lm338 regulator programming resistors
in [Design]
|
Prev: pi/e, was Re: Relay Question
Next: Changing the color of objects/primitives only ? (flat shading...) (massive parallel lookup hardware idea...)
From: Phil Hobbs on 29 Jul 2010 21:53 David Eather wrote: > On 29/07/2010 3:27 PM, Phil Hobbs wrote: >> Grant wrote: >>> On Wed, 28 Jul 2010 19:26:10 -0700 (PDT), Scott <smbaker(a)gmail.com> >>> wrote: >>> >>>> The typical LM338 power supply circuit in the application notes uses a >>>> fixed 120 ohm resistor and a variable 2k pot. The LM338 resistance >>>> formula is Vout = 1.25 * (1+ R2/R1). With R2 as 2k and R1 as 120 ohms, >>>> this yields an adjustment range from 1.2 to 22 volts. I'm building a >>>> power supply and want to increase my range to around 30V or so. >>>> >>>> To get the full range I would need a 2.7k pot. Unfortunately, all I >>>> have on hand are 2k and 5k. Using a 5k pot means the voltage will be >>>> maxxed out somewhere just over half the range of the pot. >>>> >>>> So, what I'm wondering is what happens if we substitute a larger >>>> resistor for R1. a 5K pot with a 220 ohm resistor would be just about >>>> right. I've read the datasheet, but I haven't quite figured out what >>>> effect this would have. I've seen some web articles that mention using >>>> resistors as high as 240 ohms, but they aren't specific about what >>>> side effect this may have. >>> >>> You could place a resistor in parallel with the 5k pot to bring it >>> down to 2.7k too. Always more than one solution. >>> The 120 Ohm set resistor has the advantage of soaking up the bias >>> current from the regulator. You can up that, provided you have a >>> minimum load for regulation. Add an output LED or something to eat the >>> bias or quiescent current. >>> >>> Grant. >>>> Thanks, >>>> Scott >> >> If you add an op amp, you can make the resistances anything convenient. >> Alternatively, if you don't mind a bit of nonlinearity, you can load >> down the pot with a 5.90k resistor in parallel. Or, of course, use a 2k >> pot and an 86.6 ohm resistor from output to feedback. >> >> Cheers, >> >> Phil Hobbs >> > > A serious question. You didn't suggest changing R1. Is there a > consideration I am missing? R1 is from output to feedback--change it to 86.6 ohms and use a 2k pot. It'll waste a bit more power, but it'll give you the right range. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
From: Grant on 29 Jul 2010 23:57
On Thu, 29 Jul 2010 17:32:50 -0700 (PDT), kevin93 <kevin(a)whitedigs.com> wrote: >On Jul 28, 7:26 pm, Scott <smba...(a)gmail.com> wrote: >> The typical LM338 power supply circuit in the application notes uses a >> fixed 120 ohm resistor and a variable 2k pot. The LM338 resistance >> formula is Vout = 1.25 * (1+ R2/R1). With R2 as 2k and R1 as 120 ohms, >> this yields an adjustment range from 1.2 to 22 volts. I'm building a >> power supply and want to increase my range to around 30V or so. >> >> To get the full range I would need a 2.7k pot. Unfortunately, all I >> have on hand are 2k and 5k. Using a 5k pot means the voltage will be >> maxxed out somewhere just over half the range of the pot. >> >> So, what I'm wondering is what happens if we substitute a larger >> resistor for R1. a 5K pot with a 220 ohm resistor would be just about >> right. I've read the datasheet, but I haven't quite figured out what >> effect this would have. I've seen some web articles that mention using >> resistors as high as 240 ohms, but they aren't specific about what >> side effect this may have. >> >> Thanks, >> Scott > >This approach has the disadvantage that as you rotate the pot you they >will often have momentary disconnects that will result in the max >voltage (>30V in your case) at the output. You could put a cap from adj to ground to mask that? I've not noticed the open wiper effect, but I'm usually running fixed divider or a restricted, say 5 or 10%, adjustment range. > >It is much better to have an arrangement where the wiper is not >carrying significant current and where any disconnects will cause the >output to fall to a low voltage. > >I would do it with the pot being used as a divider from a reference >and feeding the non-inv input of an opamp whose output feeds the LM338 >adj pin. > >Feedback from the output to the inv input of the opamp through a fixed >divider will give the appropriate voltage range (together with a >compensation capacitor). The LM338 will then act as a protected >output stage (overcurrent and thermal protection). A high value >resistor and a medium value cap from the non-inv opamp input will >ensure that any disconnects as the pot rotates will not cause any >disturbances at the output and if they do occur they will be in a safe >direction. > >kevin Add a small negative supply and you get adjustment down to 0V too. Grant. |