Dual Power Supply From Single (+-12V/+-1A From Single 15V)

by kuzysk in Circuits > Electronics

581 Views, 3 Favorites, 0 Comments

Dual Power Supply From Single (+-12V/+-1A From Single 15V)

schema1.bmp
1a.jpg

 Today I will write about how to make a dual supply -+12V (or other) from a single supply source. It's nothing special, but when I tried to search the internet, I found almost no practical construction.

The advantage of this power supply is that you can load only one loop and the device will work correctly. You can use a 24 volt without using a GND point and the device will also work correctly.

 History of design - For another project I'm working on, I needed a dimensionally small power supply of voltage -+12V and a current of at least 1A. I already had small experience with the LM2596 circuit, so I used its 12V version here.

 I don't think anyone will need to make PCB as I designed it, because my PCB is designed as a part of another device. This Instructables is about to show way how it can be made and to show that it works. In your project you can use LM2596's 3V3, 5V or ADJ version. (For the ADJ version, it is necessary to use two resistors and one capacitor more in each loop.)

Let's start...

Supplies

2x LM2596S12 or LM2596S5 or LM2596S3.3 or LM2596SADJ

6x Capacitors 470uF 35V

5x Diode SS54 or similar

2x Coil 33uH

2x Coil 3.3uH

4x Resistor 47k

2 x Capacitor 100nF

1x 2.1mm jack connector

1x PCB

Schematics Diagram

schema.bmp

Schematics diagram is on picture. For possibility to edit it in EAGLE just download two attached files. It is currently not possible to upload .sch file here so rename "DualSupply 12V_v1.0.sch.brd" to "DualSupply 12V_v1.0.sch".

If you compare schematics diagram to LM2596 datasheet you can see I just kept recommendation.

Although both loops (positive and negative) of the circuit look similar, they work on a different principle.

  The positive loop is a common example of a step down (or buck) converter that requires the voltage at its input to be slightly higher than the voltage at the output. That's why the +12V loop needs to be feed from 15V.

  The negative loop works differently. The principle of work is difficult to explain in a few sentences, so I will describe it when I have more time (or you can see the principle here.) Here, the principle of work is somewhat reminiscent of a step upp (or Boost) convertor, and we can produce -12V even if the input voltage is only 7V. It is important to write here that the negative this loop is significantly softer than the positive loop and maximal reachable current is about 1.5A. To achieve a current of 1A in the negative loop, needs an input voltage of at least 10V.

The PCB

2.jpg

When designing your PCB, you must not forget the important thing written in the LM2596 datasheet "HEAVY LINES MUST BE KEPT SHORT". Of course, almost all lines are "heavy marked".

As 150kHz is not super high frequency for switched power supply, probably any type of PCB design will work but difference will be at output ripple voltage.

Result by Multimeter

3.jpg
4.jpg
5.jpg
6.jpg

Here, I publish the results of the measurement of the loaded and non-loaded dual power supply I made. As you can see it works well. As load I used 10ohm resistor (25W) in both loops.(I could measure only couple of second because resistors were hot quickly) Soon I will measure the ripple voltage.

Note:

The green instrument PrisKit MT-1820 is not much precise even if it was 4 times more expensive comparing the Chinese Aneng Q1. I compared both with the most precise instrument I have (see comparing picture). So both positive and negative loop actually have almost the same voltage close to 12V unloaded and 11.55V at load 1.15A. Nice result I think.


Result by Oscilloscope

7.png
8.png
9.png

 I made a measurement on an oscilloscope. First, a more detailed measurement of the worse (negative) loop. When the source is not loaded, we see a ripple voltage of around 2mV with peaks of 7mV at a working frequency of around 160kHz.

 The next picture shows a comparison of both loops. Here we see that the positive loop has a smaller ripple voltage, which is due to the way the regulator works.

 In the last picture, both loops are loaded by a 10ohm resistor and the current is around 1.15A. Here the negative channel ripple is already quite large at value around 10mV with peaks up to 30mV. The positive loop has a half ripple.

This result is similar to cheap power source you can by on Chines market and is acceptable for most devices.