# Could LM317 in basic regulator configuration (see description) fail to regulate?

So I have a regulator using a 317 in a basic configuration. One minor change is the introduction of a 10uF electrolytic capacitor between $R_{ref}$, and ground.

$$R_{ref} = 240 \Omega$$ $$R_{adj} = 1k \;\text{linear} \times 2 \;\text{(series)}$$

$$V_i = 15.1V$$

$$V_o = 1.25 \Big(1+\frac{R_{adj}}{R_{ref}}\Big)$$

Any value of $R_{adj}$ below $R_{ref}$ should give me 1.25V.

Values of $R_{adj}$ above $R_{ref}$ hopefully will yield correspondingly increasing values at $V_o$.

What I see instead is as follows -

There is no load connected as yet. This supply being constructed to bias a varactor; the output reported below were against ground.

• $V_o=15.1V$ for $R_{adj} = 0 \Omega$
• $V_o=13.9V$ for $R_{adj} = 2K \Omega$

Do I have a bad 317?

EDIT: Turns out a loose connection was the problem. I flipped the perforated board over to take a gander at the soldering; when I flipped it back I could vary the voltage using the pots. Thank you, all!

• Can you show the exact circuit you have including the power supply you are feeding the LM317 with and any load connected to the output. – Andy aka Nov 16 '14 at 13:03
• I'll have to free-hand a sketch; couldn't find a SPICE model for the 317 to use in LTSPICE. In an hour? – Everyone Nov 16 '14 at 13:06
• I'm measuring the output against ground – Everyone Nov 16 '14 at 13:22
• Try doing this properly - exact circuit, power supply and load. show where 0V is and indicate the input voltage feeding the regulator. Take your time, no hurry. – Andy aka Nov 16 '14 at 13:35
• The LM317 is a fairly bullet proof part. You schematic looks fine, but check the pin numbers yet one more time. (I think they are different between the to-92 and the to-220) And then try a new part. – George Herold Nov 16 '14 at 15:36

The cause MAY be inadequate minimum load current - see below. Easily checked and easily fixed if this is what is wrong.

As specified in the LM317 datasheet on page 3 and in figures 2, the LM317 needs a minimum output current to regulate correctly. This is 10 mA worst case at Vin-Vout = 40V differential, 3.5 mA typical. At 15V or so differential it drops to about 1.5 mA typical and one can guesstimate it may be around 4 or 5 mA wort case.

The output current is used to power the internal elecronics, unlike earlier generation device which uses Vin to Vref voltage for powering.

In your case with no external load Iout min is about Vref/Rref = 1.25V / 240 Ohms or about 5 mA. So you are on the upper bound of worst case so in mostinstances would expect to be OK - JUST. But, as a worst case spec is not given officially at the voltage it could be closer to 10 mA than I estimated.

Using a lower Rref or a permanent minimal load resistor will soon show if this is the cause of your problem.

Connection seems a likely cause - see below.
Next most likely is a faulty IC.

With Adj grounded (lh pin) and Vin on rh pin and say 100R from centre to lh pin you should see 1.24 ish Volts from centre to lh pins.

Your 15.1V and 13.9V are 1.2V apart - which makes it sound as if you are measuring across the Vo-Vref pins.

Dim memory suggests that pinout MAY vary between TO220 and smaller pkg parts in some cases (pun noted). Check data sheet for actual part upou are using AND tell us what part is and give DS link. Standard pinout in almost all parts as per eg LM317 datasheet is left to right pins down, label side towards you AOI = Adjust Output Input .

Same for TO92 and TO89.

Vin on right hand pin
Two adjust resistors on left hand pin
Voltage out on centre pin.

Worth noting: On almost all regulators in 3 pin packages with a metallic tab the centre pin is the most negative voltage rise. So on an eg
7805 +5V regulator this is ground, on a
LM317 this is adjust and on a LM7905 -ve regulator this is Vin.
This allows substrate to be electrically connected to tab.
If the IC uses eg Silicon on Sapphire (you should be so lucky) all bets are off, but convention may be followed.

• If the resistor was a bit too high, it would 'regulate' at a slightly higher voltage. 13.9V into 240R is 58mA. It has always bugged me that the 240R is shown on the datasheet example schematics when it is clearly not guaranteed to work properly with no load (Fairchild ones can be as much as 12mA, 3.5mA typical), but I don't think it's the culprit in this case. – Spehro Pefhany Nov 16 '14 at 18:05

Part wired wrong (most likely) or bad part (less likely) I would say.

Nothing else would give you Vo = 13.9V with Radj = 0.

• I measured resistance between VAdj leg & ground varying the linear - the change reflected in the multimeter. Similarly Vout/VAdj resistance 240 also reflected clearly. I'm loath to contemplate the idea of a bad part but ... – Everyone Nov 16 '14 at 18:24
• What about the wiring?? Mixing the pins up could give you an output that's a diode drop or two below the input. – Spehro Pefhany Nov 16 '14 at 18:27
• As Spehro noted, the LM317 has a pinout somewhat different from what you might expect – EM Fields Nov 16 '14 at 18:58

From the datasheet of LM317 you should use 0.1 uF or 1 uF baypass capacitor at the input, the device is more sensitive to the absence of input bypassing when adjustment or output capacitors are used but the above values will eliminate the possibility of problems.

I suggest to you also to protect the regulator like using some 1n4001 diodes so that the regulators cannot be damaged by any reverse voltages.

• The 2200 uF is actually part of the rectifier ... I'll get the schematic done the way Andy suggested so it is easier for everyone (else) around – Everyone Nov 16 '14 at 13:54