Help me understand this. I am searching on information on building a 10 amp battery charger. On the many sites I've seen circuits using the LM317 I see the Vout to Adj pin has 240ohm connected.
On more than one forum, including this one, I read experts state that it should be 120 for a minimum current of 10Ma and that 240 ohms only applies to the more expensive LM117

In one TI datasheet I've downloaded covering LM117, LM317A @ LM317-N all the diagrams show 240ohms but the chip is labelled LM117. Unless I've missed it I didn't see a note that the diagrams don't apply to the 317

I started to believe the 120 ohm story anyway....

However when I happened upon the Wikepedia page for LM317 I see datasheets from 5 manufacturers all of whom show 240 ohms in their LM317 datasheets..This includes Texas Instruments. So where is the 120 thing coming from ?

 

Just in case I was asked I've been back searching for where I saw the info. It was mostly attributed to AudioGuru and I hadn't noticed it was always the same person. I see that person is apparently now banned..Is that the end of the story ?

 

The output voltage formula is


\({V_{out}} = 1.25\left( {1 + \frac{{{R_2}}}{{{R_1}}}} \right)\)

So it the choice of R1 will depend upon the desired output and the choice of R2 and the need to keep a minimum of say 10mA flowing.

 

The LM 317 needs 10 mA holding current and corresponds to 120 ohms at 1.2 volts

Lm117 needs only 5 mA and corresponds to 240 ohms at 1.2 volts.

 

Here is the Fairchild datasheet for the LM317:

http://www.fairchildsemi.com/ds/LM/LM317.pdf

You will note that, as you indicated, it shows a 240Ω sensing resistor.

Now let's consider a few other things that might influence the choice we actually make.

1) The adjustment pin current is spec'ed to have a current as high as 100μA.

2) The minimum load current needed to maintain regulation may be as high as 12mA.

So how do these come into play with a 240Ω resistor?

Since the voltage across this resistor is 1.25V, the current in it is 1.25V/240Ω or 5.2mA. So while this is comfortably above the typical minimum load current of 3.5mA, it isn't even half of what might be necessary to maintain regulation on a particular device. Using a 120Ω resistor gets that up to 10.4mA, which doesn't quite guarantee that the minimum load current is met at all times, but it is becoming pretty unlikely that it wouldn't me. If you were to use a 100Ω resistor, then even under the lowest spec'ed reference voltage of 1.20V you would ensure that the minimum load current was always met.

The downside of using a 120Ω, or even 100Ω, resistor is that you are wasting power and increasing heat dissipation, either of which can become a concern, in situations in which the remainder of the load will always be drawing enouigh current to maintain regulation. So it is a design decision that should be made in view of the application at hand.

As for the adjustment pin current of 0.1mA, notice that this is about 2% of the current in the feedback resistor. The variation in the reference voltage is about 4%. So while the possible variation in the reference voltage is the more significant of the two, it would be nice to make it truly the dominant error source by reducing the error due to adjustment pin current at least a bit more. By using a 120Ω resistor, or smaller, you can force this error source to be below 1%.

So if you would like to have a nice rule of thumb that let's you not have to worry about making sure that the circuit powered by your regulator draws sufficient minimum current, then use a 100Ω or 120Ω resistor. But using a 240Ω resistor will work, provided you keep the minimum current requirements in mind and ensure that they are satisfied.

 

The LM 317 needs 10 mA holding current and corresponds to 120 ohms at 1.2 volts

Lm117 needs only 5 mA and corresponds to 240 ohms at 1.2 volts.

Can you comment then on the datsheets I mentioned here: http://en.wikipedia.org/wiki/LM317 and can you put a link where it confirms what you are saying in an "application note" or similar.

I'm quite clear on calculating the output voltage. There are even several free online calculators.

sorry Wbann..you posted while i wrote.

 

@Wbahn.. I won't quote the whole reply.

Ok I think I follow you drift. The minimum load then includes an actual load ?
assumption...So if anything from an LED indicator upwards was permanently connected a 240 would be fine ?
The regulation may be less than perfect if the regulator was open circuit using 240 - but why would that matter ?
Why is it that I've yet to find a LM317 circuit on the net that uses 120 ohms ?
And the biggest why is why all the manufacturers data sheets show 240 in their typical application notes and don't make the proviso that you have ?
Sorry that's a lot of questions but let's get right to the bottom of this..

 

Ok I think I follow you drift. The minimum load then includes an actual load ?

Yes. The regulator neither knows nor cares where the current that leaves its output pin goes.

assumption...So if anything from an LED indicator upwards was permanently connected a 240 would be fine ?

Yes, as long as the minimum output current is met. There is nothing magical about 240Ω. Like all things engineering, it is a compromise. You could use 470Ω or 1kΩ. If the minimum output current is met, you are fine. Now, you still have to contend with the issue that the adjustment pin current is becoming a bigger and bigger part of the total voltage setting current, but if you account for that and can live with it, then you are good to go.

The regulation may be less than perfect if the regulator was open circuit using 240 - but why would that matter ?

In a particuilar application, it may not. But in other applications it may matter a lot because you may have devices that can be damaged or destroyed if too much voltage is applied. Remember, the output doesn't have to be open circuit in order for the load current to fall well below the minimum current. The very first product I ever designed was a radon level monitor and the entire unit, analog and digital, drew less than 1mA of current when operating.

Why is it that I've yet to find a LM317 circuit on the net that uses 120 ohms ?

You found some here, didn't you?

And the biggest why is why all the manufacturers data sheets show 240 in their typical application notes and don't make the proviso that you have ?

It's quite possible that the manufacturers don't have the option. They licensed the ability to manufacturer a part that they get to label "317" in order to tap into a ready-made market base; that comes with restrictions on what they have to include in the datasheet and what they can and can't change. Also, if the original spec sheet used 240Ω, then they are safe, legally, just copying that and incur some exposure if they change it. They also create the impression that "their" 317 part is not the same as other 317s out there. Plus, it's the easiest thing to do.

 

Thank you..I think I now understand the LM317 as well as I ever will..

 

Yes. The regulator neither knows nor cares where the current that leaves its output pin goes.



Yes, as long as the minimum output current is met. There is nothing magical about 240Ω. Like all things engineering, it is a compromise. You could use 470Ω or 1kΩ. If the minimum output current is met, you are fine. Now, you still have to contend with the issue that the adjustment pin current is becoming a bigger and bigger part of the total voltage setting current, but if you account for that and can live with it, then you are good to go....

Thanks WBahn, now I understand lm317 better

 

For low voltage regulator, energy loss is more, if R1 is less.

For example, for a 5V regulator, if R1 is 120 ohm:

5V / (360 + 120) ohm = 10.42 mA loss

or if R1 is 270 ohm:

5V / (270 + 810) = 4.63 mA loss

I always use either 240 or 270 ohm for lower voltage regulation.

 

The LM 317 needs 10 mA holding current and corresponds to 120 ohms at 1.2 volts

Lm117 needs only 5 mA and corresponds to 240 ohms at 1.2 volts.

Here's the truth of it:

The LM117 and LM317 are exactly the same part (die) but tested to different limits. The LM117 has to work over the full MIL temp range of -55 to 150C and the LM317 only has to work from -40 to 125C.

Most LM317 devices PROBABLY will work OK using the 5mA minimum load, but some won't. Ignore the spec limits at your own peril.

 

The regulation may be less than perfect if the regulator was open circuit using 240 - but why would that matter ?

Because the output voltage rises when the min load is not there. It could go up and damage downstream parts like the output cap or other devices on the output.

 

Here is the Fairchild datasheet for the LM317:

http://www.fairchildsemi.com/ds/LM/LM317.pdf

You will note that, as you indicated, it shows a 240Ω sensing resistor.

Now let's consider a few other things that might influence the choice we actually make.

1) The adjustment pin current is spec'ed to have a current as high as 100μA.

2) The minimum load current needed to maintain regulation may be as high as 12mA.

So how do these come into play with a 240Ω resistor?

Since the voltage across this resistor is 1.25V, the current in it is 1.25V/240Ω or 5.2mA. So while this is comfortably above the typical minimum load current of 3.5mA, it isn't even half of what might be necessary to maintain regulation on a particular device. Using a 120Ω resistor gets that up to 10.4mA, which doesn't quite guarantee that the minimum load current is met at all times, but it is becoming pretty unlikely that it wouldn't me. If you were to use a 100Ω resistor, then even under the lowest spec'ed reference voltage of 1.20V you would ensure that the minimum load current was always met.

The downside of using a 120Ω, or even 100Ω, resistor is that you are wasting power and increasing heat dissipation, either of which can become a concern, in situations in which the remainder of the load will always be drawing enouigh current to maintain regulation. So it is a design decision that should be made in view of the application at hand.

As for the adjustment pin current of 0.1mA, notice that this is about 2% of the current in the feedback resistor. The variation in the reference voltage is about 4%. So while the possible variation in the reference voltage is the more significant of the two, it would be nice to make it truly the dominant error source by reducing the error due to adjustment pin current at least a bit more. By using a 120Ω resistor, or smaller, you can force this error source to be below 1%.

So if you would like to have a nice rule of thumb that let's you not have to worry about making sure that the circuit powered by your regulator draws sufficient minimum current, then use a 100Ω or 120Ω resistor. But using a 240Ω resistor will work, provided you keep the minimum current requirements in mind and ensure that they are satisfied.

This really should be a sticky or something. In just a few paragraphs, it clearly describes design considerations for the LM317 in a way even newbies or the analog challenged (like me ) should be able to understand.

 

2) The minimum load current needed to maintain regulation may be as high as 12mA.

The TI data sheet says the limit is 10mA for the LM317. Fairchild must have a problem if theirs is 12 mA because NS invented the part and the correct spec has been 10mA for as long as I can remember. That is where the 120 Ohm resistor value came from since it is across 1.25V. The other makers are also on the correct 10 mA spec.

ON Semi also has 10mA.
http://www.onsemi.com/pub_link/Collateral/LM317-D.PDF

Linear Tech is also 10mA.
http://cds.linear.com/docs/en/datasheet/lt0117.pdf

 

I'd like to know WHY you choose a 317 when, if I understand properly, you want to make a Battery charger of up to 10A.
Are you planning on driving a thick bit of Silicon with it ?