Transcription

1 January 2012 page 1 Measuring Leakage Current in RF Power Transistors By Roger Butler, Sr. Product Application Specialist Richardson RFPD, Inc. Abstract The published specifications for leakage current in RF power devices are often a source of concern and confusion for engineers and technicians. This paper examines the real meaning behind the leakage current specifications and offers guidance on properly testing a device for leakage current. Introduction Around the globe, engineers and technicians using RF power devices have concerns regarding the published specifications for leakage current: what the specifications mean in terms of the part s performance in the field, and most importantly, how to properly test and verify that a given part is meeting its stated leakage current specification. Transistor Leakage Current Defined A transistor can be thought of as a simple ON/OFF semiconductor device. Ideally, the transistor only allows DC current to flow through it when it is ON (i.e. properly biased and with the proper DC supply voltages applied). By the same token, the transistor ideally allows zero DC current to flow through it when it is OFF. In reality, however, a small amount of DC current still flows through all transistors when they are in their OFF state, as long as the DC supply voltages are applied. This relatively low-level of DC OFF current is commonly referred to as transistor leakage current. Leakage current is present in every type of transistor, using any semiconductor technology Bipolar 1, CMOS, VMOS, LDMOS, GaAs, GaN, etc. Normal leakage current is the expected amount of leakage current that is within a given part s specifications. Generally speaking, normal leakage current is due mainly to imperfections and limitations in the transistor die, the details of which are beyond the scope of this paper. Leakage Current Specifications for Transistors Leakage current is specified in the vast majority of transistor data sheets, but often goes unnoticed and is rarely a cause for concern by RF power design engineers. This is because leakage current is typically very low, usually in either the low µa (micro-amps, or 10-6 amps) range or even the na (nanoamps, or 10-9 amps) range. Since leakage current is so low, it is only considered design-impacting when the transistor is used in extremely low power applications, or, in rare cases, when the transistor is used in designs where extremely tight bias current limitations exist. The other instance in which leakage current receives attention is when a transistor unit in the field is found to be out of specification with regard to its published leakage current specification. Understanding Stated Leakage Current Specifications Figure 1 shows an excerpt of an actual data sheet for a Field Effect Transistor (FET). For FETS, leakage current is usually specified for both drain-to-source current (I DSS ) and for gate-to-source current (I GSS ).

2 January 2012 page 2 Fig. 1: Example of leakage current specifications for a FET device. Notice that the specifications for leakage current are dependent on certain conditions of the transistor device under test (DUT), as follows: As noted in the table header, the device s electrical characteristics are all to be tested at a case operating temperature of T C = 25 C (unless otherwise noted). The Zero Gate Voltage Drain Leakage Current, I DSS, is specified twice as a maximum value, with two different drain-tosource voltages (V DS = 66 V DC and V DS = 28 V DC ). In each case the gate-to-source voltage is specified at zero volts (VGS = 0 V DC ), i.e. the gate and the source of the DUT are shorted together to properly test I DSS. I DSS Max (i.e., the maximum allowable value for drain leakage current), is clearly specified for this device as up to 10 times higher with V DS = 66 V DC than it is with V DS = 28 V DC. Gate-Source leakage current, I GSS, is specified once, with V GS = 5 V DC and V DS = 0 V DC, which means the drain and the source of the DUT are shorted to test I GSS. The bottom line is that stated leakage current specifications only apply to devices tested within the stated testing conditions. Proper Conditions for Measuring Leakage Current To determine whether a given transistor sample meets its stated leakage current specification, absolute care must be taken to properly test and evaluate the device. As explained in the previous section, always test the device per the manufacturer s stated test specifications (temperature, voltages, shorts, etc.). Furthermore, it is important to adhere to proper testing conditions: The device itself must be free of foreign substances, dirt, dust, and other contaminants indeed, the leakage current attributed solely to contaminants on the board (even solder flux and fingerprints) can be higher than the leakage current through the DUT itself. Isolate the DUT. The device cannot be properly isolated and tested when soldered into a board with other parts connected. The other parts will alter the test results because they are part of the test. Always use properly calibrated laboratory-grade test equipment. (See below.) Never use a battery operated multimeter to test leakage current. Comply with all industry Electrostatic Discharge (ESD) requirements. Employ proper lab grounding techniques (this applies to equipment and technicians). Ensure that technicians are trained properly for the task.

3 January 2012 page 3 Use a proper, calibrated test fixture and shielded test leads. Providing shielding from light (for the DUT) and some filtering/shielding from other noise sources (AC line, RF, etc.) may be necessary to detect low na currents. Proper Test Equipment The preferred method for testing leakage current in RF transistors is either using a semiconductor device analyzer or a curve tracer. Fig. 2 Calibrated curve tracer The curve tracer is a single piece of test equipment, shown in Figure 2, that is able to provide a graphical display of the current versus voltage. It allows the technician or engineer to test the transistor to the test conditions stated in the data sheet. Besides leakage current, it is also able to measure current gain and breakdown voltages. Many of the latest models of semiconductor device analyzers available today offer curve tracers as an option. Common Field-Testing Mistakes Engineers and technicians know that a working transistor with too much leakage current, that is to say a working transistor with out-of-spec leakage current, can indeed be a problem in the field. Such a device can cause early field failures, exhibit poor performance, be an unnecessary drain on the battery (in very low power applications), and even induce noise into the channel. When problems occur in a circuit, it is good troubleshooting procedure to properly test the key devices to make sure they are performing within the manufacturer s specifications. Here are a range of field-testing mistakes commonly encountered: Mistake #1: Technicians testing an RF power transistor with a battery powered multimeter (using either the ohm meter or

4 January 2012 page 4 diode setting). Result: It is impossible to properly test leakage current with a battery-powered multimeter. The output voltage and the meter s impedance are completely unknown. This leads to invalid, unreliable readings, and will most likely cause damage to the RF transistor. Remember, the DUT must be tested under the correct voltages and the correct conditions to guarantee reliable and repeatable results. A battery-powered multimeter can only reliably be used to test whether a DUT that is already clearly defective is either open or shorted. Mistake #2: Technicians attempting to test an RF power transistor for leakage current while the DUT is still in-circuit (i.e. still plugged- or soldered-into the circuit). Result: Trying to test leakage current while the device is in-circuit, even if the circuit is powered-down, will provide erroneous and invalid measurements. The device cannot be properly isolated and tested when plugged- or soldered-into a circuit with other parts connected to it. These other parts will alter the test results because they are indeed part of the test. Also, the leakage current attributed solely to contaminants on the printed circuit board (even solder flux and fingerprints) can be higher than the leakage current through the DUT itself. Mistake #3: Static discharge is applied to the DUT through improper device handling. Result: Static discharge can and likely will permanently ruin an otherwise good device. Mistake #4: Improper grounding, shielding, and/or isolation methods are employed with the DUT. Result: Testing with improper shielding, grounding, and isolation will lead to erroneous and invalid measurements; and it may also damage the part. Mistake #5: Uncalibrated test equipment is used. Result: Using uncalibrated test equipment will produce erroneous and invalid measurements. Mistake #6: Devices are not tested at the proper temperature. Result: If the parts are not tested to all manufacturers specifications, then there is no way to prove that a part does not meet its specifications. In all of these examples, improper leakage current testing leads to unfortunate consequences. Both good and bad devices are being damaged unnecessarily, making further analysis impossible. Some devices that are actually within specification are being thrown away, rejected or returned because of erroneous test results. (This also applies to the incoming inspection stage, in some cases.) Finally, root causes of other RF transistor problems are being masked undetected and undiagnosed when leakage current is improperly tested and labeled the cause of a problem in a circuit. Conclusions All RF components and subassemblies used in complex and critical designs (i.e. military, avionics, broadcast, etc.) should be parametrically evaluated in order to make sure that they meet published specifications. Component testing must be performed properly. Carefully review testing conditions and methods to insure that leakage

5 January 2012 page 5 current testing is safe, calibrated, reliable, and repeatable. 2 Notes 1. Leakage current for bipolar junction transistors (BJT) is commonly referred to as ICEO, the collector-emitter cutoff current (base open). This is one reason why you seldom find the words leakage current in older transistor data sheets and data books. With the advent of Field-Effect Transistors (FETs), and the subsequent FET technology advancements (VMOS, LDMOS, etc.), BJTs have decreased dramatically in usage as RF power transistors. 2. Richardson RFPD helps customers to perform DC testing (including leakage current) and/or RF testing on components and assemblies that we supply. In our testing facilities, fully automated test systems are employed to test volume production parts, including those parts on tape and reel. Critical component testing can increase field reliability, improve product performance, and save cost. Richardson RFPD uses trained technicians, fully-calibrated test equipment, and adheres to all manufacturer and industry testing specifications. More information is available at richardsonrfpd.com.