LM317 5A Power Supply

Papabravo

Joined Feb 24, 2006
21,159
Hi, I've got a question. How can I further improve this circuit: http://www.circuitdiagram.org/lm317-5a-variable-power-supply.html

1.) Will this actually output 5A?
2.) Should I add a 10uF cap in parallel with the 5k ohm pot and diodes on LM317's (In, Out) and (Adj, Out) to protect it from the capacitors.
3.) Is it alright if I replace the 3,300uF cap with 10,000uF, 0.33uF with 100nF, and 100uF with 2,200uF?
You need to think very carefully about power dissipation in this design. Even with a properly designed heatsink you might have to dissipate 160 watts or so. Can you and the part handle that much?
The θjc for a TO-220 is 3°C/watt so 160 watts will result in a temperature rise of 480°C. The θja is like 34°C/watt so that is 5440°C temperature rise over ambient. That is nearly an order of magnitude hotter than molten steel.

IMHO this is a really crappy way to make a 5A power supply.
 

Thread Starter

Charles Mendoza

Joined Jun 12, 2017
5
You need to think very carefully about power dissipation in this design. Even with a properly designed heatsink you might have to dissipate 160 watts or so. Can you and the part handle that much?
The θjc for a TO-220 is 3°C/watt so 160 watts will result in a temperature rise of 480°C. The θja is like 34°C/watt so that is 5440°C temperature rise over ambient. That is nearly an order of magnitude hotter than molten steel.

IMHO this is a really crappy way to make a 5A power supply.
Thanks for your response sir. I know it's a really crappy circuit and I knew immediately that heat will be my worst enemy here, but I'm limited with the components I currently have. I'd be glad if you could suggest a good circuit design.
 

AnalogKid

Joined Aug 1, 2013
10,989
1.) Will this actually output 5A?
2.) Should I add a 10uF cap in parallel with the 5k ohm pot and diodes on LM317's (In, Out) and (Adj, Out) to protect it from the capacitors.
3.) Is it alright if I replace the 3,300uF cap with 10,000uF, 0.33uF with 100nF, and 100uF with 2,200uF?
1. Yes, if the power transformer can source 8 A of current.

2. Yes, yes.

3. Yes, yes, why?

Note: this is not a good circuit. For any particular output voltage setting, the actual output voltage will vary by over 1 V as the output current changes from 0 A to 5 A. The LM317 datasheet has a better circuit, but it uses a PNP power transistor.

ak
 

Papabravo

Joined Feb 24, 2006
21,159
Thanks for your response sir. I know it's a really crappy circuit and I knew immediately that heat will be my worst enemy here, but I'm limited with the components I currently have. I'd be glad if you could suggest a good circuit design.
That is a really lousy excuse for doing something so bovine. At least check out the datasheet for the design of the PNP alternative. A better design would involve selecting a transformer with multiple output windings. These would be used to drop the input voltage to the regulator so you can limit the power losses in the worst case of a low output voltage and a high current.
 

dl324

Joined Mar 30, 2015
16,846
Thank you everyone for all your response. I've checked out the datasheet and I've decided to go with the PNP design.
This circuit can be used to substitute a less expensive power NPN.

From National Semiconductor Voltage Regulator Handbook:
upload_2017-6-12_7-54-17.png

The circuit you referenced is a horrible design. It and another variant pop up all too frequently...

Instead of throwing stuff at a design in the hopes that it will improve it, you should read the LM317 datasheet. The Nat Semi version has a lot of good information.
 

Bordodynov

Joined May 20, 2015
3,177
I do not understand the obsessive desire to use an old chip with additional transistors. There are much better microcircuits of the stabilizer for high currents.
Arguments against LM317: 1. Increases the minimum voltage drop and, accordingly, the power dissipated. 2. It is necessary to enter additional elements to limit the current.
Arguments for the use of new chips:
1. . Simple electronic circuitLT108x.png . 2. The best characteristics of the stabilizer.
 

ian field

Joined Oct 27, 2012
6,536
Hi, I've got a question. How can I further improve this circuit: http://www.circuitdiagram.org/lm317-5a-variable-power-supply.html

1.) Will this actually output 5A?
2.) Should I add a 10uF cap in parallel with the 5k ohm pot and diodes on LM317's (In, Out) and (Adj, Out) to protect it from the capacitors.
3.) Is it alright if I replace the 3,300uF cap with 10,000uF, 0.33uF with 100nF, and 100uF with 2,200uF?
Most application notes show the current sampling resistor in series with the input pin, and a PNP external pass transistor on a positive regulator.
 

Thread Starter

Charles Mendoza

Joined Jun 12, 2017
5
Just wondering, is it possible to adjust the current like so:


(Not sure if that's where I should place D1 to protect the ICs from C6 when discharging)

Correction: R7 is connected to Vout of U2 and the other end on VR4's middle pin.
 
Last edited:

MrAl

Joined Jun 17, 2014
11,396
Hi, I've got a question. How can I further improve this circuit: http://www.circuitdiagram.org/lm317-5a-variable-power-supply.html

1.) Will this actually output 5A?
2.) Should I add a 10uF cap in parallel with the 5k ohm pot and diodes on LM317's (In, Out) and (Adj, Out) to protect it from the capacitors.
3.) Is it alright if I replace the 3,300uF cap with 10,000uF, 0.33uF with 100nF, and 100uF with 2,200uF?

Hi,

The pass transistor version has one advantage over the single IC chip solutions. That is that the chip itself like the LM317 does not heat up as much if you implement some alternate means of current limiting. That's not enough though for me to use a pass transistor instead of a single IC chip that can handle the current.

First, you should go with an IC chip that is rated for 5 amp or even 8 amp. That way you dont need a pass transistor, and i assume you want a simple power supply.

I built one of these a few years ago and used an LM338T which are cheaper than the LT versions and handles 5 amps.

Yes it's not the best way of doing it, but after all it is one way to do it.

Second, instead of the 3300uf cap, you should go with 3300uf per amp of output, or min 2200uf per amp output current.

Third, you need a pretty big heat sink if you intend to run at high current and large input/output differential voltages.

The LM317 and similar all drift more than expected when the chips heat up due to higher current outputs. Always measure your output after some time after the load has been applied.
 

bertz

Joined Nov 11, 2013
327
Hi, I've got a question. How can I further improve this circuit: http://www.circuitdiagram.org/lm317-5a-variable-power-supply.html

1.) Will this actually output 5A?
2.) Should I add a 10uF cap in parallel with the 5k ohm pot and diodes on LM317's (In, Out) and (Adj, Out) to protect it from the capacitors.
3.) Is it alright if I replace the 3,300uF cap with 10,000uF, 0.33uF with 100nF, and 100uF with 2,200uF?
You would do well to follow this series of videos on power supply design.
 
Top