it, try explaining more fully or reccommending a particular "at length" article. You will have a much better chance of not seeming like arrogant bastards.

"Does not squirrel crack nuts on bough of oak tree."
Lao Fu 1410 1620

Best Regards
Clear Ether