transisitor testing

Mr Do

Well-known member

Donor 6 years: 2012-2013, 2015-2016, 2019, 2025
Joined
May 29, 2006
Messages
1,394
Reaction score
32
Location
Basking Ridge, New Jersey
Can a lm323k 5volt regulator be tested in circuit?
How is it done?
 

Attachments

  • 2011-12-14 16.19.52.jpg
    2011-12-14 16.19.52.jpg
    76.1 KB · Views: 5
Last edited:
A reg is a different creature to a transistor

It keeps voltage at a constant level despite differing loads. So it is not NPN or PNP

It is positive or negative. If there is no constant voltage coming from the output of a regulator, then it is likely failed. I needs to be tested in the circuit, while the input power and related components are applied around it.

A load is not critical, so if there is a load that has shorted and it keeps blowing the reg, then you may remove the load/s and test the reg without the load, it will still provide a constant voltage

In your case the LM323 is a 5 Volt positive regulator
 
That is an IC... a chip... not a transistor...

And if it doesn't have some small capacitors close to the inputs and outputs then it can oscillate and fail to maintain a good regulated voltage.

Also, if it doesn't have an adequate heat sink it will get NOWHERE near it's rated output current as it will go into thermal shut down.

It either outputs the right voltage or it doesn't. If it doesn't, look for broken/bad caps right next to it... heatsink issues... or just replace it. They are very cheap.
 
It may be obvious, but also check that the voltage going into the regulator is good (above its dropout voltage), or else it won't be able to output the correct voltage.

Here's a good overview of regulators:

 
Last edited:
How does a non-EE figure out when to add a heatsink? And what kind of heat sink?

Not an easy task.
First, you must figure out how much power the device is dissipating. That's the easy part.

Then you need to know what your final desired temperature is, factor in the part's junction to case thermal resistance (theta JC) and the heatsink compount (case to sink) thermal resistance (theta CS) to arrive at the required heatsink (sink to ambient) thermal resistance (theta SA). Then flip through thousands of available heatsinks for the appropriate thermal resistance....

There is no easy route.
 
Typically a 7805 wont need a heatsink if the circuit is designed correctly. They are thermally protected in case they do get overloaded so it will shut itself off until it cools back down. If you see it doing this you either have too much input voltage or the regulator is ready to fail. These do not go bad very often. You can test it in circuit by looking for +5 on the far right leg(iirc).
 
Last edited:
Typically a 7805 wont need a heatsink if the circuit is designed correctly. They are thermally protected in case they do get overloaded so it will shut itself off until it cools back down. If you see it doing this you either have too much input voltage or the regulator is ready to fail. These do not go bad very often. You can test it in circuit by looking for +5 on the far right leg(iirc).

Uh... yeah it will...

If you want more than 100 to 200ma of current out of it you need the heatsink. Too small of a heatsink and it will shut down when it gets too hot.

To get anywhere close to its rated output current of 1A you need a heatsink of the proper size.
 
It depends. If you put 8V into a LM7805 and pull 100ma out then the 5V regulator is acting like a 300 mW heater. Not so bad. It might handle that without overheating. But if you put 35V in and take 100ma out it's a 3 Watt heater. It will almost certainly shut itself off without a heatsink of some type (3W and the 7805 temperature has to stay below 125C). If you put 35V in and take an Amp out (these are the rated maximums) you need a big heatsink to safely dissipate the 30 Watts of heat coming off the 7805.

[EDIT]
Whoops. Nevermind...channelmanic beat me to it. I should read the entire thread before posting.
 
Bit of an example based on the original poster's LM323K.

This part has an RthJa of 35C/W.
If ambient temperature is 25C and using a max temp of 100C (don't touch that!).
You have 100 -25 = 75C leadway. Using the 35C/W then you can dissipate a hair over 2W before things become too hot.

If the input voltage is 12V and the regulator is outputting 5V --> 7V difference.

2W / 7V = 0.285 amps max without a heatsink before it starts cooking itself.
Pretty bad for a part rated at 3 amps.

The TO-220 package parts (i.e. LM7805CT) are far worse.
Using same temps and voltages for the 7805 - you can only dissipate 1.5W tops with the 7805 which gives you about 200mA tops before it starts overheating.

Ed
 
Last edited:
Mr Do, sorry for hijacking your thread. But this is all great information. It explains why the 7905 in my Robotron power supply can't put out 600 mW of power for my Mario Brothers PCB with the Nintendo to JAMMA adapter. I will look at an alternative approach.
 
Back
Top Bottom