LM317 Variable power supply

[poff1]

Hi,

Great looking site!

I have a few q's for this project:

-Which of the following should I get?

+ 1.2 V to + 37 V NSC 8 / SOIC (Small outline IC) LM317LM
+ 1.2 V to + 37 V NSC 3 / TO-92 LM317LZ
+ 2 V to + 37 V ON 8 / LDO (IND TEMP) LM317LD

-What kind of transformer am I looking for (I've never done this sort of thing before)

Encapsulated?

-What kind of bridge should I get?

A parts list would be a good addition to the project entry!

Also, I'll be ordering over the internet from somewhere like www.farnell.com, so links to suitable parts would be great.

How about casing?

Thanks!

Project Link: http://www.electronics-lab.com/projects/power/010/index.html

P

 

[mixos1]

Regulator:

The best is to use LM317T which is in a plastic package TO-220 and has a load current 1.5A. Like this: http://www.allelectronics.com/cgi-bin/category.cgi?category=search&item=LM317T&type=store

Transformer:

You need something like this: http://www.allelectronics.com/cgi-bin/category.cgi?category=790500&item=TX-2815&type=store


Bridge:

You need any rectifier bridge that can handle 2A. Nothing special. Example: http://www.allelectronics.com/cgi-bin/category.cgi?category=search&item=FWB-206&type=store

Casing:

Imagination is the only way to find a good case

 

[poff1]

Thanks a lot for your reply!

For the rectifier as it's 28V DC output, I can use a 50V 2A one right?

Also, is it easy to use another transformer value?

I'm having trouble finding 220/28V, but I see lots of 220/24...

Would it mean changing a lot of values for caps etc?

The transformer should handle 2A too right? So 28*2=56VA min?

Sorry if I'm mixed up!

P.S:

BTW, does the regulator generate a lot of heat?

The TO-3 package would be good on a heatsink, but only if it actually generates heat...

P

 

[mixos1]

A 50V 2A rectifier would be nice.

If you use a 24V tranformer then the max output voltage that you can get is 21V (because the Vin must be around 3V higher that Vout).

For that value (21V) and the use of this equation: Vout = 1.25V * (1 + R2/R1) + Iadj * R2 (that describes the LM317) => R2=[(Vout/1.25) - 1] x R1 (because Iadj=100uA typical) and if R1=240 => R2= 3,8k

But it's dificult to find a 3,8k pontesiometer. For that reason use the nearest value.

Nothing else must be changed.

A transformer must be capable to deliver 2A at minimum. Yes.

It is reccomended that you use a heatsilk. If you consume more than 1W then you must use it.

 

[poff1]

Cool,

Can I use one of those transformers in series?

Like:

2x115V/2x15V

How would I connect it to the mains (230V) supply?

 

[Chevees1]

For safetly running I recomend add 2 universal Si-diode, e.g. 1N4007.
1. add diode to output - anode to GNG and cathode to +V
2. add diode like bridge over LM317 - anode to Vout and cathode to Vin


If You want using this power supply with minimum outgoing current (e.g. LCD-panel) You must add some loading, e.g. serial LED with resistor to output (You will got signaling power on ). Otherwise this power supply will be very unstable (LM317 need min. 5mA load).

Project Link: http://www.electronics-lab.com/projects/power/010/index.html

 

[Pinky1]

Hi.
I guess you are referring to " http://www.electronics-lab.com/projects/power/010/index.html" project

As Mike suggested you need LM317T that can handle 1.5A.
I have some more suggestions to make.

Firstly, about the transformer. LM317 can handle 3V to 37V but you should go to 30-32V.
Also, LM317 needs input to be +3V from output approximately.
A transformer 230V/28V is not needed. 230V/24V will not give you 21V as Mike said because the 24V is the AC voltage that the secondary coil of the transformer has.
After the bridge you will have 24*1.4 (sqrt(2)) DC voltage so you will have 33.6V which is enough to give you 30V.

Secondly, everyone uses R1=240 without looking at LM317's specs.
The LM317 datasheet suggests 240 but if you want to avoid the worst case scenario (load=0.Vout=1.25) since LM317 needs at least 10mA current flowing, with 1.25 Reference Voltage you must have R1<125, so many circuit makers (e.g velleman) use R1=120

Also for better output you can put a 10uF tantalum capacitor accross the 4.7k potensiometer.
For safety you can also put 2 protection diodes (e.g 1N4001,1N4007). One accross R1 facing towards LM317's output and one from LM317's output towards LM317's input so that when you cut off the switch the capacitors do not discharge through LM317.

The last i want to say is that LM317 dissipates extra voltage as heat and also it gives 1.5A only
when (Vin-Vout)<15V, so if you have 24*1.4=33V input and you want 5V as output you will not have 1.5A and LM317 will be very hot.
An easy solution to this is to get a transformer 230V/2*12V and use a switch in order to use only 12V or 24V as input, so if you want e.g 9V you can have 12*1.4 as input which is far more good than 24*1.4

Some of what i suggested are extreme but i believe make the circuit better.

Thank you.

 

[poff1]

Thanks for posting.

I've been trawling google and am compiling my "own" schematic.

You've helped a lot with the transformer ratings!

P

ps. I'll post later when I've finished

 

[eagle20031]

This one is quite interesting project.
how u r controlling the o/p voltage.
if u r using conventional pot method then i have one suggestion.
u can use digially controlled one. attaching
digital up down counter section for adj. pot selection.
i hope u will find it userful.
use circuit in datasheet of lm317t(digitally selected o/ps)

Click here for higher resolution of schematic

 

[MP1]

Actually, my data sheet lists 3.5 mA as minimum typical load current for the LM317.
In referencing a typical 2X16 display data sheet: It is true that the logic current on most LCD displays only draw 1 mA typical, but the lighting draws considerably more. This can be in the 100 mA to 200 mA range. Even higher on some of the STN LCD displays. You must have a contrast voltage which draw current and backlighted displays are even more current hungry. If this is a serial display, you are also powering a microprocessor chip. Ususally a PIC with a program in it to add the serial option. I certainly wish I could find a display with such low spec as 3.5mA or less. This would be my all time favorite for all of my future microprocessor projects. LCD displays are the one item that drains my battery powered circuits the worst. Unfortunately, everything I design with a microprocessor seems to have a display. I love them ;D.
You will not have a problem with a LM317 for powering an LCD display. Also note, one never powers only a display. You will need other circuitry to send data to the LCD display. Typically, you will add the current draw of the microprocessor and any other peripherals in the circuit to the total draw. Microprocessor, Display, Contrast circuit, Analog to Digital Converters, switches, ram, or eprom. All will add to the current consumption. Try using a 9V battery to power a circuit with a display. It will not last for very long. You will find that you need to go to a battery type with a much better storage capability. This is due to the large current consumption of the display.
The LM317 and LM337 are great for bench power supplies. Use the LM317 for the variable positive and the LM337 for the variable negative and you have a very nice variable bipolar bench supply. If you want to get less than the minimum 1.2 VDC on a variable adjustment, just add a couple of diodes on the output. This will bring it down to zero for the low end. Also note that this will bring down the high end by the same amount, so compensate accordingly on your calculations for the resistors.

MP

 

[MP1]

Hmmmm.....not really sure where this belongs. I have uploaded a voltage calculator for LM317. See the download area as soon as it is available.
It is a nice calculator for figuring R1 and R2 for a particular output voltage when using the LM317. Makes less work. ;D

MP

 

[mixos1]

The LM317 calculator is here:

http://www.electronics-lab.com/articles/LM317/index.html

be sure to check it .

 

[AnalogX]

hi mates im planning to build that one, after having on hand first the lm317t is there no effect if im building it with lower current says 750ma? sorry for my ignorance :-\

 

[MP1]

You cannot get more current out of this one than the spec. You can, however, put two of these is parallel and get more current out. Also depends upon what you have available on the input side.

As far as the build instructions, there is no difference.

MP

 

[bogdan2]

just a suggestion. if you cannot find a 2 x 12v transfoermer, you could use a power suply made from 2 LM317 regs in series. though, the maximum output voltage is a bit lower.
i think that having that switch is the best way.
also, if you could use a 3A or 5A bridge, then you will have less power lost on the bridge.
Pinky, 24*1.4=33.6, but you forgot about the loss on the birdge.
thats why i suggest to use a 5A bridge. you loose less on it, since it has a lower dropout voltage.
you could use a 4700uF cap for the filtering of the rectified ac. this means lower ripple voltage.
and don't forget about the minimum load resistance. you could use a 120R resistor as suggested, but at 30V you have 7.5W disipated on it, not very good. i think that a single LED with a resistor in series is enough. and add a 5K resistor at the output too.

 

[mikratt]

Hi I am new hear on this fourm and I am trying to build one of these power supplies. Instead of using a Transformer and a Bridge I wanted to use an AC adapter that has an output of 24V DC 1A. The problem is I get the varying voltage from 20V DC to about 8V DC and very little amperage about 100 milliamps max. I need to get between .3 amps to 1 amp. I don't know what is wrong any suggestions.
Thanks
Mikratt

 

[Firefox2]

Hey man, that is one of the problems I am having. I am using AC to DC Wall plugs. My power is fine going in, but it seems to lose alot of voltage and the amps suck once it goes thru. I am getting frustrated and thinking about trying the AC input, but dont see where it should matter. I figure probally something simple being missed.

Hope someone can shed light on this.

Ken

Hi I am new hear on this fourm and I am trying to build one of these power supplies. Instead of using a Transformer and a Bridge I wanted to use an AC adapter that has an output of 24V DC 1A. The problem is I get the varying voltage from 20V DC to about 8V DC and very little amperage about 100 milliamps max. I need to get between .3 amps to 1 amp. I don't know what is wrong any suggestions.
Thanks
Mikratt

 

[MP1]

You are losing a bit when you pass the DC voltage through the rectifier bridge. The rectifier bridge is for the AC input. Just tie into the circuit after the 4 diodes that make up the rectifier bridge or somewhere on the trace that goes to the input pin of the LM317. As long as you have enough current to begin with, your problems should be solved.

MP

 

[mikratt]

Hi MP, I am not using a rectifier bridge. I have eleminated the Transformer and the rectifier bridge by using an Wall Wart AC Adapter with an Output of 24DC 1Amp.

Sorry you