317 resistance values

It's no big deal. The 240 ohm value dumps enough current to satisfy the minimum load requirement. If you have a decent load, that's taken care of.

Very high values, many K ohms, might affect dynamics. Or make some DC error; the adjust pin current is typically 50 uA.

Try it!

Two issues. First, you usually want the voltage divider to sink the datasheet minimum load current. That prevents the output from floating up out of spec and maybe blowing up something downstream.

IIRC the minimum load for the guaranteed performance is 10 mA, so you want a 120-ohm resistor on the top of the divider unless you're sure you'll always have enough of a load to maintain regulation.

Second, the 317's ADJ pin sources a reasonably-constant 50 uA, so you want the voltage divider impedance low enough that this won't limit the voltage accuracy. In order to achieve that value, the 317 dumps all its quiescent current (used for biasing internal nodes) into the output pin, which is why the minimum current is so high.

The LM78xx series works well down to zero load current, because its quiescent current goes out the ground pin.

Cheers

Phil Hobbs

snipped-for-privacy@gmail.com wrote: =====================

** The choice of 240 ohms means a tad over 5mA flows in the path to negative common. This swamps the approx 50uA flow from the " adj" pin itself by 100 to 1. The output voltage will then be as calculated to within 1%.

Using lower values for both resistors would improve the error margin slightly. While making the values higher would do the opposite.

..... Phil

240 means 5.2mA nominal from the regulator. Worst-case minimum to maintain regulation is 10mA (Vi-Vo = 40V) but typical is 3.5mA so it is usually enough under sensible conditions. With a (more sensible) Vi-Vo of 10V or less, minimum load current is in the 1.4-1.7mA typically, so 5.2mA is pretty safe.

The current flowing out of the ADJ pin is nominal 50uA so it will have an effect on the voltage even if the minimum load current is made up elsewhere. It typically varies from 40 to 57uA or so over the military temperature range for the original NS part.

Given you *have* to draw the minimum load current or the output may rise out of regulation you may as well keep to the 240 ohms or so in most circumstances. If worst-case Vin-Vo is very high (especially with possible very low Ta) you may wish to go as low as half that.

onsdag den 29. december 2021 kl. 02.50.18 UTC+1 skrev Spehro Pefhany:

some manufacturers, I think ST is one of them, specify 3.5mA typical, 5mA max

Looks like that's only for the LM217. LM317 is 10mA max.

Similarly NS has the 117 at 5mA max, but the 317 and 317A are 10mA.

On a sunny day (Tue, 28 Dec 2021 15:15:28 -0800 (PST)) it happened Sid 03 snipped-for-privacy@gmail.com wrote in snipped-for-privacy@googlegroups.com:

It is not very critical really. It sometimes helps to use a bit different value to make a divider that gets to the required voltage. The reference can change between 1.2 and 1.3 V according to my datasheet. I am using 220 Ohm almost everywhere it seems, no 240 ohm available :-)

Although 240 ohm's a common cookbook current set resistance for a LM317 voltage regulator, LMx17 inventor Bob Dobkin often uses 120 ohms. [1] Presumably the drawback of a lower value is less efficiency and greater power loss. For an LM317 current regulator, current set resistors of a few ohms are used by both Bob and others. [2]

Note.

[1] Danke,