[SOLVED] Problem with LM317 Output Voltage

I am using an LM317 to provide an output voltage varying between +1.5 to 17 Volts, for use in a DC power supply.

The input to the IC is from across a 4700uF capacitor and is about 35V.

I used a 390ohm resistor along with a 5k variable resistor.

Now my problem is that when the IC is powered by the main it only provides an output in the range of +8.9 to 17 Volts (with no load attatched) but whenever I turn off the power to the IC, it provides the desired output between 1.5 to 17 volts. (Since the capacitor is large and there is no resistive load attatched to the circuit, the capacitor takes an appreciative time to discharge during which I am able to vary the value of the variable resistor to see that the IC is providing the desired output). I have also tried using a combination of 270 ohm resistor with a 3k variable resistor but faced a similiar result.

Any suggestions would be very helpful.

Are you sure about the connections, can you post the schematic?

You are using this right? **broken link removed**
double check the pinout of LM317

Alex

This is the circuit I am using, and yes I did double check my pinouts, still the result stays the same i.e. LM317 regulates from 8-9 to 17 volts

Datasheet recommends a small bypass capacitor across Vin to ground (read: parallel to the 4700uF you mention), for example 100nF ceramic. Also the 390 Ohm is a bit on the high side, datasheet recommends 240 Ohm (this to achieve minimum load of ~ 5mA), but 390 Ohm should still work I think.

And 35V is very high input if you need 17V max output - you might run into thermal problems when you start loading this circuit.

Other than that it should work with circuit shown. So either you don't have things connected as circuit shows, or the LM317 you have is broken.

EDIT: you don't happen to have it screwed onto a cooling plate that's connected to housing, or something like that :?: Cooling tab is normally connected to center pin (=Vout for LM317 in TO-220 package).

I second RetroTechie
The pinout of LM317 is NOT the same as the 7805 etc.
Check the pinout from the datasheet,then your connections. The problem is likely to be there.
The chances of a bad 317 are usually low.

Problem Solved. I changed the IC, lowered the input voltage to 25V and used a 3300uF capacitor.

My conclusion is that the problem was caused by a faulty IC plus using a high voltage as the input.

Thank you all for your suggestions.

I still cant understand though that why use a bypass capacitor,other than the fact the datasheet mentions it, after the filter capacitors? Isn't the filter capacitor able to act as a bypass capacitor on its own, owing to its large capacitance?

The bypass cap becomes more imp in -ve regulators.
If one removes this cap from 7905 (-5V) regulator, there is erratic output with too much of oscillations at the output. A simple 100n stops this and every thing is restored
Raoof Khan

hassanzia said:

I still cant understand though that why use a bypass capacitor,other than the fact the datasheet mentions it, after the filter capacitors? Isn't the filter capacitor able to act as a bypass capacitor on its own, owing to its large capacitance?

Click to expand...

Frequency response is very different for a large electrolytic capacitor like 3300uF and a small capacitor like 100nF ceramic or a few uF tantalum. For example the crackling noise you often hear on speakers when you switch on or off equipment, has lots of high-frequency components that a 3300uF would hardly filter, but a 100nF ceramic would. More on that here.

So you put a few different types capacitors in parallel, which makes the LM317 less likely to behave weird as a result from noise on its input.

Good to hear you figured out what was wrong... :smile:

ark5230 said:

If one removes this cap from 7905 (-5V) regulator, there is erratic output with too much of oscillations at the output. A simple 100n stops this and every thing is restored
Raoof Khan

Click to expand...

Can you please explain the reason for this erratic behaviour? I'm just being curious as isn't the regulator IC supposed to prevent exactly this sort of thing happening, to some extent, at least?

What has been shown above, is perfect and explains what you need. Why these capacitors are not built into the design of the regulator is a question. The obvious reasons of size and material can be understood. If there is something special and extraordinary in this relation, experts would be able to point.
Raoof Khan