MOSFET-熱阻測(cè)試.doc

Measurement of MOS FET devices by booster deviceMOS FET devices can be measured using three different electrical setups.1. Reverse diode measurement In the first setup we use the reverse diode of the device for heating and measurement also. The power step is applied using the so called current step method by switching from a higher Iforce+Isense current level to Isense measurement current and using the forward voltage as temperature sensitive parameter.to T3Ster channelIsenseIforceBoth Iforce and Isense can be programmed in the booster which is a unipolar device of positive polarity. The device voltage is typically below 1V.2. MOS diode measurement In the second setup we connect the Gate to the Drain pin and measure this three terminal device as a so called MOS diode. We apply a similar current step as described above and measure the forward voltage of the MOS diode.to T3Ster channelIsenseIforceThe device voltage is typically 2 . 5V corresponding to gate threshold voltage.3. Measurement with heating on channel resistance and sensing on reverse diodeIn this third setup the electrical connection is a bit more complex. to T3Ster channelIsenseIforceUGIn the heating phase the UG voltage is at high level ensuring continuous heating current through the channel resistance RDSon. The positive heating current comes from the booster and a negative sense current is generated by an other equipment, e.g. T3Ster.At the start of the cooling Iforce is switched off and UG also goes to zero thus closing the transistor. The continuously flowing negative Isense causes the reverse diode to open.The device voltage is typically low on the channel resistance and -0.5V on the reverse diode at cooling.A practical problem is that the booster output is of devices of large surface in order to switch hundred amperes. This can cause leakage at negative voltages, resulting in loss of the measurement current from external source and noise in the measurement.A simple “tandem” circuitry can stop this leakage while withstanding the current levels of the measurement.to T3Ster channelIsenseIforceUGT1T2DUTT1 and T2 are transistors of the same type and characteristics. T1 is the device under test. T2 behaves as a control switch which is ON during heating but OFF during cooling such isolating T1 during the cooling measurement.During the heating Iforce is switched on and the gates are connected to a high voltage level i.e. 10 V. The Iforce Isense heating current flows through the channel of T1 generating heating power on RDS,on. For measurement the heating current is suddenly switched off and the gate voltages are also switched to 0 V. The Isense current flows through the reverse diode of T1 (and not through T2) and the forward voltage of the reverse diode is used for transient measurement.4. Comparison of measurement resultsWe carried out the above described three measurements on a MOS FET device in TO220 type package without disassembling the mechanical setup. After that we repeated the set of measurements with a different boundary condition.The measurement parameters for the three methods were the following:TypeUhigh VUlow VIforce AIsense APower step WReverse diode0.920.56100.019.22MOS diode2.571.9720.015.13RDSon (sat)0.41 0.57100.014.14In the figures below the comparison of the curves corresponding to the three different setups can be seen. In all charts revdio (blue curve) corresponds to setup 1, Mosdio (red curve) to setup 2 and sat (green curve) to setup 3. After about 1 K/W the three curves fit well, but before this point the curve belonging to sat (green curve) is significantly different. This hints that the heating surface and such the heat spreading from the surface is different when the channel resistance is heated. TTEL: 158-2079-5808Using the measurement results corresponding to the two different boundary conditions we can identify the junction-to-case thermal resistance as it can bee seen in figure below.The green and black curves diverge at around 0.8 K/W. It can be stated that the three curves are similar only outside of the package; inside the package in case of setup 2 the heat spreading is different.