Part 5 of our video tutorial series on power supplies for beginners and non-electronics engineers compares the different types of power supply.
Here in part 5 of our series of video tutorials for beginners on power supplies, we will finish off by comparing the different types of power supply we have seen, both Linear and Switching, and see the advantages and drawbacks of both the switching and linear power supply.
Time: 0:00sHello I’m Chris Richardson, and if you made it to part 5 then you probably know by now that I am an electronics engineer focused on power supplies. This is the fifth part in a series of web seminars for power supply enthusiasts, or hobbyists who aren’t necessarily trained as electronics engineers.
So far we have gathered some low cost equipment to test power supplies. Looked at unregulated power supplies. Tested various linear regulators and tested various switching regulators. In this section we will compare examples of these different power supplies and examine them to see which fits best in different applications.
Time: 0:31sThere are many, many different types of power supplies out there. From tiny systems using milliwatts, (mW) or even microwatts, (uW) in areas, like the so called “energy harvesting” field, to megawatts, (MW) in electrical generation and distribution. Selecting the most appropriate device for your application is therefore a critical step in power supply use and design.
Time: 0:48sRegardless of the type of power supply being tested, accurate power efficiency testing requires one ammeter and one voltmeter for the input to your supply, and then another ammeter and another voltmeter for each output. For very low power circuits, generally under one-tenth of a watt (1/10W or 0.1W), special equipment is needed because the ammeter and voltmeter always consume some power and would distort those measurements.
I am going to focus on power supplies of at least one watt of output power, since that special equipment is definitely not on my list of affordable devices that I talked about in part one.
Time: 1:18sKelvin Sensing refers to measuring the input voltage and output voltage directly at the inputs and outputs of your power supply. The demo boards I have been using always include test points right next to the input capacitors and also the output capacitors for this purpose.
If you use the voltage readout of a lab power supply or trust the ATX box to give exactly 12 volts or 5 volts, your measurements will be wrong, since voltage is lost due to resistive drops in the connecting cables. The ammeter itself also uses a series resistor and some voltage is lost there too.
Time: 1:48sFor the first efficiency experiment, I am going back to my unregulated power supply here. It is being used with a linear current source set to draw 500mA. The blue multimeter is measuring the input current, (IIN) and the orange multimeter is measuring the input voltage, (VIN).
When switched “ON” the input is drawing 30.8mA at 226 volts rms and now I will switch things around and look at the output current, (IOUT) and output voltage, (VOUT).
Time: 2:16sThe same circuit, but now when I switch it “ON” I am going to measure the output current at 510mA and output voltage at 6.2mA. Remember that this is a semi-regulated circuit. In practice if we were going to calculate lots of efficiency points, we would vary the load. However, this linear current source is actually logarithmic in the way it adjusts with this potentiometer so it makes it a little bit more difficult.
Now I am repeating the experiment but instead of using the unregulated power supply, I am using a regulated switching power supply. This is the output current and output voltage and when I turn it “ON”, again 510mA. But it is not actually as well regulated as I expected it to be since its 6.5 volts and here we actually have 7.07 volts. Nonetheless, we can take these two data points and make an efficiency plot.
Now I am testing the input current and the input voltage to my switching dc-to-dc power supply, and we can already see that the input current is much lower whereas the input voltage is almost exactly the same. So we know that the efficiency is going to be much better and now we have two more data points so let’s go ahead and calculate.
Time: 3:26sIn the previous slides and videos, we saw that switching regulators are vastly more efficient than linear regulators in most cases. So it should not be too much of a surprise that linear regulators used in the same conditions of VIN, VOUT and IOUT, dissipate a lot more power and their components get a lot more hotter than equivalent switching regulators.
Still the lower electrical noise, simplicity and low cost of NPN regulators and LDO’s (Low Dropout Regulators) make them my preferred choice whenever it is reasonable to use them. My criteria are the following:
This assumes that you do not have the space or the budget for a heatsink, and in my experience there is really space or money for heatsinks. It might be surprising to you that many heatsinks that can dissipate more than one watt cost more than the power supply control chip.
Time: 4:29sTo talk about heat, I have here my synchronous buck converter. This is probably the most efficient of all the switching power converters and its delivering about 27 or 28 watts. I am powering it from 12 volts from my ATX power supply here. I have almost exactly 5 volts as the output voltage, about 5.5 amps as the output current and that’s thanks to a group of power resistors who’s total resistance is just under 1Ω.
To get an idea, the ambient temperature in the room is somewhere between 27 and 28oC. One of the power resistors is quite hot at close to 50oC. I always think of that as anything over 50oC is uncomfortably to the touch.
If I start to measure the temperature of some of the power components, the switching MOSFET is the component that gets the hottest, and it is barely over 30oC. This is the synchronous power MOSFET, slightly cooler at 29oC.
The power inductor is also an element that can heat up a lot and barely over 30oC. The last thing we will measure is an aluminium electrolytic input capacitor that is barely heating up at all. So that means it should have a nice long lifetime.
Time: 5:57sIf you watched the section on linear regulators, then you will remember that this discrete linear regulator (showing circuit board) has a big, big heatsink and has a control chip and a discrete power transistor. So we will compare this to the buck converter that we just did.
The blue multimeter is the output current, (IOUT) and the orange multimeter is the output voltage, (VOUT), so actually the load here which is altogether 1Ω is drawing so much current it is actually collapsing to output voltage a little bit here. But this is still a good thermal test with the ambient temperature here about 27oC. The only thing that really matters in a linear regulator is the chip itself, which in this case the discrete pass element.
If I put the temperature probe on it it is nice and hot, probably over 100oC. I have the tip of the thermocouple right at the junction of the heatsink and the tab of that discrete power MOSFET. There is a huge difference, but remember that none of the components of the buck regulator were getting to be over 31 or 32oC, or so. I will turn it “OFF” as it is overheating.
Time: 7:16sLinear regulators beat switching regulators without any doubt when it comes to low conductive noise. That goes for their inputs which we see at the top of the screen here, and those may be subject to legal limits, and for their outputs which are often sensitive to noise. For example, most digital circuits are sensitive to noise of certain frequencies.
Three ways to reduce voltage ripple when your power dissipation or voltage transformation forces you to use a switcher are:
Time: 8:26sTo compare power supply voltage ripple, I have both the buck regulator on the left and my LDO with the discrete power transistor on the right. Each one is using +12 volts in from the ATX power supply. Each one has the same load, two 8Ω power resistors in parallel to make a total of a 4Ω load, and on the previous slide you saw the input voltage ripples.
Time: 8:47sNow we are using the oscilloscope to measure the two output voltage ripples. Again, 5 volts output for the LDO and 5 volts output for the buck regulator and if we look at the oscilloscope, the buck regulator ripple in yellow and the LDO ripple is in blue.
At first you may say that they look almost the same, but most of the ripple we see in blue is the result of noise coupling from the buck. If we take out the probe we can see that the LDO noise is much much lower.
Time: 9:20sLinear regulators also radiate far less noise than the switching regulator, even a switcher the processes far less power. When I design power supplies it usually takes me as long or longer to design the filters and the electromagnetic noise reducing circuits as it does to design the switcher itself.
If you know someone who has taken their product for UL laboratories testing in the United States, or CE testing in Europe, then they may have stories of spending days on trial and error fixes for conductive or more commonly radiated noise that exceeded the limits. This is another reason why I use linear regulators whenever I can.
Time: 9:52sTo demonstrate electromagnetic interference (EMI), I have here an AM (Amplitude Modulated) radio tuned to somewhere around 600kHz which is in the range of the switching frequency of this buck regulator. Right now it is not “ON” and we can here some Spanish talk radio. When I turn “ON” the circuit, sure enough noise. When I get the radio closer to the source, which is the switching node and the inductor, the more interference we hear.
Once again for the EMI test I have the AM radio tuned to around 600kHz or so and we saw earlier with the buck regulator that as soon as I turned it “ON”, we got nothing but static noise. When I turn the the linear regulator “ON” we have approximately the same amount of output power, but although there is a little bit of interference, we can hear the AM radio just fine even if we get close to the power supply, and that’s the beauty of linear regulators, there is very little radiated noise.
Time: 10:59sOne more test. I now have the same radio but now it is tuned to FM (Frequency Modulated) 92MHz and sounds good. If I get the antenna (aerial) close then the interference starts again. This happens because not only does the switching inverter operate at maybe 400kHz to 1MHz, but it has a lot of harmonics and also has high frequency noise. Those extend and also interfere with the FM band.
Now we are going to test the linear regulator. I have here my FM radio and with the antenna close, I will switch it “ON” but have to be quick because this gets very hot as you can see it is drawing lots of current and voltage here, and no matter what I do we can continue to hear the Spanish radio.
That concludes part 5 of power supplies for non-EE’s. This is the last video tutorial in the series for now and it has been my pleasure making these videos. On behalf of myself and electronics-tutorials.ws I sincerely hope that you have learnt something and thanks again for watching.
End of transcription.
You can find more information and a great tutorial about the different types of power supplies by following this link about: Switch Mode Power Supply.