Introduction

When you look at a MOSFET datasheet, you'll almost always come across the term "on-resistance RDS(on)".
On-resistance is the drain-source resistance when the MOSFET is turned on, and it is one of the important characteristics to consider when thinking about the conduction losses of a MOSFET.
Even in existing articles on the fundamentals of MOSFETs, MOSFETs are described not as complete ON/OFF switches, but as devices whose on-resistance changes depending on the gate-source voltage.
Link: MOSFET Basics -Semiconductor Business -Macnica 

So, "If we actually prepare a MOSFET and measure the drain voltage (VDS) and drain current (ID), will we be able to measure the same on-resistance as in the datasheet?"
This time, we young engineers actually tried our hand at the measurement. To put it bluntly, it wasn't that easy.

Consider the on-resistance measurement circuit.

For this measurement, we used a MOSFET in a TO-220F package that was available in the laboratory.
Upon reviewing the datasheet, it is specified that the maximum value of the on-resistance RDS(on) under the conditions VGS = 4.0 V and ID = 10 A is 60 mΩ, as shown in the figure below. Therefore, in this study, we considered the measurement circuit using these VGS and ID values as target conditions.

However, I stumbled right away.
Initially, I thought, "Wouldn't it be enough to apply a DC voltage (Vgs=4.0V) to the gate and measure it?" I also tried to measure it without considering the current-limiting resistor.
However, after researching various sources, I arrived at the following measurement circuit.

Our measurement circuit

Based on the datasheet, we assumed the MOSFET's on-resistance to be approximately 60mΩ, and set Vin=11.6V and R=1.1Ω as a guideline to achieve an ID of approximately 10A.
However, the actual current value will vary depending on factors such as resistance tolerance, wiring resistance, power supply voltage fluctuations, and MOSFET temperature.

Regarding the equipment used

The following equipment was used for this measurement.

Measurement overview diagram

Day 1: I thought I got the measurement! ...But it turned out I hadn't measured it correctly.

These are the waveforms obtained on Day 1. I will provide further details about each waveform.
1ch: Vgs (yellow waveform)
2ch: Vds (blue waveform)
3ch: Ids (red waveform)
Channel 4: Vds/Ids (orange waveform) *Channel 4 was calculated using the Math function of the oscilloscope.

In the first measurement, an on-resistance value of 65.304 mΩ was obtained.

However, upon examining the waveform, it became clear that the channel measuring VDS and ID was clipping, suggesting that a proper waveform might not have been acquired in the first place.

For this measurement, VDS was set to 200 mV/div and ID to 200 mA/div.
The display range was too narrow for the measurement target, and the VDS and ID waveforms exceeded the oscilloscope's display and measurement range.
In this state, parts of the waveform are not acquired correctly, and the calculated waveform VDS/ID cannot be trusted.

This was the first major lesson learned from this experiment.
Just because a number is displayed on an oscilloscope doesn't mean that the number is correct.
You need to check whether the measurement is reliable, including the measurement range, clipping, and probe settings.

Day 2: Review the oscilloscope range and try again.

Therefore, I reviewed the oscilloscope settings and range, and measured again to avoid clipping. The waveform is shown in the figure below.

The second time, VDS ≈ 0.57V and ID ≈ 9.25A.
The result was RDS(on) ≈ 0.57 ÷ 9.25 ≈ 62.4 mΩ.

However, some questions remain.
According to the datasheet conditions, the maximum RDS(on) value at VGS=4.0V and ID=10A is 60mΩ.

In contrast, the measured value was approximately 62.4 mΩ.

"The measurements should have improved, so why are they still higher than the datasheet values?"
Among us young engineers, we've hit a dead end here.

In the second part, with advice from experienced engineers, we will examine each possible cause of the discrepancy in the measured values.
The second part is scheduled to be released around January 2027.

Inquiry

If you are interested in the topics discussed in this article, please feel free to contact us.