Dear All,
I am working on failure analysis of products.
In one faulty module we use LM317 for 5.17 V generation.
The reported problem is that the output voltage is not observed and sometimes it is observed (During vibration)
After analysis it is observed output pin of LM317 is not soldered so sometimes it is get connected and sometimes it is not.
But problem is that when output pin is firm with soldering pad that time at the output voltage observed 10.48V instead of 5.17V.
When output not observed that time output voltage is occurred 0.25V.

When I soldered output pin then everything is working fine and observed 5.17 v exact.

Now i want to know why this happen in that instance I got sometime 10.48 V and sometime 0.25V.

I attached herewith snap of circuit.
Please help me friends it is very interesting problem.
your suggestions and comments regarding this problem is must.

 

On which node is the solder joint failing?

 

output node
Any idea what exactly happen

 

I cannot read clearly but it looks like you have a 47 microfarad cap on your 150 ohm resistor. This seems really big and means your LM 317 will be flip-flopping output voltage until the cap stabilizes. It may be latching up and going into protection mode - just a guess.

If it happens again, try removing that cap or making it reasonably small (0.1uF).

It may also depend on your load so I cannot determine why 10 volts in some conditions. No use diagnosing the past if a clean solder joint works now.

 

I just can't get enthusiastic about this one. "I assembled it wrong and it worked wrong. Now I fixed it and it works right. Why did it have random, wrong outputs when it was assembled wrong?"

Because it was assembled wrong.

 

I just can't get enthusiastic about this one. "I assembled it wrong and it worked wrong. Now I fixed it and it works right. Why did it have random, wrong outputs when it was assembled wrong?"

Because it was assembled wrong.

Yeah, I spent 20 years diagnosing these type of problems. They give you a schematic.... except that a circuit error or manufacturing defect means it is not actually connected as shown or is intermittently connected or in many cases, something was touching something it shouldn't making a connection not intended.

good luck

 

I just can't get enthusiastic about this one. "I assembled it wrong and it worked wrong. Now I fixed it and it works right. Why did it have random, wrong outputs when it was assembled wrong?"

Because it was assembled wrong.

Dear friend I asked reason behind why 10.48 v observed sometimes and why 0.25 V observed sometimes.

If I pull up output pin and measure voltage at PCB output pad then I got 10.48 v in this condition I never found 0.25 V at any instance at PCB output pad

 

When the output of the regulator is not connected there is still a path from the 24 volt supply through the 150 ohm resistor.

 

Good catch! Using a load resistor on the test setup would eliminate that issue because it will load the circuit more than the ADJ pin and 150 ohm resistor can supply.

Alternatively, if the open circuit pin is the ADJ pin instead, the LM317 will turn hard on and put out close to the full supply voltage even with a load.

 

When the output of the regulator is not connected there is still a path from the 24 volt supply through the 150 ohm resistor.

I agree with you friend but again question is there.
Why sometimes 0.25V and sometimes 10.48V

 

I agree with you friend but again question is there.
Why sometimes 0.25V and sometimes 10.48V

Because you have it connected incorrectly one way to get 0.25V and you have a different incorrect way to get 10.48 volts.

As I said in my first post...
1) we would need to know what your load is so we can calculate how the voltage is divided.
2) it doesn't matter because it was incorrect anyhow. If you disagree, carefully follow the circuit paths and use ohms law to calculate the voltage dividers and, from that, the voltage at each node in your circuit. Just remember, the LM317 will drop 1.25 volts across the adjust to out pin (if load is 120 ohms or less) and assuming the input is 2 volts above the output (google "drop-out voltage").

 

Instead of simulating, why not breadboarding the circuit?

Disconnect every component and check when you get those values.

Never say never...but I do not feel I would come to do that ever.

I tend to agree with Nro. 12:

I just can't get enthusiastic about this one. "I assembled it wrong and it worked wrong. Now I fixed it and it works right. Why did it have random, wrong outputs when it was assembled wrong?"

Because it was assembled wrong.