Post by thread:24VDC to 5VDC Supply

No matter what you do, you are going to get heat from power dissipation. Several solutions. DC to DC converter. Power resistor to drop the voltage (but this will be dependant on current draw), or several steps of step down regulation, say to a 15 volt, or a 12 volt to the 5V (or skip intermediate steps, again, depending on current draw). Have you measured the current being pulled? Do you have the 317 on a heat sink? {Quote hidden}> ---------- > From: Ryan Bloom[SMTP:.....RyanBKILLspam@spam@MPL.NET] > Sent: Tuesday, September 22, 1998 11:36 AM > To: PICLISTKILLspamMITVMA.MIT.EDU > Subject: 24VDC to 5VDC Supply - Slightly [OT] > > I need to regulate 24vdc down to 5vdc for a PIC project. > I've been using LM317's in TO-220 packages but they seem to go into > thermal > runaway when I draw any current. > Does anyone have any good methods/schematics for this. > > Thanks in advance, > > -Ryan > __________________________________________ > Ryan Bloom .....ryanbKILLspam.....mpl.net > Meridian Partners Limited > Internet Systems Group, Director > 425.688.9211 1.800.MPARTNER > http://www.mpl.net > "Your LINE On The World" >

>I need to regulate 24vdc down to 5vdc for a PIC project. >I've been using LM317's in TO-220 packages but they seem to go into thermal >runaway when I draw any current. >Does anyone have any good methods/schematics for this. You can't draw much current without a BIG heatsink. It's not in runaway, it's called thermal shutdown. You're dropping 19V across the reg, so your current in amps, times 19, gives you reg dissipation (aka wasted power) in watts. OUCH! Digikey has some switchers that are literally drop-in replacements for those regs, that will run nice and cool. Your wallet may feel the pinch though, at $12-$20 PT5101-ND will take 35V (max) in and deliver 1A out, with about .5-.6W dissipation. Compare that with 19W above. http://www.digikey.com

>Put a 12V or 15V zener to ground before the regulator. At 1A, this would have to be a 12 or 15W zener respectively, and then the zener needs heatsinking or it will self-desolder. This is just moving the problem. >Most regulators don't like to drop that much in one stage. ALL linear regulators suffer from the same problem. Current times (Vin - Vout) = Heat. There's no other option for linears. Sticking a zener in front of it just moves the problem. It's not a "preference", it's physics. http://www.digikey.com has some three-terminal regulators by Power Trends that are 85% efficient, so for the same 5W output, you only have to get rid of a fraction of a watt as heat. PT5101N-ND as an example.

Either get a bigger heat sink for the LM317 or go to a switcher. National has their "Simple Switcher" line, and SOMEONE (who I don't recall), has a switcher 7805 replacement (all in one piece. Finally, how much current are you trying to draw? The typical PIC application is pretty low power, so maybe the LM317 is oscillating instead just dissipating too much power. A 220nF to ground at both input and output can help. Also, any reason you're not using a 7805? It requires two less resistors than the LM317 (and the ideal design has zero parts...) . I see someone else has suggested dropping voltage prior to the regulator. They suggested a zener. The 7805 can take 35V input, so if you are indeed driving the regulator with +24V, that should be ok. The LM317 can take an input-output voltage differential of 40 volts, so you should be ok there. Adding the suggested zener in series with the input will move some of the dissipation to the zener, reducing the regulator heat sink size, but then requiring a heat sink on the zener. I don't think additional parts are the real solution here. Also, if the load current is relatively constant, you could use a resistor instead of the zener, again moving dissipation to the resistor from the regulator. However, I'd first check to see if the regulator is oscillating (look with a scope), then add a couple 220nF capacitors. Also, you can calculate the dissipation (about (Vin-Vout)*LoadCurrent), multiply by the thermal resistance of the heat sink (degreesCtempRise/Watt), add to ambient temperature and see where your case temperature lands. If too high, add heat sink or go to a switcher. Harold Harold Hallikainen EraseMEharoldspam_OUTTakeThisOuThallikainen.com Hallikainen & Friends, Inc. See the FCC Rules at http://hallikainen.com/FccRules and comments filed in LPFM proceeding at http://hallikainen.com/lpfm On Tue, 22 Sep 1998 10:36:15 -0700 Ryan Bloom <RyanBspam_OUTMPL.NET> writes: {Quote hidden}>I need to regulate 24vdc down to 5vdc for a PIC project. >I've been using LM317's in TO-220 packages but they seem to go into >thermal >runaway when I draw any current. >Does anyone have any good methods/schematics for this. > >Thanks in advance, > >-Ryan >__________________________________________ >Ryan Bloom @spam@ryanbKILLspammpl.net > Meridian Partners Limited > Internet Systems Group, Director > 425.688.9211 1.800.MPARTNER > http://www.mpl.net > "Your LINE On The World" > _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]

Although my options sound limited, thanks for all the help. I was hoping to avoid the distributed solution (BIG resistor or zener). I will look into the switching type and see if it is cost prohibited. BTW - I'm using a 317 because I need between 5 and 5.5vdc (I forgot to mention that). Thanks again, -Ryan -----Original Message----- From: Harold Hallikainen [KILLspamharoldhallikainenKILLspamJUNO.COM] Sent: Tuesday, September 22, 1998 1:08 PM To: RemoveMEPICLISTTakeThisOuTMITVMA.MIT.EDU Subject: Re: 24VDC to 5VDC Supply - Slightly [OT] Either get a bigger heat sink for the LM317 or go to a switcher. National has their "Simple Switcher" line, and SOMEONE (who I don't recall), has a switcher 7805 replacement (all in one piece. <Snip>

On Tue, 22 Sep 1998, Dave VanHorn wrote: > http://www.digikey.com has some three-terminal regulators by Power Trends > that are 85% efficient, so for the same 5W output, you only have to > get rid of a fraction of a watt as heat. PT5101N-ND as an example. Isn't there a bunch of schematics to make a 7805 a switcher with one transistor ? I've never used this, but it used to be in the National data books. Has anyone tried one of these as a fix on an existing design ? OTOH how about using a 555 as power oscillator at 1/4 duty cycle followed by a D, L and C to feed a normal series regulator with 8-9 V ? This is one that I've used when I had no choice (had to modify an existing design that had a closed plastic case and fried regulators on a regular basis). I don't know about passing FCC etc with it however. Peter

Check National Semiconductors simple switcher line. 8 pin dip, choke, diode, and caps. Should be able to get something running for under $6. Also, check Linear Tech for the same type of parts. We use these swithers for dropping anywhere from 6 to 35 volt inputs down to 5 volts with negligable heat. Roger Ryan Bloom wrote: {Quote hidden}> I need to regulate 24vdc down to 5vdc for a PIC project. > I've been using LM317's in TO-220 packages but they seem to go into thermal > runaway when I draw any current. > Does anyone have any good methods/schematics for this. > > Thanks in advance, > > -Ryan > __________________________________________ > Ryan Bloom spamBeGoneryanbspamBeGonempl.net > Meridian Partners Limited > Internet Systems Group, Director > 425.688.9211 1.800.MPARTNER > http://www.mpl.net > "Your LINE On The World"

>> http://www.digikey.com has some three-terminal regulators by Power Trends >> that are 85% efficient, so for the same 5W output, you only have to >> get rid of a fraction of a watt as heat. PT5101N-ND as an example. > >Isn't there a bunch of schematics to make a 7805 a switcher with one >transistor ? I've never used this, but it used to be in the National data >books. Has anyone tried one of these as a fix on an existing design ? There's something in the databook about using an external follower, basically, again, just moving the heat problem, not getting rid of it. >OTOH how about using a 555 as power oscillator at 1/4 duty cycle followed >by a D, L and C to feed a normal series regulator with 8-9 V ? This is one >that I've used when I had no choice (had to modify an existing design that >had a closed plastic case and fried regulators on a regular basis). I >don't know about passing FCC etc with it however. I'm not clear on how much power he needs, I was assuming large-ish due to his comments on heat. Anything with much current demand would blow the 555 approach, plus using transistors to switch large caps on and off isn't very efficient either. An inductor is the right tool here.

On Tue, 22 Sep 1998, Dave VanHorn wrote: > >Isn't there a bunch of schematics to make a 7805 a switcher with one > >transistor ? I've never used this, but it used to be in the National > data > >books. Has anyone tried one of these as a fix on an existing design ? > > There's something in the databook about using an external follower, > basically, again, just moving the heat problem, not getting rid of it. Nonono. I mean the one with the follower, the coil, the recovery diode and a resistive divider in the output to induce hysterezis. This is one of those weird designs where the regulator stays linear for low Vi/Vo and starts switching when current is demanded. {Quote hidden}> >OTOH how about using a 555 as power oscillator at 1/4 duty cycle > followed > >by a D, L and C to feed a normal series regulator with 8-9 V ? This > is one > >that I've used when I had no choice (had to modify an existing design > that > >had a closed plastic case and fried regulators on a regular basis). I > >don't know about passing FCC etc with it however. > > I'm not clear on how much power he needs, I was assuming large-ish due > to his comments on heat. > Anything with much current demand would blow the 555 approach, plus > using transistors to switch large caps on and off isn't very efficient > either. An inductor is the right tool here. I was suggesting the use of a 555 as switching regulator (series, open loop) before the actual reg. If the 555 is used to give 100 mA it should give about 300 mA available at the input of the linear regulator with the given voltage drop. Peter

>Nonono. I mean the one with the follower, the coil, the recovery diode and >a resistive divider in the output to induce hysterezis. This is one of >those weird designs where the regulator stays linear for low Vi/Vo and >starts switching when current is demanded. Eww.. I don't think I've seen that one. Deliberately destabilizing the reg with added phase shift is what it sounds like. It could be reasonable I guess. You'd get better efficiency at the low delta than a conventional switcher. >I was suggesting the use of a 555 as switching regulator (series, open >loop) before the actual reg. If the 555 is used to give 100 mA it should >give about 300 mA available at the input of the linear regulator with the >given voltage drop. An external switch here? The 555 gets fairly warm at high temps.

> 78XX switcher Ok, I'll try a famous ASCII schematic from memory here: Q1: PNP e c Coil Power O-*--V /----*---nnnnn----*------O Load + | === | A | \ | +-|<|-+ | R1 / | K | | \ | === | / | +-----+ | | | i | | o | +---*---* Reg.*--------* | | | +--*--+ \ g | / | R2 \ | / | | R3 +-----------*--/\/\/\--+ | === GND The circuit starts to oscillate when Ur1 > Ube Q1. R2 and R3 set the hysterezis of the O/P voltage, which can be as small as 25 mV but 50 mV is a good value in my experience. The output voltage is the nominal voltage of the regulator and increases to that + 1/2 the hysterezis when switching. The switching frequency must be low enough to be within the range of the 78XX. A 317 cannot be used here easily (as it has a constant Ignd). The O/P smoothing cap is not drawn in this schematic. It influences the switching frequency (together with the load current and Vi/Vo). In general, only regulators that sink the output transistor's base current through the gnd terminal can be used in this scheme. 78XXs are that. If R1 is choosen cunningly the SOA and thermal protections of the regulator will be maintained and also apply to Q1 (mount on common heatsink). Peter

Try using a switch mode regulator. You can use a simple switcher from National or look into Linear Technology. Linear Technology has a program called switcher Cad that will help you design a switching power supply very simply. At 10:36 AM 9/22/98 -0700, you wrote: {Quote hidden}>I need to regulate 24vdc down to 5vdc for a PIC project. >I've been using LM317's in TO-220 packages but they seem to go into thermal >runaway when I draw any current. >Does anyone have any good methods/schematics for this. > >Thanks in advance, > >-Ryan >__________________________________________ >Ryan Bloom TakeThisOuTryanbEraseMEspam_OUTmpl.net > Meridian Partners Limited > Internet Systems Group, Director > 425.688.9211 1.800.MPARTNER > http://www.mpl.net > "Your LINE On The World" > > Larry G. Nelson Sr. RemoveMEL.NelsonTakeThisOuTieee.org http://www.ultranet.com/~nr

It's called a DC-DC Converter -----Original Message----- From: pic microcontroller discussion list [EraseMEPICLISTMITVMA.MIT.EDU] On Behalf Of Michael Hagberg Sent: 25 September 1998 07:09 To: RemoveMEPICLISTEraseMEEraseMEMITVMA.MIT.EDU Subject: Re: 24VDC to 5VDC Supply - Slightly [OT] will a switcher supply regulate DC in to DC out? or do you need AC in? michael You may leave the list at any time by writing "SIGNOFF PICLIST" in the body of a message to RemoveMELISTSERVspam_OUTKILLspamMITVMA.MIT.EDU. -----Original Message----- From: Larry G. Nelson Sr. <RemoveMEnrTakeThisOuTspamMA.ULTRANET.COM> To: EraseMEPICLISTspamspamBeGoneMITVMA.MIT.EDU <RemoveMEPICLISTKILLspamMITVMA.MIT.EDU> Date: Thursday, September 24, 1998 8:53 PM Subject: Re: 24VDC to 5VDC Supply - Slightly [OT] >Try using a switch mode regulator. You can use a simple switcher from >National or look into Linear Technology. Linear Technology has a program >called switcher Cad that will help you design a switching power supply very >simply. > > >At 10:36 AM 9/22/98 -0700, you wrote: >>I need to regulate 24vdc down to 5vdc for a PIC project. >>I've been using LM317's in TO-220 packages but they seem to go into thermal {Quote hidden}>>runaway when I draw any current. >>Does anyone have any good methods/schematics for this. >> >>Thanks in advance, >> >>-Ryan >>__________________________________________ >>Ryan Bloom ryanbSTOPspamspam_OUTmpl.net >> Meridian Partners Limited >> Internet Systems Group, Director >> 425.688.9211 1.800.MPARTNER >> http://www.mpl.net >> "Your LINE On The World" >> >> >Larry G. Nelson Sr. >spamBeGoneL.NelsonSTOPspamEraseMEieee.org >http://www.ultranet.com/~nr >

At 09:35 25/09/98 +0100, you wrote: >It's called a DC-DC Converter > >-----Original Message----- >From: pic microcontroller discussion list [KILLspamPICLISTspamBeGoneMITVMA.MIT.EDU] On {Quote hidden}>Behalf Of Michael Hagberg >Sent: 25 September 1998 07:09 >To: EraseMEPICLISTEraseMEMITVMA.MIT.EDU >Subject: Re: 24VDC to 5VDC Supply - Slightly [OT] > >will a switcher supply regulate DC in to DC out? or do you need AC in? > >michael > >You may leave the list at any time by writing "SIGNOFF PICLIST" in the > body of a message to @spam@LISTSERV@spam@spam_OUTMITVMA.MIT.EDU. > >-----Original Message----- >From: Larry G. Nelson Sr. <spamBeGonenrKILLspamMA.ULTRANET.COM> >To: .....PICLISTspam_OUTMITVMA.MIT.EDU <TakeThisOuTPICLIST.....TakeThisOuTMITVMA.MIT.EDU> >Date: Thursday, September 24, 1998 8:53 PM >Subject: Re: 24VDC to 5VDC Supply - Slightly [OT] > > >>Try using a switch mode regulator. You can use a simple switcher from >>National or look into Linear Technology. Linear Technology has a program >>called switcher Cad that will help you design a switching power supply very >>simply. >> >> >>At 10:36 AM 9/22/98 -0700, you wrote: >>>I need to regulate 24vdc down to 5vdc for a PIC project. >>>I've been using LM317's in TO-220 packages but they seem to go into >thermal >>>runaway when I draw any current. >>>Does anyone have any good methods/schematics for this. >>> >>>Thanks in advance, >>> >>>-Ryan >>>__________________________________________ >>>Ryan Bloom TakeThisOuTryanbKILLspamspammpl.net >>> Meridian Partners Limited >>> Internet Systems Group, Director >>> 425.688.9211 1.800.MPARTNER >>> http://www.mpl.net >>> "Your LINE On The World" >>> >>> >>Larry G. Nelson Sr. >>.....L.NelsonRemoveMEieee.org >>http://www.ultranet.com/~nr >> > >