Loading ...
Sorry, an error occurred while loading the content.
 

LM317 thrmal considerations

Expand Messages
  • lists
    Recently designing a small PSU (not done one for a while but wanted a non-standard V-out) I chose an LM317. I worked out the design and the final stage was
    Message 1 of 7 , Jul 20, 2014
      Recently designing a small PSU (not done one for a while but wanted a
      non-standard V-out) I chose an LM317. I worked out the design and the
      final stage was choosing a heatsink.

      I looked at a data sheet (Texas instruments- revised 2013, downloaded off
      the net) and it gave the thermal resistance, junction to case, for a TO220
      package at 19 deg C/watt.

      I thought "That's ridiculous, even with only a 3V forward drop it can
      barely produce its rated 1.5A".

      In my circuit, at maximum load, the calculated power dissipation would
      have been 8.95W (call it 10W). With an ambient of 30C, I'd have had a
      junction temperature of 220C!

      I could not believe that anyone would design a power regulator in such a
      way.

      After redesigning using an LM723, with 2N3055 external pass transistor, (I
      have loads of them) I had an idea and, after a rummage around in the loft,
      found an old Natsemi databook which gives a far more realistic 4deg C/watt
      and a maximum dissipation of 15W.

      TI aren't exactly a "tinpot" company so what's going on?

      I know I will have to look carefully at the manufacturers of all my LM317s

      --
      Tools With A Mission
      sending tools across the world
      http://www.twam.co.uk/
    • Dave Neale
      I think you fill find that is the thermal rating without a heatsink. If you add one the picture will change dramatically. Regards, Dave Neale Managing Director
      Message 2 of 7 , Jul 20, 2014

        I think you fill find that is the thermal rating without a heatsink. If you add one the picture will change dramatically.

        Regards,

         

        Dave Neale

        Managing Director

        Revelation Designs (Pty) Ltd.

        Work - +27 (0)11 312 5646

        Cell    -  084 542 3037

         

      • lists
        In article , lists Stuartlists@orpheusinternet.co.uk [Electronics_101] ... Correction. Junction to case is
        Message 3 of 7 , Jul 20, 2014
          In article <542a2fda1fStuartlists@...>,
          lists Stuartlists@... [Electronics_101]
          <Electronics_101@yahoogroups.com> wrote:
          > I looked at a data sheet (Texas instruments- revised 2013, downloaded
          > off the net) and it gave the thermal resistance, junction to case, for a
          > TO220 package at 19 deg C/watt.

          Correction. Junction to case is 17C/watt, junction to ambient is specified
          as 19C/Watt, which in itself is strange as I cannot see a TO220 pack
          having a case to ambient of 2C/watt

          --
          Tools With A Mission
          sending tools across the world
          http://www.twam.co.uk/
        • Andy
          I just glanced at Fairchild s LM317 datasheet in a TO220 case. Junction to case: 5 C/watt Junction to ambient: 80 C/watt (I assume that means with no
          Message 4 of 7 , Jul 20, 2014
            I just glanced at Fairchild's LM317 datasheet in a TO220 case.

               Junction to case: 5 C/watt
               Junction to ambient: 80 C/watt  (I assume that means with no heatsink)

            TI's datasheet (LM117/LM317A/LM317-N, TO220):

               Junction to case: 4 C/watt
               Junction to ambient: 50 C/watt  (datasheet says no heatsink)

            TI's *other* datasheet for the LM317 has the same numbers you quoted.  I don't know what they've been smoking.

            Regards,
            Andy


          • lists
            In article , Andy ai.egrps@gmail.com [Electronics_101] ... Thanks for the confirmation that
            Message 5 of 7 , Jul 21, 2014
              In article
              <CAN_saP+2sSsv-LbO13hq2f+z0A1vdi+gYXQKd9rwG7XNFq5V-A@...>,
              Andy ai.egrps@... [Electronics_101]
              <Electronics_101@yahoogroups.com> wrote:
              > I just glanced at Fairchild's LM317 datasheet in a TO220 case.

              > Junction to case: 5 C/watt Junction to ambient: 80 C/watt (I assume
              > that means with no heatsink)

              > TI's datasheet (LM117/LM317A/LM317-N, TO220):

              > Junction to case: 4 C/watt Junction to ambient: 50 C/watt (datasheet
              > says no heatsink)

              > TI's *other* datasheet for the LM317 has the same numbers you quoted. I
              > don't know what they've been smoking.

              Thanks for the confirmation that I'm not going crazy.

              What you have also highlighted is the importance of checking the datasheet
              from the manufacterer who's part you are using. Tja figures of 80C/watt
              and 50C/watt are significantly different.

              Now I've done the 723/2N3055 design I'll run with it because I can set the
              current limit to be exactly what I want and, because the 2N3055 has only
              1.5C/watt thermal resistance, I have a greater choice of heatsink.

              --
              Tools With A Mission
              sending tools across the world
              http://www.twam.co.uk/
            • Donald H Locker
              Is it possible the other datasheet is for a part with an isolated package? Isolated package (isolated tab, in the case of a TO-220) has no electrical
              Message 6 of 7 , Jul 21, 2014
                Is it possible the "other" datasheet is for a part with an isolated package? Isolated package (isolated tab, in the case of a TO-220) has no electrical connection to the package mount, but incurs a high penalty in the heat transfer department.

                Donald.

                ----- Original Message -----
                > From: "Andy ai.egrps@... [Electronics_101]" <Electronics_101@yahoogroups.com>
                > To: "Electronics 101" <Electronics_101@yahoogroups.com>
                > Sent: Sunday, July 20, 2014 11:30:05 PM
                > Subject: Re: [Electronics_101] LM317 thrmal considerations
                >
                > I just glanced at Fairchild's LM317 datasheet in a TO220 case.
                >
                > Junction to case: 5 C/watt
                > Junction to ambient: 80 C/watt (I assume that means with no heatsink)
                >
                > TI's datasheet (LM117/LM317A/LM317-N, TO220):
                >
                > Junction to case: 4 C/watt
                > Junction to ambient: 50 C/watt (datasheet says no heatsink)
                >
                > TI's *other* datasheet for the LM317 has the same numbers you quoted. I
                > don't know what they've been smoking.
                >
                > Regards,
                > Andy
                >
              • lbeaty
                I generally speed wire up a project and use an LM317 or LM7805 TO220 package. Then I feel the TO220 for temp. If really hot, I just bolt a piece of aluminum
                Message 7 of 7 , Jul 21, 2014

                  I generally speed wire up a project and use an LM317 or LM7805 TO220 package.  Then I feel the TO220 for temp.   If really hot, I just bolt a piece of aluminum to the tab.   Cools it right down.   After all, this is just a prototype.   When laying out the PC board, I always put the TO220 on the edge so it can be bolted to the chassis below the board or to a side chassis area.

                   

                  Larry

                   

                  From: Electronics_101@yahoogroups.com [mailto:Electronics_101@yahoogroups.com]
                  Sent: Monday, July 21, 2014 4:19 AM
                  To: Electronics_101@yahoogroups.com
                  Subject: Re: [Electronics_101] LM317 thrmal considerations

                   

                   

                  In article
                  <CAN_saP+2sSsv-LbO13hq2f+z0A1vdi+gYXQKd9rwG7XNFq5V-A@...>,
                  Andy ai.egrps@... [Electronics_101]
                  <Electronics_101@yahoogroups.com> wrote:

                  > I just glanced at Fairchild's LM317 datasheet in a TO220 case.

                  > Junction to case: 5 C/watt Junction to ambient: 80 C/watt (I assume
                  > that means with no heatsink)

                  > TI's datasheet (LM117/LM317A/LM317-N, TO220):

                  > Junction to case: 4 C/watt Junction to ambient: 50 C/watt (datasheet
                  > says no heatsink)

                  > TI's *other* datasheet for the LM317 has the same numbers you quoted. I
                  > don't know what they've been smoking.

                  Thanks for the confirmation that I'm not going crazy.

                  What you have also highlighted is the importance of checking the datasheet
                  from the manufacterer who's part you are using. Tja figures of 80C/watt
                  and 50C/watt are significantly different.

                  Now I've done the 723/2N3055 design I'll run with it because I can set the
                  current limit to be exactly what I want and, because the 2N3055 has only
                  1.5C/watt thermal resistance, I have a greater choice of heatsink.

                  --
                  Tools With A Mission
                  sending tools across the world
                  http://www.twam.co.uk/

                Your message has been successfully submitted and would be delivered to recipients shortly.