Posted on 01 September 2019

A Discussion of the Active Clamp Topology

Free Bodo's Power Magazines!




Primary switches are switching in a lossless manner

The Active Clamp is a topology being favored for a significant number of power converters. There is a misconception, however, that the main FET turns on in a zero voltage condition. The transition of the clamp FET to the main FET is the subject of this article.

By John Bottrill, Senior Applications Engineer, Texas Instruments


To generate the waveforms presented in this paper, a schematic of the UCC2891 Engineering Verification Model (EVM) was used (see Figure 1). The EVM was tested with 50V on the input, and two different loads on the output. The first was at a load of 0.5 amps. The second was at a load of 6.0 amps. Current loops were put in series with the primary and secondary windings so that the currents through the transformer could be measured. Understanding the effect requires a detailed look at the idealized transformer. This will be used to explain the effects seen in the waveforms.

Schematic of the UCC2891EVM

The magnetizing inductance represents the magnetic flux in the core when a current is present. Increasing the current in a winding by applying a voltage to it increases the magnetic flux. Allowing current out of a winding by magnetically inducing voltage, results in a decrease of the magnetic flux. In short, anything that stores energy increases the magnetic flux, and anything that reduces the stored energy reduces the stored magnetic flux.

Figure 2 shows an idealized transformer block where the magnetizing and leakage inductance are separate from the actual transformer.

Idealized Transformer

When power is being delivered to the load, current flows through both the magnetizing inductance and the primary of the transformer. The current in the primary is transformed and conducted out the secondary to the output inductor.

When the unit is in the clamp mode, the magnetizing current first flows into the clamping capacitor then back out of the clamping capacitor. However, there is no current in the primary as the winding polarity is such that the current flow is blocked by the secondary side switching element (Q3 and Q4). In addition to the idealized transformer, a detailed model of the switches also will be needed to help explain the effect (see Figure 3).

Detailed Model of the Switch

This FET model has an inherent body diode. So if a reverse voltage is applied across the FET, it appears as a diode. In addition to the diode, the inherent capacitances are shown. The presence of these capacitances are needed to explain some of the effects being observed.

Figure 4 shows the switching waveform of the converter at a load current of 0.5 amps and over slightly more than a full cycle.

Figure 5 shows the same picture with an output current of about 6.0 amps.

The difference between the waveforms of Figures 4 and 5 is the output current, which only occurs once Q2 turns on (see Figure 1). By that time, the effect we are interested in has already happened.

Waveforms at 0.5 amps out

The effect we are looking at occurs between the time Q1 turns off and Q2 turns on. To look at this more closely, we need to expand this section of Figures 4 and 5. Let’s look closely at this section in Figure 6, which is an expansion of the waveforms under the conditions of Figure 5.

Waveforms at 6.0 amps out

Figure 6 shows that when Q1 turns off, the voltage across the primary winding starts to decrease. Once the secondary side leakage inductance is overcome, the secondary side voltage starts to decrease, and the current starts flowing in the secondary side winding. In the transformer model, this current has to come through the ideal transformer. Therefore, it robs the magnetizing current from the magnetizing winding. This in turn causes the total current in the primary winding to effectively decrease by the amount of current that is circulating through the magnetizing inductance and the primary of the ideal transformer. This decrease in current in the primary is also a decrease in the leakage current in the primary.

Expansion of Figure 5

Figure 7 demonstrates this by showing the same time frame, but focusing on the primary current – rather than the secondary current.

Same image as Figure 6 except primary current instead of secondary

As can be seen in a comparison of the two pictures, when the current on the output begins to increase, the current on the primary likewise decreases. The primary decreases by about 100mA. The current on the secondary increases to about 600mA. The transformer is a six-to-one turns ratio, so that confirms the direct correlation.

If the time between Q1 turning off and Q2 turning on were sufficiently long, the voltage across the primary would collapse completely. The magnetizing current would then flow through the internal diode of the FET and through the output choke to the output capacitors. Because the magnetizing current is much smaller than the load current, it is not possible for it to result in a voltage across the transformer. If the transformer secondary voltage were to collapse, any magnetizing current would simply supplement the current coming through the output inductor that is being drawn through Q5, Q7 and Q8. Because current would be coming through these transistors body diodes, the transformer secondary winding would be clamped to 0V.

Where is the secondary current coming from the transformer going?

There is still a positive voltage across the secondary of the transformer with pins seven and ten positive, so it is not going through the body diodes of Q3 and Q4. It is going to the parasitic capacitive elements of the secondary switches more than anywhere else. If we look at the current waveform, the time it flows and the change in voltage, we can calculate the effective secondary capacitance at about 12nF.

Q3 and 4 are HAT2165H and have a Coss of 1200pF at 10 V. This goes down as the drain to source voltage decreases. The Crss for the same 10V is about 400pF. At about 3 V, Coss is over 2000pF and rapidly rising. Crss is over 500pF and also rapidly rising, according to the data sheet. This, however, does not account for the 12nF. This same transformer winding, however, is connected to the gate of Q5, Q7 and Q8 and they have a Ciss capacitance of 5nF each. Each would be 15nF. All in all, the 12nF calculated is well within the margin of error for the amount of capacitance from the data sheet, and we have determined that the current is going into the capacitance of the FETs.

What will the current on the primary switch look like when Q2 turns on and why?

Looking at the waveform in Figure 7, the current in the primary side of the transformer is negative to the direction of current flow once Q2 is turned on. The change of current in that winding is not instantaneous (thanks to the leakage inductance), so the FET is fully on and low impedance before that current materially changes direction. In fact, from the waveforms for the first 5ns or so, the current through the primary is negative to the direction that it will eventually be as Q2 remains on.

However, there is the drain to source capacitance of Q2, and drain to source capacitance of Q1 that will immediately discharge through Q2. These are fairly small values so though the switching will not be 0V switching. It will be close to zero current switching with only parasitic capacitive elements providing current through the FET switch, hence, the losses.

There is a large primary current spike that is coincident with the outputs transitioning and the Q5, Q7 and Q8 conducting momentarily in the forward direction, either from the gate not being discharged fast enough or from body diode conduction. In conclusion, both primary switches are switching in a lossless manner. Q1 is switching in a 0V condition, and Q2 is switching in a zero current state – if you ignore the drain to source capacitance discharging currents.



VN:F [1.9.17_1161]
Rating: 0.0/6 (0 votes cast)

This post was written by:

- who has written 791 posts on PowerGuru - Power Electronics Information Portal.

Contact the author

2 Responses

  1. avatar Hunter says:

    I am learing the acitve clamp topology, and I found your document is very clear, so would you like to send it to me thank you.

    VA:F [1.9.17_1161]
    Rating: 0.0/5 (0 votes cast)
    • avatar PowerDahl says:

      Hello Hunter. We are glad that you found the information usefull.

      If you would like to print or download the article in PDF format, click on the "Print PDF" button in the green box above the comments. Also, you can download the full magazine where this article first appeared on the Bodo's Power Systems website (there is logo with link in our partner companies banner at the very top of this webpage - issue Sep 2007).

      VN:F [1.9.17_1161]
      Rating: 0.0/5 (0 votes cast)

Leave a Response

You must be logged in to post a comment.