Making sense of mosfet datasheets

In fact, every page of a mosfet datasheet contains valuable information for the designer. What's not always obvious is how to interpret the data provided by the manufacturer. This article offers an overview of some key mosfet specifications, how they're expressed on datasheets and the fine print you need to be aware of to understand them. Like most electronic components, mosfets are affected by their operating temperature. So it is critical to be aware of the test conditions under which the specifications apply. It is also important to know whether the specifications you're looking at in a 'product summary' are 'maximum' or 'typical' values, since some datasheets from various manufacturers do not spell this out. Voltage ratings The primary characteristic by which a mosfet is identified is its drain-to-source voltage rating Vds, or 'drain-source breakdown voltage,' which is the maximum guaranteed voltage a mosfet can withstand with the gate shorted to the source. Vds is stated as an 'absolute maximum rating at 25°C', but it is important to be aware that this absolute is temperature-dependent and usually the datasheets include a 'Vds temperature coefficient.' You should also be aware that the maximum Vds rating is defined for the combined value of the dc voltage plus any voltage spikes and ripples existing in the power rail. For instance, if you're using a 30V device in a 30V power rail with 100mV, 5ns spikes, then you're exceeding the absolute maximum limit for the device and possibly are entering avalanche mode. The mosfet's reliability might not be guaranteed in this scenario. The temperature coefficient changes the breakdown voltage noticeably. For example, some mosfets with a 600V rating have a positive temperature coefficient that turns them into something near a 650V mosfet when running close to their maximum junction temperature. Some design rules call for 10% to 20% derating factors. In those designs. it may be beneficial to consider that the breakdown voltage actually gets 5% to 10% higher than the 25°C rated value and thus adds a corresponding design margin that can be useful in real life designs. Equally important for proper selection of mosfets is understanding the gate-to-source voltage Vgs used in on resistance Rds(on) ratings. This is the voltage at which a mosfet is guaranteed to be fully on at a given maximum Rds(on). This is why on-resistance ratings are always tied to the Vgs levels, and they are the only voltages at which the on-resistance is guaranteed. One important design consequence is that you cannot drive a mosfet fully-on with a voltage lower than the lowest Vgs used for the Rds(on) rating; for example, to be driven fully-on from a 3.3V microcontroller, you would need a mosfet with on-resistance rated at Vgs = 2.5V or lower. Applications requiring mosfet operation in the linear region would have a Vgs lower than those used in the datasheets for on-resistance ratings. Such applications are possible, but require special considerations due to potential linear operating mode thermal instability. The on-resistance of a mosfet is always specified at one or more gate-source voltages. The maximum Rds(on) limits can be 20% to 50% higher than typical values. Rds(on) maximum limits are typically shown at a 25°C junction temperature, yet at maximum junction temperatures, Rds(on) can go up by 30% to 150% (see fig 1). Given such variations and since no minimum value is typically guaranteed, Rds(on) based current sensing is not a very accurate technique. On-resistance is important for both n- and p-channel mosfets. Qg is a key selection criterion mainly for n-channel mosfets used in switchmode power supplies, since it affects switching losses. These losses have two aspects: one is due to the transition time of the mosfet turning on and off; the other is the energy required to recharge the gate capacitance at every switch cycle. Since Qg depends on the gate-source voltage, the switching losses can be reduced by using a lower Vgs. As a way to quickly compare mosfets being considered for switchmode applications, designers often use a single-number formula that encompasses the conduction losses represented by Rds(on) as well as the switching losses represented by Qg: Rds(on) x Qg. This 'figure of merit' (FOM) summarises a device's capabilities and either typical or maximum values can be used to compare mosfets. To ensure accurate comparisons among devices, you need to make sure that the same Vgs level is used for the Rds(on) and Qg and that typical and maximum values haven't been accidentally mixed together in the formula. A lower FOM would allow you to expect better performance in switching applications, but does not guarantee it. The best comparison results can be achieved only in actual circuits which might need to be tweaked for each mosfet individually in some cases. Most mosfets have two or more continuous drain current ratings in the datasheets based on different test conditions. You should look carefully at the datasheet to see whether the rating applies at an indicated case temperature (such as TC = 25°C) or ambient temperature (such as Ta = 25°C). Which of these is most relevant will depend on the specifics of the device and the application (see fig 2). Of course, the ambient temperature is measured not in the room where the product is installed, but rather inside the enclosure near the mosfet installed on the pcb. For small surface mount devices used in handheld applications, the most relevant current rating is likely going to be the one taken at a 70°C ambient temperature, while for larger devices that are used with heat sinks and forced air cooling, the current rating at Ta = 25°C may be more realistic. For some devices, the current level that can be handled by the silicon at its maximum junction temperature is higher than the level limited by its package; such 'silicon limited' current ratings are supplied additionally to 'package limited' ratings in some datasheets to give you an idea of the robustness of the silicon. Similar considerations apply to continuous power dissipation ratings, which are based not only on temperature but also sometimes on time. Consider a device with continuous Pd = 4W at Ta = 70°C for t = 10s. What constitutes a 'continuous' period of time will vary according to the thermal solution, so for convection cooling you need to use the Normalised Thermal Transient Impedance graphs on the datasheet to estimate the power dissipation rating after 10 or 100s or after 10min. As shown in fig 3, this particular device has a thermal impedance coefficient of ~0.33 for a 10s pulse, which means that, once the package is thermally saturated after about 10min, it will only be capable of dissipating about 1.33Wm rather than 4W, although the same device could dissipate about 2W if it is well cooled. Leo Sheftelevich works for Vishay Siliconix.