Power Supplies for Beginners, Part 3

Power Supplies for Beginners, Part 3

Previously in part 2 of our series of video tutorials for power supplies for beginners, we explained how to test and use Unregulated Power Supplies and showed how an unregulated power supply has difficulty in controlling its output. Here in part 3 of our video tutorial series we will look at Linear Power Supplies and shows how series and shunt regulators are much better at controlling their output.

Watch our Part 3 video tutorial to learn more!

Following Transcript of the Video Tutorial

Time: 0:00sHello I’m Chris Richardson, and I’m an electronics engineer focused on power supplies. This is the third part of a series of web seminars for power supply enthusiasts who aren’t necessarily trained as power electronics engineers.

So far in parts one and two we gathered the basic equipment for testing power supplies without spending a fortune and then we found and tested some older unregulated power supplies. Now its time to evaluate and test the oldest and most basic type of regulated power supply known as a Linear Regulator.

Zener and Shunt Regulators

Time: 0:27sThe diagram on the left is a discrete linear regulator made up of a zener diode and a resistor, R_S. All the resistor does is limit the current. If it wasn’t there, the input supply would either melt the zener with too much current, or the input supply itself would run into its own current limit.

Such a circuit is very, very cheap, but the tolerance to the output voltage depends upon the zener voltage, V_Z and that depends on the load current, the temperature, and the natural distribution of V_Z itself from part to part. To me it’s debatable whether or not this type of supply is regulated, but its a good introduction to the true regulated circuits coming later. Since the active element is in parallel with the load, we say that it “shunts the load”, hence the name Shunt Regulator.

TL431 Shunt Regulator Circuit

Time: 1:08sOn the right is a similar circuit that uses a genuine integrated circuit for better precision. The TL431 and its variants are everywhere in the power supply world, but are not often used as shunt regulators like we see here. The TL431 is so similar in its function to a true zener diode, that the symbol is often drawn like a zener as I have shown.

Resistor R_S still limits current, and still must meet the minimum and maximum limit, but now resistors R_TOP and R_BOTTOM divide down a portion of V_OUT and feed it back the the reference pin. Inside the TL431 are active transistors, and the Ref pin allows V_OUT to vary from the reference voltage up to V_IN minus about 1 volt. That 1 volt is the so called drop-out voltage and we will discuss that in detail in the next slides in the video segments.

Time: 1:53sHere I’m showing the very basic TL431 based shunt regulator solution. Here on this piece of proto-board (prototyping board) I have the actual TL431 and R_S current limiting resistor and on the backside is a blue ten-turn precision 50kΩ potentiometer.

So this is both, R_TOP and R_BOTTOM, or R_A and R_B, so if I was to turn this dial I would adjust the output voltage. I have adjusted it to 5 volts which would be typical for something like an Arduino.

Time: 2:33sThis is the unregulated power supply to a phone that my cat decided to kill, so if I switch it ON, then we can see here we have just under 10 volts input and 5 volts here at the output.

Multimeter Voltage Readings

The next test I am going to do is apply a load and show that this regulator still maintains the output voltage under load.

Linear Shunt Regulator Under Load

Time: 2:55sHere is the same circuit but now its loaded by 75Ω’s that two 150Ω power resistors placed in parallel. You can see that the input voltage has dropped a little bit but the output voltage is still maintained.

Once again, the shunt regulator based on the TL431 is connected to its 66mA or 75Ω load, and what I want to show here on the screen is the ripple in yellow, that’s the input voltage, and how nice and smooth the same volts per division the output voltage is. So that’s really what the shunt regulator is doing for us.

TL431 Shunt Regulator Voltage

Another example would be if you are using an Arduino power supply, then this could clean up a voltage that’s both too high and has too much ripple and make it nice and smooth and give the 5 volts that the Arduino would want.

Time: 3:43sHere’s the last test for the shunt regulator and is a good explanation of why shunt regulators aren’t used except for very low power situations. So I have changed things around, now this multimeter is actually measuring the input current. So the unregulated power supply comes in here, gets measured as we are now using this as an ammeter, and goes back into the circuit.

Right now I have the load connected and we can see its drawing about sixty milliamps, (60mA). What I am going to do now is disconnect the load and you can see that there’s a transient but then the current goes straight back to normal. The upper voltage stays the same, but because of this resistor here, the shunt regulator is always drawing load current. So if your circuit is operating at no-load you are still using power and some might say wasting power.

Series Linear Regulators

Time: 4:33sThese more sophisticated linear power supplies are known as “series regulators”. As the name implies, a transistor operating in its linear active region goes in series with the load. The current limiting resistor R_S is not needed here and that saves some power.

Time: 4:48sThe circuit on the left is similar to the zener shunt regulator in that it produces one and only one output voltage value. However, inside the 7809 there are series regulators with fixed output voltages are a pair of feedback voltage divider resistors like R₁ and R₂ in the right hand circuit.

In both types, the feedback circuit adjusts the voltage across the active terminals of the transistor. This transistor is often called the pass element since it passes the current from the input to the output. The active voltage is continually adjusted to maintain the desired output voltage, and such a circuit is also called a voltage or a potential divider.

Another way to think about this is to imagine a resistor divider with the top resistor, R_TOP is actively adjusted and the bottom resistor, R_BOTTOM is the load.

The LM317 Series Linear Voltage Regulator

Time: 5:30sHere I am showing the LM317 series linear voltage regulator. It is the same pcb with the device itself. This is a minimum load resistor this device needs between two and three milliamps to regulate properly, but that’s a lot less than the 60mA we were drawing with the shunt regulator, and again here is the 10-turn potentiometer to adjust the output voltage, and I have adjusted it to give 5 volts at the output.

It’s finally time to make some actual use of that ATX power supply we turned into a bench power supply. So I am going to use the 12 volt input here, and there is a very noisy fan so you can tell that I am actually using it, and here is 12 volts at the input and 5 volts at the output.

Time: 6:23sOnce again, the LM317 series linear regulator is going to be connected the the four ohms of power resistors here, and if we look at the oscilloscope screen, again in yellow, that’s the input voltage, this is now the output of a switching regulator.

Notice that the time division is much tighter because this is not 100Hz ripple its probably 100kHz ripple and some of the noise gets to the output, but the blue, which is the output, is much smoother.

LM317 Series Linear Regulator Voltage

Time: 6:51sA series regulator is generally capable of a lot more current than a shunt regulator and a LM317 is capable of more than 1 amp. I have switched things around here and now the orange multimeter is measuring my output voltage while the blue multimeter is measuring the output current.

Here I have these two 8Ω power resistors in parallel to give a load of about 4Ω. So when I connect them in, the circuit starts to draw over one amp, and now the control here is ok, but keep in mind that there is no output capacitors here and we are not doing kelvin sensing (4-terminal sensing) of the load.

Time: 7:28sI will have to do this experiment quickly as my linear regulator is quickly overheating, so here it is with the load and we can see more clearly how much ripple the switching power supply has and that’s due to the circuit overheating.

Low Drop Out Series Linear Regulator

Time: 7:40sThe 780x series and LM317 regulators are often called “NPN Regulators” because their pass elements are npn bipolar junction transistors. They are great parts and some of their designs are over 40 years old and still going strong.

But their biggest drawback is their large drop out voltage. That’s the minimum difference needed between the input voltage and the output voltage to keep the circuit regulating properly, and its about 2.5 volts for NPN regulators. An LDO, which stands for “low drop out regulator”, uses PNP transistors or more commonly MOSFET’s to allow the maximum output voltage to get very close to the minimum input voltage. Some parts get close to 100mV’s or less.

That’s perfect for modern circuits where you might want to drop from 1.8 volts down to 1.5 volts. One example is the circuit shown on the left with the simplified block diagram of the inside of the LTC3025 shown on the right. In theory this circuit can regulate down to a drop out voltage which is just above the load current multiplied by the ON-resistance of MOSFET, M1 (V = I_LOAD*R_ON).

Time: 8:43sI am now showing a genuine low dropout voltage regulator built with the LT1575 and this one is not that common as it has the control chip here and actually has a discrete power transistor and then a MOSFET. You can see this giant heatsink clearer when I do the heat experiment and we will compare this to the LM317 that had no heatsink.

Right now there is no load and I have all four of these 8Ω power resistors in parallel to give me a 2Ω load. The input voltage is nominally the 3.3 volts coming from the ATX bench power supply, and I am also using the 12 volt to actually power the control section and that is the way that this chip is actually able to get to such a low drop out voltage as its the higher voltage that actually powers and drives the gate of that nMOSFET.

So what happens when I actually connect the load, (2Ω) we can see that the input voltage collapses. Now that is not because the ATX power supply isn’t capable of giving all the current, the problem is the voltage drop in all these long thin wires.

What I want you to notice is that the output voltage only drops to 2.1 volts. It is suppose to be 2.8 volts as it is with no load, but with a load we can see that the drop out about 100mV’s or so.

Time: 10:09sI have setup my low drop out regulator once again to show that the input current and the output current are related almost directly in a linear regulator. The yellow multimeter is the input current and the blue multimeter is the output current and this device is an LM317 being used as a constant current source.

You can see that there is almost no current being drawn at the output, the input is drawing about 150mA and that’s due to this minimum load resistor down here. However as I start to increase the load on the output, notice that the input current is equal to the output current plus the 150mA, or so. As I turn it all the way up to its maximum, you can see that they track.

Heat – How Much is Acceptable

Time: 10:50sIn part two, we saw that the bulky line transformers of the unregulated power supplies barely heated up at all even at their maximum loads. With linear regulators heat is a much more immediate concern. The power dissipation is easy to predict as its equal to V_IN minus V_OUT multiplied by the load current, (V_IN-V_out)*I_L.

Engineers take the worst case into account, the maximum input voltage, V_IN(max) and the maximum load current, I_L(max). As a general rule, semiconductor packages can take about 1 watt before they can get too hot. Anymore than 1 watt, then a heatsink or forced air-flow (fan) is needed.

But how hot is too hot but that depends on many factors like ambient temperature, the air flow, the presence of sensitive components nearby, like aluminium electrolytic capacitors, but also something more basic, how long the power supply needs to last.

Low Drop Out Regulator Heat Test

Time: 11:38sI am running a heat and power dissipation test on my discrete LDO regulator, so I am back to my 2Ω load and I am now using a full 5 volts from my ATX power supply at the input. I have the 2.8 volts out and the output load current is 1.32 amps. The thermocouple is put inside the clip of this big heatsink and reads only 35^oC.

I will remove it and stick it on the tab which is the hottest part of the MOSFET to see if I can see what the temperature there is. Now that’s heating up much more quickly. The tab is probably getting to somewhere around 40^oC or so, and to me that is fine as long as it doesn’t get to over 50^oC then I think it is perfectly safe.

Time: 12:39sThat concludes part three of power supplies for non-EE’s. Stay tuned for part four where we will look at switching power supplies and by far the most exciting topic.

On behalf of myself and Electronics-Tutorials.ws, thanks for watching and how to see you for part four.

End of video tutorial transcription.

You can find more information and a great tutorial about linear power supplies by following this link: Variable Linear Power Supply.

In part 4 of our video tutorial about power supplies for beginners, we will look at using Switching Power Supplies and see how Buck and Boost converters can increase (boost) or decrease (buck) the output voltage.