What slows down design progress? One category of "being stuck" has to do with being fooled – thinking that the data is saying something that it is not. To demonstrate how one might be fooled in this way in power electronics, the following situation is posed as a quiz.
The following circuit was devised to measure the inductance of T2.
The magnetic device under test is a common-mode power-line filter transformer (or "choke"). The MOSFET is driven by a Tek PG508 pulse generator. The on-time is adjusted as the current, measured on channel 2 of a Tek TDS360, is observed. Channel 1 of the 'scope is connected to the drain. The following waveforms were observed.
The inductance can be calculated from
The transformer inductance was measured with two different RLC meters (an ESI model 253 and a B&K 875A). Both gave values of around 4 mH, obviously in disagreement with the test-setup value. What is wrong here?
A more careful observation of the above waveforms provides the first clue. Theoretically, a constant voltage applied to an inductor should cause a ramp of current, but the ramp does not take off until about 25 μs after the voltage is applied. Why?
The error is one of scale, like viewing a familiar object too close or too far away, thereby making it unrecognizable. The scale factor of channel 2 is set at 1 A/div. If the vertical scale of channel 2 is decreased to 200 mA/div, then the next oscillograph can be observed.
Now the current ramp is evident, after the initial diode switching spike just to the right of the left-most (dotted) vertical cursor. The current rises linearly to about 100 mA (2.5 minor vertical divisions) in 20 μs at the center of the display, with an applied voltage (see channel 1) of about 14 V. Calculating inductance, the value is now
This value is much closer to the 4 mH measurements of the RLC meters. The larger value from the meters is probably due to their measurement at a lower value of current excitation than 0.1 A, with even less saturating effect than shown in the oscillograph above. This particular transformer (which is typical of common-mode line-filter transformers of the 25 mm EE core size) shows a knee in its current curve at about 300 mA. This is revealed by, again, selecting the right scale-factor for observing it, as shown below.
At 200 mA, the curve turns sharply upwards, and is clearly beginning to saturate already at around 100 mA.
Although the first oscillograph correctly displays magnetics behavior at a large (1 A/div) scale factor, it shows the transformer core deeply in saturation. The first inductance calculation is more indicative of its leakage inductance than its magnetizing inductance, demonstrating that ferrite material and air have permeability values differing by orders of magnitude. The lesson: beware of being fooled by interpreting phenomena at the wrong scale-factor.
