# Power supply LM317

#### SpNw

Joined May 27, 2020
6
Hello guys, I'm doing a mini project, and the objective is to make a small linear source (24V / 1.5A maximum) using a regulator (LM317) but I was dealing with a problem, my power dissipation at wrost (in LM317) if I am correct, its something like 46.48 watts (P = 32.5V - 1.25V) x 1.5A = 46.48 watts, so its extremly hard to LM317 hold that power dissipation, a huge huge heatsink will be needed, I tried to reduce it by dividing the current through 2 darlington transistores, what would reduce the dissipation, I want to know what do you think about this (I'm new in this game dont blame me ) ty everyone.

Questions:
How can I make a current limiter until 1.5A?
And what is the power dissipation in R1 and R4?

#### Audioguru again

Joined Oct 21, 2019
2,875
Your idea ruins the voltage regulation because the base-emitter of the output transistors have a higher voltage loss when the current is high.
This is on the datasheet:

#### Attachments

• 28.7 KB Views: 18

#### Papabravo

Joined Feb 24, 2006
15,520
Twinkle twinkle,
Little star
Power's equal
I squared R.

A cute way of remembering to square the current and multiply that by the value of the resistor. The units are watts.

#### SpNw

Joined May 27, 2020
6
Twinkle twinkle,
Little star
Power's equal
I squared R.

A cute way of remembering to square the current and multiply that by the value of the resistor. The units are watts.
P = 0.053w right?

#### Papabravo

Joined Feb 24, 2006
15,520
P = 0.053w right?
What numbers did you use? I think the power in R1 and R4 should be larger than 53 mw.

#### MisterBill2

Joined Jan 23, 2018
8,244
Hello guys, I'm doing a mini project, and the objective is to make a small linear source (24V / 1.5A maximum) using a regulator (LM317) but I was dealing with a problem, my power dissipation at wrost (in LM317) if I am correct, its something like 46.48 watts (P = 32.5V - 1.25V) x 1.5A = 46.48 watts, so its extremly hard to LM317 hold that power dissipation, a huge huge heatsink will be needed, I tried to reduce it by dividing the current through 2 darlington transistores, what would reduce the dissipation, I want to know what do you think about this (I'm new in this game dont blame me ) ty everyone.

Questions:
How can I make a current limiter until 1.5A?
And what is the power dissipation in R1 and R4?

View attachment 208311
You have just discovered for yourself the problem of power dissipated in a series pass regulator. It is a real challenge and large heat sinks are one option. The transistors wrapped around the regulator can help, but not the way you show it. The way that works uses a PNP transistor with the emitter fed by the input voltage, the collector tied to the LM317 output, and the base connected to the LM317 input terminal. In the most useful version there are series resistors but what I just described is the basic circuit.
Some power supplies reduced the input voltage when lower outputs were selected. That also works.

#### dl324

Joined Mar 30, 2015
12,684
The Nat Semi datasheet shows how to use LM317 as a switching regulator:

#### Attachments

• 737 KB Views: 7

#### MisterBill2

Joined Jan 23, 2018
8,244
The Nat Semi datasheet shows how to use LM317 as a switching regulator:
View attachment 208320
Leave out R5, C3, and D1 and you have the classic "wrap around" circuit. AND you can add several PNP power devices in parallel with the transistor, I think that the 2N3772 is a higher current choice, and you can have a much more powerful regulator. BUT it will still waste a lot of power with low voltages out at high currents. The fix there is to be able to reduce the supply voltage by reducing the AC input by changing the transformer connections.

#### MisterBill2

Joined Jan 23, 2018
8,244
This worked for me, replace R2 with 5k pot, and use the PNP to drive as many NPN outputs as you want,

http://www.bowdenshobbycircuits.info/page12.htm#317pass.gif
That circuit works, but it is now a linear regulator instead of a switching mode regulator. That means a lot more power dissipation than in the switching mode design shown in post #7. So you are back to less efficient operation. So tghere is the trade-off.

#### Ramussons

Joined May 3, 2013
920
Twinkle twinkle,
Little star
Power's equal
I squared R.

A cute way of remembering to square the current and multiply that by the value of the resistor. The units are watts.
And some one recollected it as:

"Little Star Up In The Sky,
Power Equals R Square I"

#### MisterBill2

Joined Jan 23, 2018
8,244
Yes, that is possible IF the TS is able to produce a suitable circuit board for the small surface mount part and IF the TS is able to create a suitable layout for the switch-mode supply and IF the TS does not need more current than the specified 1.5 amps. AND IF that part, at 4.39 Euros remains available, then it could be done that way. And by the way, that is a single-sourced part, only made by one company.
The LM317 is a mature part produced by multiple sources with an application track record of more than 20 years. Even better, it is easily adapted to much higher current applications by the simple means of adding shunt transistors. It is easily heat-sinked and produces very little noise if the application recommendations are followed. It requires no circuit board to be usable, and even the expanded mode application is simple to construct. PLUS, it costs less.

#### andrewmm

Joined Feb 25, 2011
1,142
Yes, that is possible IF the TS is able to produce a suitable circuit board for the small surface mount part and IF the TS is able to create a suitable layout for the switch-mode supply and IF the TS does not need more current than the specified 1.5 amps. AND IF that part, at 4.39 Euros remains available, then it could be done that way. And by the way, that is a single-sourced part, only made by one company.
The LM317 is a mature part produced by multiple sources with an application track record of more than 20 years. Even better, it is easily adapted to much higher current applications by the simple means of adding shunt transistors. It is easily heat-sinked and produces very little noise if the application recommendations are followed. It requires no circuit board to be usable, and even the expanded mode application is simple to construct. PLUS, it costs less.
All true , and the designers choice, just giving choices,

#### MisterBill2

Joined Jan 23, 2018
8,244
Referencing post #1, and how to reduce power dissipation in the linear regulator, The simple way is to reduce the AC voltage by changing transformer connections. This is how it was done in a number of llab-grade power supplies at one time. Automated tap selection changing as the voltage control knob was turned. And at least one model had a range selection switch. Another brand of variable supplies was the SORENSON brand, which used an SCR phase control scheme for both control and regulation. I am not sure if a modern mains powered triac light dimmer switch would function as a primary voltage adjuster, but it would certainly be an interesting experiment worth trying. And quite simple to do, as well .

#### Hamlet

Joined Jun 10, 2015
435
I am not sure if a modern mains powered triac light dimmer switch would function as a primary voltage adjuster, but it would certainly be an interesting experiment worth trying. And quite simple to do, as well .
I tryed a dimmer specifically made for dimming low-voltage lighting transformers; although better than your standard
dimmer, it was still very twitchy without a load, especially at lower settings. As I recall, a small 50w load helped, but
I was unwilling to used a ganged log/linear-pot arrangement to gain simultaneous control over both pre-regulator and final lab
output. I junked the idea and used a transformer with 120,240 inputs, and a DPDT switch.

#### MisterBill2

Joined Jan 23, 2018
8,244
I tryed a dimmer specifically made for dimming low-voltage lighting transformers; although better than your standard
dimmer, it was still very twitchy without a load, especially at lower settings. As I recall, a small 50w load helped, but
I was unwilling to used a ganged log/linear-pot arrangement to gain simultaneous control over both pre-regulator and final lab
output. I junked the idea and used a transformer with 120,240 inputs, and a DPDT switch.
A switch was going to be my next suggestion.

#### SpNw

Joined May 27, 2020
6

#### Audioguru again

Joined Oct 21, 2019
2,875
I removed as much of the dark grey background as I can for better contrast.

#### Attachments

• 155.9 KB Views: 13

#### MisterBill2

Joined Jan 23, 2018
8,244
The circuit in post #18 looks reasonable, but how does the simulator get two different vol6tages on the output line? 20.4035 at the regulator, and 23.7434 at the output. I find that a bit confusing.

Also confusing is the use of a six digit voltmeter.