9.2.2.1.1 Excitation Voltage

For this example, the resolver impedance is specified between 2Vrms and 7Vrms up 20kHz maximum frequency. Since, the resolver attenuation is ~0.5V/V and most data acquisition microelectronics run off of 5V supply voltages. An excitation input of 6Vpp (or 2.12Vrms) will be chosen to give the output readout circuitry enough headroom to measure the secondary side outputs (~3Vpp).

The excitation coil can be driven by a single-ended opamp output with the other side of the coil grounded or differentially as shown in Figure 28. Differential drive offers higher peak to peak voltage (double) on to the excitation coil, while not using up as much output voltage headroom from the opamp. Leading to lower distortion on the output signal.

Another consideration for excitation is opamp power dissipation. As described in Power Dissipation and Thermal Reliability, power dissipation from the opamp can be lowered by driving the output peak voltages close to the supply and ground voltages. With ALM2402Q1's very low V_OH/V_OL, this can be easily accomplished. Please see Figure 1 for V_OH/V_OL values with respect to output current and Output Stage for further description of the rail-rail output stage.

9.2.2.1.2 Excitation Frequency

The excitation frequency is chosen based on the desired secondary side output signal resolution. As shown in Figure 34, the excitation signal is similar to a sampling pulse in ADCs, with the real information being in the envelope created by the rotor. With a GBW of 600kHz, ALM2402Q1 has more than enough open loop gain at 10kHz to create negligible closed loop gain error.

Along with GBW, ALM2402Q1 has optimal THD and SR performance (see Typical Characteristics) to achieve 6Vpp (or 3Vpp from each opamp). The signal integrity can also be observed in the Typical Characteristics section.

9.2.2.1.3 Excitation Impedance

Knowledge of the primary side impedance is very important when chosing an opamp for this application. As shown below in Figure 29, the excitation coil looks like an inductance in series with a resistance. Many time these values aren't given and must by calculated from the cartesian or polar form, as it is given as a function of frequency or phase angle. This calculation is a trivial task.

Once the coil resistance is determined, the maximum or peak-peak current needed from ALM2402Q1 can be determined by below:

Equation 1.

In this example the peak-peak output current equates to ~120mA. Each opamp will handle the peak current, with one sinking max and the other sourcing. Knowledge of the opamp current is very important when determining ALM2402Q1's power dissipation. Which is discussed in the Power Dissipation and Thermal Reliability section.

Figure 29. Excitation Coil Implementation

The primary side of a resolver is inductive, but that typically is not all the opamps driving the coils see. As shown in , many times designers will add a resistor in series with a capacitor to illiminate crossover distortion. Which happens due to the biasing of BJTs in the discrete implementation. With ALM2402Q1's rail-rail output, this is rarely needed. This can be seen in the waveforms shown in the section.

It is also common practice to add EMC capacitors to the opamp outputs to help shield other devices on the PCB from the radiation created by the motor and resolver. When choosing C_EMC, it ismportant to take the opamp's stability in to account. With ALM2402Q1 having bery good phase margin at and above 200nF, no stability issues will be present for many typical C_EMC values.

9.2.2.3.1 Improving Package Thermal Performance

θ_JA value depends on the PC board layout. An external heat sink and/or a cooling mechanism, like a cold air fan, can help reduce θ_JA and thus improve device thermal capabilities. Refer to TI’s design support web page at www.ti.com/thermal for a general guidance on improving device thermal performance.