« ESR Compensation | Main | Electrolytic Capacitor Failure »

June 05, 2001

Magnetics Saturation

Question: Can you discuss magnetic core saturation?

The cause, effects, and prevention of core saturation (cores ratchet up the hysteresis loop until they saturate) will be a problem/solution topic in the future but in the interim I get email questions on the subject. Here is a set of recent questions.

Original Question: How do you ensure the proper resetting of the flux in the core for a current transformer used in a single ended forward converter? In the forward converter, the current through CT is unidirectional. The flux must be reset during the period when no current is flowing. How this can be implemented? N. S. 05-17-2001

Answer: Cores will usually reset themselves to the residual flux density and if the core is gapped this is often enough. Also parasitic or intentional capacitance my reset the core. For any circuit and magnetic, you need to look at the hysteresis loop and see if it is doing what you want in a controlled manner. If not, you have to add reset circuitry to force the hysteresis loop to do what you wish. You can look at the loop in either a Spice simulation of the circuit and in the lab by measuring it in the circuit. Both are recommended to make sure you are getting an accurate picture of reality.

Question (continued): In many literatures it has been described that, in a duty ratio controlled push-pull converter, any imbalance can result in the core saturating and lead to a device failure. It is my understanding that the core can be saturated only for a small period of imbalance only. If the core is driven by time T in one direction and T + delta T in the other direction, the saturation should only be for the delta period only. There should only be current spikes for very small duration at end of one cycle. Is my understanding is correct?

Answer: Yes. But that small duration many allow tens or hundreds of amps to flow, destroying parts or severely degrading field reliability. One mistake made is the designer only looks at this on a lab breadboard, where wiring and current-limited supplies may limit the current spike. In production, the impedance may be lower and the source may not be current limited.

Question (continued): Also, can I say that a voltage-mode-controlled push-pull converter is quite safe in a low power application, because the current spike for the duration of saturation is restricted by the winding and other resistance and could be very small?

Answer: It is what it is. If the parts are rated to handle it and it does not interfere with the circuitry (which may not like large spikes or the EMI created by them) then it may be OK. Personally I don't like spikes in my circuits or magnetics saturating unless that is the circuit function you want.

Question (continued): The device can fail only in high power applications where the current surge results in excessive power loss and consequently heating of the devices.

Answer: Not true. The parts usually fail by voltage over-stress or secondary breakdown, or by the control loop commanding them to fail due to noise induced loss of control. Any devices can fail this way, whether designed for milliwatts or Megawatts.

Question (continued): Give your view about the saturation phenomena in voltage-mode push-pull operation of a duty-controlled switch mode power supply.

Answer: Don't let it saturate or limit by design the current spike caused by saturation.

Posted by Jerrold Foutz at June 5, 2001 08:59 PM