Troubleshooting LM317T Heating in Variable Power Supply - Tips and Tricks

  • Thread starter Thread starter Hitesh
  • Start date Start date
  • Tags Tags
    Heating
AI Thread Summary
The discussion focuses on troubleshooting overheating issues with the LM317T in a variable power supply setup, where the user experiences protection mode activation after five minutes under load. Key suggestions include using a larger package and heatsink for the LM317T, as well as considering a Buck DC-DC converter for better efficiency and reduced heat generation. Participants highlight that the linear regulator approach is inefficient, leading to significant heat dissipation, and recommend redesigning the circuit to achieve variable output voltage with a Buck converter. Additionally, there are mentions of alternative configurations and components that could improve performance and reliability. Overall, the conversation emphasizes the need for better thermal management and efficiency in the power supply design.
Hitesh
Hello . I am trying to make power supply using LM317T. As i need Variable output voltage from 14 V to 27.6 V.
I made a power supply of 29V/1A using Power Integration IC LNK6775E. At the output of 29V i am using LM317T to vary voltage.
But after 5 minutes IC is going in protection mode if apply a load of 1A and set output voltage to 14V.
I am also using Heatsink, but there is too much heat.
How to control this heat.
And is it reliable ?
 

Attachments

  • IMG20171011173325.jpg
    IMG20171011173325.jpg
    35.3 KB · Views: 563
Engineering news on Phys.org
Welcome to the PF. :smile:

15W is a lot of heat to get rid of. What package is your LM317 in? When you say LM317T, does that imply you are using a TO-220 package? You probably need a larger package than that, with a bigger heatsink. Using a linear regulator at that current level and that large of a voltage drop may not be the best approach.

Have you looked at using a Buck DC-DC converter for the output regulator instead?
 
  • Like
Likes Hitesh, davenn, jim hardy and 1 other person
15 volt drop at 1 amp is 15 watts
To hold temperature rise to 100 degrees you need thermal resistance junction to ambient of 100deg/15watts = 6.67 deg/w or less
and observe 5 of those are used up internally
leaving only 1.67 for case to ambient which is your heatsink.
http://www.st.com/content/ccc/resou...df/jcr:content/translations/en.CD00000455.pdf
upload_2017-10-11_12-31-33.png


to keep temperature rise less than 100 degrees you need a heatsink that can beat 1.67 deg/watt.
Here's a snip from
http://www.st.com/content/ccc/resou...df/jcr:content/translations/en.CD00000455.pdf
nov-heat-sink-fig-2.jpg


about a square foot of 1/8 " aluminum should do nicely.

Alternatively you might add resistance in series with the IC when asking it to handle a lot power...
A BIG light bulb might work well.
Got an old halogen Torchiere lamp ?
upload_2017-10-11_12-44-8.jpeg
 
  • Like
Likes Hitesh, dlgoff and berkeman
You also made a regulator that's 50% efficient, which is not good.

Why not just use the LNK6775E for your voltage regulation. I'm assuming you are using a potentiometer or similar variable resistor in the feedback path on your linear regulator. Just include the potentiometer on your LNK6775E instead. you already have feedback resistors, that how you get the 29 volts. just redo your calculations to determine the resistance swing and tolerance you need.
 
  • Like
Likes jim hardy, Hitesh and berkeman
berkeman said:
Welcome to the PF. :smile:

15W is a lot of heat to get rid of. What package is your LM317 in? When you say LM317T, does that imply you are using a TO-220 package? You probably need a larger package than that, with a bigger heatsink. Using a linear regulator at that current level and that large of a voltage drop may not be the best approach.

Have you looked at using a Buck DC-DC converter for the output regulator instead?
Thanks for reply. But i need variable output voltage from 14 V to 28 V. By using Buck converter , it will decrease output voltage. And can i get variable output voltage using Buck DC DC converter? or what circuitry i need to be added?
 
For just a one-of-a-kind home project,
Try a search on "LM317 as switching regulator"

the old National datasheet from before TI took them over had this image
upload_2017-10-12_10-15-35.png


i never tried it so don't know if it's well behaved.
 
  • Like
Likes davenn and dlgoff
jim hardy said:
i never tried it so don't know if it's well behaved.
I don't get it. What switches?
 
berkeman said:
I don't get it. What switches?

If i understand the circuit,

this positive feedback path gives it enough hysteresis that Mr LM317 will oscillate between full-on and full-off.

lm317Switcher.jpg


For lower current i wonder if the inductor with a series diode could be moved down to R2's position , eliminate Q1 & R1 ?
I don't know, just speculating.

As i say i never built one. Looks like an easy introduction to switchers and inductor sizing, though.

old jim
 
  • Like
Likes donpacino
  • #10
I can't recall the exact circuit topology but I have seen a switcher built around a 317. From what I recall it was a pretty reliable setup.
 
  • #11
Why build when you can buy and have it delivered for less than $2.

http://www.ebay.com.au/itm/Voltage-Regulator-LM2596-3A-Adjustable-DC-To-DC-Step-Down-Converter-Buck-Module/232274773224?

It should handle one amp without a heatsink.
Replace the trim-pot on the PCB with an external pot of same resistance on three short twisted wires.
 
  • Like
Likes jim hardy
Back
Top