Reading a MOSFET datasheet: check gate drive and thermal conditions before using RDS(on)
Read MOSFET threshold voltage, on-resistance and thermal ratings correctly, with IRFZ44N and IRF7726 datasheet examples and a conduction-loss calculation.
If a MOSFET runs hotter than expected, first check whether the on-resistance used in your calculation is specified for the gate drive in your circuit. The small milliohm figure on the first page is not enough. Gate–source voltage, test current and junction temperature determine whether that value applies.
This distinction is easy to miss when a 3.3 V microcontroller switches a load. A threshold-voltage specification does not establish the gate voltage needed for low-loss operation.
Crossing the threshold does not establish on-resistance
Consider the IRFZ44N part number visible in the photograph. The manufacturer's IRFZ44NPbF datasheet specifies a gate threshold voltage, VGS(th), of 2.0–4.0 V, tested at a drain current of 250 µA. Its maximum on-resistance of 17.5 mΩ is specified at VGS=10 V, ID=25 A and a junction temperature of 25°C. That resistance test also specifies a pulse width of no more than 400 µs and a duty cycle of no more than 2%. IRFZ44NPbF datasheet, page 2
VLSI Korea view
“The threshold starts at 2 V, so 3.3 V is enough” does not follow. The conditions for conducting 250 µA are different from those for carrying several amperes with low loss. The 17.5 mΩ figure therefore cannot simply be inserted into a thermal calculation for 3.3 V gate drive.
Check VGS at the MOSFET terminals, rather than relying on the microcontroller's nominal output voltage. VGS is gate voltage minus source voltage. Establish the source potential and circuit reference, then look for a maximum on-resistance specified at the available drive. The manufacturer's explanation likewise defines threshold voltage at a specified small drain current. Infineon datasheet explanation, section 2.10
Even 26 mΩ and 40 mΩ on the same part need context
The P-channel IRF7726 makes the distinction explicit. Its datasheet gives two maximum on-resistance specifications.
| Device | Maximum RDS(on) | VGS | ID | TJ |
|---|---|---|---|---|
| IRF7726PbF | 26 mΩ | −10 V | −7 A | 25°C |
| IRF7726PbF | 40 mΩ | −4.5 V | −6 A | 25°C |
Test conditions for both rows: pulse width ≤400 µs; duty cycle ≤2%. Source: manufacturer datasheet, page 2.
When evaluating −4.5 V drive, the relevant entry is 40 mΩ. Selecting the smaller 26 mΩ number merely because it looks better substitutes the −10 V test condition for the actual drive. IRF7726PbF datasheet, page 2
Dividing 40 by 26 does not establish the loss increase in a real circuit, either. The test currents differ, and both values come from short-pulse tests at 25°C. A circuit comparison must account for load current and operating temperature. These two devices are examples of how to read specifications, not recommended substitutes: IRFZ44N is N-channel, while IRF7726 is P-channel.
Calculate losses after choosing the applicable resistance
If resistance can be approximated as constant during conduction, average conduction loss is:
Pcond ≈ IRMS2 × RDS(on)
Here, RMS current is evaluated over the entire interval, including the time when the device is off.
For a sensitivity example, assume 5 A RMS and 20 mΩ at the operating conditions. Conduction loss is 0.50 W. Changing only the assumed resistance to 30 mΩ gives 0.75 W. These are hypothetical inputs used to explain the calculation, not measurements or guaranteed values for either device discussed above.
The equation alone cannot settle the thermal question. The manufacturer's explanation treats on-resistance as temperature-dependent and junction-to-ambient thermal resistance as dependent on board layout. Combining a room-temperature resistance with a thermal resistance measured on a different board does not validate your design. Infineon datasheet explanation, sections 2.1, 2.2 and 2.9
In practice, check resistance at the expected temperature, calculate loss, and estimate temperature through the actual cooling path. Check that the result is consistent with the starting assumption. Also distinguish a curve labelled “Typical” from a specified maximum limit.
The condition hidden beside a 49 A rating
The IRFZ44NPbF datasheet lists 49 A continuous drain current at VGS=10 V and TC=25°C. TC means case temperature, not ambient air temperature. At a case temperature of 100°C, the listed current is 35 A. IRFZ44NPbF datasheet, page 1
VLSI Korea view
“The load is 10 A and the part is rated for 49 A” is therefore not enough to finish the heatsink or board design. You need the loss at the load current and a path for that heat to leave. Check whether the circuit can maintain the case-temperature condition behind the rating.
For each candidate, record actual VGS, maximum on-resistance at that drive, expected junction temperature and cooling conditions beside the part number. Missing entries then point to specific work. If the available drive lacks a matching resistance specification, reconsider the part or gate-drive circuit. If thermal conditions are uncertain, establish the board conditions and measurement plan before claiming margin from the resistance figure.
This article focuses on reading conduction-loss specifications and ratings. PWM transition losses, inrush current and safe operating area during linear operation require separate checks. A low on-resistance does not complete them.
Sources checked September 29, 2026. The referenced documents are IRFZ44NPbF Rev. 2.1, dated August 11, 2026, and IRF7726PbF, dated February 22, 2005. English edition prepared September 30, 2026. Written with AI assistance and checked against manufacturer documentation. The explicitly labelled calculation example uses hypothetical inputs, not bench measurements.
Disclosure
This article is provided for informational purposes only and does not constitute a recommendation to buy or sell any security. Figures and factual statements are based on publicly available sources, including company filings, news reports, and market research, and may contain errors in the source material or in the preparation of this article. The author may own shares in companies mentioned. All investment decisions and their outcomes are the reader's responsibility.