Discrete device current detection method

Holding Current Testing Methods for Discrete Semiconductor Devices

A thyristor turns on just fine. The gate pulse fires, the device latches, everything looks good. Then the load current dips below a critical value and the device drops out of conduction without warning. No error signal. No gradual decline. Just off. That critical current is the holding current, and if you do not measure it correctly, your circuit will misbehave in ways that are almost impossible to debug.

Holding current (IH) is the minimum anode current required to keep a thyristor, SCR, or TRIAC in the on-state after it has been triggered and the gate signal has been removed. Below IH, the device reverts to the off-state. It sounds like a simple number. It is not. The value shifts with temperature, with the rate of current decay, and with the device's internal carrier distribution. A test that ignores these factors gives you a number that works in the lab and fails in the field.

This article covers how to test holding current properly, what causes the measurement to drift, and the setup mistakes that turn good data into fiction.

Why Holding Current Matters More Than Latching Current

Everyone talks about latching current. Almost nobody talks about holding current. That is a problem.

Latching current (IL) is the minimum current needed right after turn-on to keep the device latched while the gate pulse is still present. Holding current (IH) is the minimum current needed after the gate pulse is gone. IL is always higher than IH, typically by a factor of 2 to 4. Once the device is fully on, it needs less current to stay on than it did to get on.

In a real circuit, the current through the device is never constant. It ripples. It dips during zero-crossing in AC applications. It sags under transient load. If the current dips below IH even for a microsecond, the device turns off. And when it turns off, it does not turn back on until the next gate pulse. In an AC phase-control circuit, that means a missing pulse. In a motor drive, that means a dead leg. In a power supply, that means output voltage collapse.

Testing holding current tells you the true operating margin of your device. Latching current tells you whether the device will turn on. Holding current tells you whether it will stay on. Staying on is the hard part.

The Physics Behind Holding Current

A thyristor is essentially two transistors connected in a positive feedback loop. When the gate triggers the device, both transistors turn on and keep each other on. The holding current is the minimum collector current needed to sustain that feedback loop.

Below IH, the collector current of one transistor drops too low to provide base drive to the other. The loop breaks. Both transistors turn off. The device returns to the blocking state.

This mechanism is temperature-dependent. At high temperature, carrier lifetime increases, which means the transistors stay on with less base current. IH decreases with temperature. At low temperature, carriers recombine faster, and you need more current to keep the loop alive. IH increases with temperature drop.

The datasheet gives you a typical IH at 25°C. That number is almost useless for design. You need to know IH at your maximum operating temperature and at your minimum expected load current. The gap between those two values is your real design margin.

How to Test Holding Current Correctly

The test concept is simple. Trigger the device, remove the gate signal, reduce the anode current slowly, and measure the current at which the device turns off. The execution is where things fall apart.

Setting Up the Test Circuit

The test circuit needs three things: a trigger source, a controllable anode current, and a way to detect when the device turns off.

Use a pulse generator to apply a gate trigger pulse. The pulse must be wide enough to fully latch the device. A pulse that is too short will not establish the feedback loop, and the device will drop out before you even start measuring IH. Use a pulse width of at least 100µs for most thyristors. For high-current devices, go up to 1ms.

The anode current must be controllable and stable. Use a programmable current source or a DC supply with a series resistor and a variable voltage. The current must be adjustable in small steps, ideally 1mA or less near the expected IH value. If your current steps are too coarse, you will miss the exact turn-off point and your IH reading will be off by 10% or more.

To detect turn-off, monitor the anode-cathode voltage. When the device is on, Vak is low, typically 1V to 2V. When the device turns off, Vak jumps to the supply voltage. That voltage step is your turn-off indicator. Use a comparator or an oscilloscope to catch the exact moment of transition.

The Current Ramp Method

The most reliable method ramps the anode current down slowly after the device is latched.

First, trigger the device with a gate pulse. Confirm it is on by checking Vak. Then remove the gate signal. Wait 10µs to let any transient settle. Then begin reducing the anode current at a controlled rate, typically 1mA/ms to 10mA/ms.

Watch Vak on the oscilloscope. The moment Vak jumps from the on-state voltage to the off-state voltage, note the anode current. That current is your holding current.

The ramp rate matters. If you ramp too fast, the device does not have time to respond, and you measure a current higher than the true IH. If you ramp too slow, the device heats up during the test, and IH shifts lower because of the temperature rise. Use a ramp rate of 5mA/ms as a starting point. Adjust based on the device size and the expected IH value.

Run the test at least five times per device. Holding current has measurement noise. The turn-off point is not always sharp. Sometimes the device flickers between on and off for a few microseconds before settling. Take the lowest stable reading, not the first one.

Temperature Effects on Holding Current

Temperature is the dominant variable in holding current testing. Ignore it, and your data is worthless.

How IH Shifts Across the Temperature Range

For most silicon thyristors, holding current decreases as temperature rises. The typical temperature coefficient is -0.5%/°C to -1%/°C. A device with IH of 10mA at 25°C might have IH of 7mA at 125°C. That 3mA difference is enormous in a low-current application.

At low temperature, IH increases. A device that holds at 10mA at room temperature might need 15mA at -40°C. If your circuit only provides 12mA, the device works in summer and fails in winter.

Test at three temperatures minimum: the minimum operating temperature, 25°C, and the maximum operating temperature. Plot IH versus temperature. The slope tells you how much margin you need at the cold end to survive the hot end.

For automotive applications under AEC-Q101, test at -40°C, 25°C, 85°C, and 125°C. For industrial applications under IEC 60747, test at -25°C, 25°C, and the maximum rated temperature.

The Danger of Self-Heating During the Test

When you run current through the device to measure IH, the device heats up. The junction temperature rises above the ambient temperature, and IH drops. You end up measuring a lower IH than the device actually has at the intended temperature.

Use short test pulses. Apply the holding current for no more than 10ms before measuring. Let the device cool for at least 100ms between measurements. This keeps the junction temperature close to the ambient temperature and gives you a reading that reflects the real condition.

If you must run a continuous current test, measure the case temperature and calculate the junction temperature using the thermal resistance from the datasheet. Correct the IH reading to the actual junction temperature. This is more work, but it is the only way to get an accurate number under continuous operation.

Common Errors That Corrupt Holding Current Data

The most frequent mistake is measuring IH immediately after triggering. The device needs time to establish the internal carrier distribution. If you start reducing the current within 10µs of the gate pulse, the feedback loop is not fully stable, and the device turns off at a current higher than the true IH. Wait at least 50µs after the gate pulse before you start the current ramp.

The second mistake is using a resistor to set the anode current instead of a current source. A resistor lets the current change as Vak changes. When the device turns off, Vak jumps, and the current through the resistor changes instantly. You lose control of the current at the exact moment you need it most. Use a true current source with fast compliance response.

The third mistake is not accounting for gate leakage current. Some thyristors have significant gate-cathode leakage at high temperature. That leakage acts like a small gate current and keeps the device partially on even after you remove the gate pulse. Your measured IH will be lower than the true value because the device is not fully dependent on anode current to stay on.

To check for this, measure the gate-cathode leakage at your maximum test temperature. If it exceeds 100µA, your IH measurement is contaminated. Use a device with lower gate leakage or correct the reading by subtracting the leakage contribution.

The fourth mistake is testing only one device per lot. Holding current varies with process spread. A single device might measure 8mA. The next one from the same lot might measure 12mA. If you design for 8mA, half your production will fail. Test at least five devices per lot. Use the maximum IH value for your design, not the typical.

Defining Pass/Fail Criteria for Holding Current

A holding current test without clear limits is just a measurement. It does not tell you whether the device is acceptable.

Setting the IH Limit Based on Your Application

The datasheet gives you a maximum IH. That is your starting point. But the maximum IH on the datasheet is measured under specific conditions that may not match your application.

For production testing, set your fail limit at the datasheet maximum IH at the worst-case temperature. If the datasheet says IH max is 10mA at 25°C, and your application runs at 125°C where IH is typically 7mA, use 10mA as your fail limit. Do not derate the limit. The device must hold at the datasheet maximum under all conditions.

For characterization, record the actual IH at each temperature and plot the curve. Look for devices that are close to the limit at high temperature. Those devices have no margin and will fail in the field when the temperature rises.

Sample Size and Statistical Treatment

Test five devices per lot for production. Test ten for incoming inspection. Test twenty-five for characterization.

Plot the distribution. If the spread is tight, your process is stable. If the spread is wide, your process has variation that your circuit must tolerate. A bimodal distribution means you have two populations in the lot, possibly from different process runs. Do not ship that lot without sorting.

Record the temperature, the gate pulse width, the current ramp rate, the Vak at turn-off, and the raw waveform for every sample. Holding current is sensitive to every one of these variables. If you do not record them, you cannot reproduce the test when a field failure comes back.

Holding current is the parameter that keeps your thyristor on when everything else wants to turn it off. Test it at the right temperature, with the right ramp rate, after the right delay, and with enough samples to see the spread. A device that passes every other test but fails holding current will drop out of conduction at the worst possible moment. That is the failure you do not want to find in the field.