A good MOSFET datasheet is excellent when it contains detailed information. However, too much information might mean there is too much data to sort through. If you want to build a low-frequency switch, what is the minimum information you need to understand? As it turns out, you only need a few pieces of data from a table and to look at a couple of MOSFET curves to find most of what you need to know.
Figure 1 - N-Channel MOSFET Symbol
MOSFET stands for metal-oxide-semiconductor-field-effect-transistor. (Though if you look in a textbook, it probably says "s" stands for silicon.) The silicon layers form either N-Channel and P-channel types, which all have three terminals: gate, drain, and source. Last, two common functional types are "enhancement mode" and "depletion mode." These work best for switching applications. Enhancement mode FETs are normally "off" while depletion mode FETs are normally "on."
There are some additional details I'm glossing over. If you'd like to understand MOSFETs in more detail, check out The Learning Circuit 31: How FETs Function.
Key MOSFET parameters
When designing in a MOSFET as a switch, there are two parameters people tend to evaluate:
- Vgs-threshold. The voltage drop from gate to source that controls the channel between the MOSFET's drain and source.
- Rds-ON. The amount of resistance between the drain and source when the MOSFET is active.
Knowing the Vgs-threshold is critical because many high power MOSFETs have a Vgs in the range of 10 to 15 volts. Such a high threshold means you need a driver circuit when used with 3v3 or 5 volt embedded controllers like an Arduino, Raspberry Pi, or Beaglebone. Instead of a driver, you could try to find a MOSFET with a lower voltage threshold. Sometimes manufacturers call these transistors "logic level" MOSFETs. An example is the IRFZ44N.
Figure 2 - IRFZ44N's Vgs-threshold range
Figure 2 shows the IRFZ44N's Vgs-th. Notice how it is a range from 2.0 to 4.0 volts. Do not make a mistake pre-bald engineer made many years ago. This value is not the operational range for Vgs. This range is telling you anything above 4.0 volts guarantees a MOSFET is ON and anything below 2.0 guarantees the MOSFET is OFF. A voltage in-between may or may not turn on the MOSFET. But. There is even more to the explanation of the threshold voltage.
Side note. Many people, including experienced engineers, make the mistake of looking at the "Absolute Maximum Ratings" table. It usually comes first and looks similar to the electrical characteristics table. The values in the absolute max table tell the conditions that can (and do) damage a component. Do not operate at or exceed these values!
The other critical parameter is the resistance between drain and source, or Rds-ON. Like any resistance, the current through it generates heat and a voltage drop. If there is enough heat, a heat sink may be required. Do not overlook the voltage drop across the drain and source, Vds, due to the resistance.
So when using a MOSFET as a switch, you should model it as a resistor when turned-on, not a short. But how much resistance?
Figure 3 - IRFZ44N Datasheet Header showing (minimum) Rds-ON
The first page of the IRFZ44N's datasheet says the Rds-ON is 17.5 mOhm. Well, that was easy! Or was it? First, let me address the tiny amount. Someone asked me once: "Isn't 17.5 milliohms practically 0 ohms?" Well, no. No, it is not. Not only that, but the resistance from drain-to-source varies under a variety of conditions! Including, the applied voltage from the gate to the source (Vgs).
So, Vgs and Rds-ON have an essential relationship. (Later, you'll see the current through the drain matters too.) The bad news is that like most things in engineering; it isn't a simple relationship. It takes graphs to explain what is going on and a keen designer to interpret them correctly.
The good news is that you only need to look at two graphs. (And I would argue only one of those.)
MOSFET Graphs: where is Rds-On?
In the datasheet, look for a graph called "Transfer Characteristics" and one called "Output Characteristics." The output characteristics graph may also be called "on-region." Neither of these graphs has an axis for Rds-ON. Actually, most MOSFET datasheets do not. However, that's okay, between output and transfer characteristics, the resistance information is there.
Compared to the other stuff in the datasheet, the transfer characteristics graph is easy to read. It contains the drain current for when applying a specific voltage to the gate. Take note that this graph starts at 4.0 volts, which was the maximum value stated in the electrical specifications table.
Figure 4- Transfer Characteristics
To me, this graph is a quick way to see whether or not you need a driver for a particular MOSFET. If I were driving the IRFZ44 from a 3.3 volt microcontroller, and my load required at least 10 amps, I would be using a driver to increase the gate voltage. While 4 volts is enough to turn on this particular MOSFET, its on-resistance is going to be (relatively) high.
Output (On-Region) Characteristics
The output characteristics graph looks busy. But there is a ton of useful information packed into it. It shows the voltage drop across Vds, based on the amount of current going through the drain, which is also based on the Vgs voltage. There is an additional detail available.
Figure 5 - Output Characteristics
The orange section of the graph shows the linear operating region. The condition shown is the ideal "area" to operate a MOSFET when used as a switch.
RdsON=Vds =400mV =66.67mOhm
You can use Ohm's law to find the resistance with a known voltage and a known current. Since this graph shows the drain-to-source voltage (Vds) at a certain drain current (Id), it contains the on-resistance (Rds-ON). The blue line shows at 4.5V, the Rds-ON for the MOSFET will be 66 milliohms.
Measuring Output characteristics
You might notice that the datasheet uses the word “typical” for most of these measurements. That word is significant. It means individual transistors perform differently. To get a specific transistor’s MOSFET curves, you need to measure it.
How? Well, with something called a curve tracer. That device is test equipment which shows performance characteristics across different conditions. For example, they apply voltage to a MOSFET's gate while measuring the current through its drain. Sweeping across a range of voltages creates enough data points to draw, or trace, graphs.
The equipment used to draw diagrams found in component data sheets is somewhat expensive. However, there is a low-cost tool that can provide much the same functionality, though with some limits.
Recently, the element14 community sent me a DCA Pro from PEAK Electronics. It is a semiconductor analyzer. (Previously I wrote about the ESR70, which measures capacitance and ESR.) This small portable device runs about 150 USD. For a wide range of semiconductors, it can determine pinouts, measure key characteristics, AND draw parameter curves. It works with didoes, zeners, LEDs, BJTs, MOSFETs, IGBTs, JFETs, SCRs, Triacs, and EVEN voltage regulators. (Really!)
Determining pinouts and testing components is the primary reason to get the DCA Pro. There is a bit of an issue with the curve tracing. Look at Figure 6, which is comparing the output characteristics graph from the datasheet and the DCA pro, side-by-side. Notice how their shapes are quite a bit different. Both graphs have the same axes titles, and both have multiple Vgs lines. However, that's the only similarities. But why?
Figure 6 - Output Characterization Comparison
The reason is the limited current of the DCA Pro. It can only drive a few milliamps, up to 10-12 mA, through the drain. Let's combine that fact with three points from this post:
- This graph shows the voltage drop from drain to source.
- The Vgs voltage sets the resistance between drain and source.
- MOSFETs obey Ohm's Law when operating in their linear region.
Since our semiconductor analyzer can only source 10 milliamps, the measured Vds is significantly limited. In the case of this IRFZ44, we are comparing drain currents of milliamps to amps. So the displayed data is correct but seems off. (Also, the datasheet is using a logarithmic scale while the DCA Pro software uses linear scaling.) Does this comparison mean the DCA Pro is worthless for measuring power semiconductors? No, not at all. I very much like the DCA Pro and think it is worth its price. Just do not expect it to replace a $10,000 curve tracer.
More Details on the DCA Pro
If you're interested in learning more about the device, check out my full video review. In the video, I compare how it performs with a variety of semiconductors. I was happy to see that it could correctly identify components like an N-Channel Depletion Mode MOSFET. You can also see the DCA Pro software in action. Over on the Workbench Wednesday's post, I have a zip file with example data from components I measured that you can use in PEAK's software.
As a standalone device, the DCA Pro is already fantastic. As a curve tracer, it is better for low power devices like diodes. However, even for large current transistors, it can still hint at a transistor's performance. With the sweep in Figure 7, look at how this transistor follows the min and max Vgs thresholds from the datasheet. Right at 3.00 volts, Vgs is just high enough to turn on the MOSFET barely. At 2.75 and below, the MOSFET Is entirely off. At 3.175V and higher, it turns on. Just like the datasheet says. Pretty cool, huh?
Figure 7 - DCA Pro Finding Vgs Threshold
MOSFETs are a complicated device. Using this information, you should be able to look at a couple of graphs and tables in a MOSFET to get a good idea of how to operate them. For the cases it isn't enough, what other parts of the datasheet do you need help reading? Let me know in the comments below.