Design of electromagnetic ultrasonic excitation power supply based on DDS technology

introduction

Electromagnetic ultrasound is a non-contact ultrasonic detection method that does not require any physical contact with the object being tested, does not require coupling agents, and can be applied to the detection conditions of the object being tested under high temperature, high speed, and rough surface. Because of the non-contact characteristics, the excitation power supply used to excite the electromagnetic ultrasonic transducer is an extremely important part. The excitation power supply must generate electric pulses with high peak current and narrow pulse width characteristics. For different objects being tested, using appropriate parameters to excite electromagnetic ultrasound to maximize the electrical/acoustic conversion efficiency of the electromagnetic ultrasonic transducer is also one of the keys to improving the signal-to-noise ratio. Therefore, it is very necessary to design an excitation power supply with adjustable pulse train frequency, number, and phase. This paper designs an electromagnetic ultrasonic excitation power supply based on DDS technology.

1 Composition of electromagnetic ultrasonic excitation source

The electromagnetic ultrasonic excitation power supply mainly includes a DDS signal generating circuit, a pulse train control circuit, a power amplifier circuit, and an impedance matching circuit, as shown in Figure 1. In order to facilitate the adjustment of the frequency and phase of the excitation pulse and the control of the number of excitation pulses, the host computer communicates serially with the single-chip microcomputer to set the parameters of the excitation power supply. The single-chip microcomputer controls the DDS chip AD9850 to generate an adjustable square wave signal with a frequency of 1 kHz to 2 MHz, and the single-chip microcomputer controls the programmable logic device (CPLD) MAX7064 to complete the setting of the number and phase of the pulse train. Since the pulse signal generated by the signal generating circuit has a weak power and a low voltage amplitude, which is insufficient to drive the VMOS tube, a driving circuit is added between the pulse generating circuit and the power amplifier circuit to amplify the signal. The control circuit composed of the signal generator circuit and the driving circuit controls the opening and closing of the VMOS tube. When the VMOS tube circuit is turned off, the high-voltage power supply charges the capacitor through the charging resistor; when the VMOS tube is turned on, the capacitor, the VMOS tube and the probe (including the impedance matching circuit) form a discharge circuit, so that a high peak value narrow pulse width electric pulse can be obtained at both ends of the probe.

In order to maximize the efficiency of electrical/acoustic conversion, an impedance matching circuit is added between the power amplifier circuit and the transducer, which consists of an impedance matching transformer and a capacitor. The power amplifier circuit adopts a half-bridge power amplifier method, in which the power switch uses a MOSFET module.

2 Excitation source hardware implementation

2.1 DDS principle and circuit signal generation circuit

In order to obtain the best electrical/acoustic conversion, the excitation frequency should be consistent with the resonant frequency of the probe, so the frequency of the control signal must be flexible. The signal generator circuit is designed using a single-chip microcomputer and direct digital frequency synthesis (DDS) technology. DDS technology is a phase increment technology that uses digital control signals. It has the advantages of high frequency resolution, good stability, and the ability to flexibly generate a variety of signals. The waveform generator based on DDS changes the output frequency by changing the value of the phase increment register △phase (the number of degrees per clock cycle). Whenever the output latch of the N-bit full adder receives a clock pulse, the frequency control word latched in the phase increment register is added to the output of the N-bit full adder. After the output of the phase accumulator is latched, it is used as an addressing address of the waveform memory. The content in the waveform memory corresponding to the address is the amplitude value of a waveform synthesis point, which is then converted into an analog value output through D/A conversion. When the next clock arrives, the output of the phase accumulator is added with the frequency control word again, so that the address of the waveform memory is at the next amplitude point of the synthesized waveform. Finally, the phase accumulator retrieves enough points to form the entire waveform. The output signal frequency of DDS is calculated by formula (1):

Where: Fout is the output frequency; △phase is the frequency control word; FCLK is the reference frequency.

The frequency resolution of DDS is defined as:

Where: △Fout is the frequency resolution.

Since the frequency of the reference clock is generally fixed, the number of bits of the phase accumulator determines the frequency resolution. The more bits, the higher the frequency division. With the single-chip microcomputer STC89C516 as the control core, the parallel input method is used to realize the writing of the AD9850 control word, and the frequency of the square wave is controlled through the serial communication of the host computer. The input clock of AD9850 uses a 50 MHz active crystal oscillator, and the output frequency range can be from a few hertz to a few megahertz, but the output frequency range of the entire system is determined by some time constants in the post-stage power amplifier circuit, so the frequency range is 1 kHz to 2 MHz adjustable. Connect the P1 port of the single-chip microcomputer to the parallel input port of AD9850, and P3.6 and P3.7 complete the input/output control of the single-chip microcomputer to AD9850. After the AD9850 control word is written, the sine wave signal of the corresponding frequency is output by IOUT. In order to prevent the output frequency from being interfered by the high-frequency ramp wave, a two-stage type LC low-pass filter is selected, whose dynamic range bandwidth is 0-40 MHz, to send the pure sine wave to the comparator port of AD9850, and finally output the square wave from QOUT. The DDS signal generation circuit diagram is shown in Figure 2.

2.2 Pulse train control circuit

In order to adjust the resonance point of electromagnetic ultrasound, the number of control signals is required to be flexible. Since the electromagnetic ultrasonic transducer (EMAT) uses an electromagnet, the phase of the excitation source should be consistent with the 50 Hz power frequency phase of the electromagnet and can be adjusted between 0 and 180 degrees. A single-chip microcomputer is used to control the programmable logic device (CPLD), and the number and phase of the pulse train are controlled inside the CPLD. Finally, the host computer communicates with the single-chip microcomputer to generate a pulse train with adjustable frequency, number, and phase. The P0 and P2 ports of the single-chip microcomputer are connected to the CPLD as address and data interfaces, and P3.4 and P3.5 are used as control ports. When the single-chip microcomputer writes the number and phase of the pulse train into the CPLD, it outputs two complementary unipolar square wave signals, HO and LO.

2.3 Design of power amplifier circuit and impedance matching circuit

In order to increase the intensity of electromagnetic ultrasound, the power of the excitation signal needs to be further amplified. Since the intensity of electromagnetic ultrasound is proportional to the square of the current, the power amplifier circuit can be used to amplify the signal current.

The power amplifier circuit uses a high-power tube (MOSFET) to form a half-bridge power amplifier circuit. MOSFET has the characteristics of fast switching speed, high voltage resistance, good high-frequency characteristics, high input impedance, low driving power, and no secondary breakdown problem. The gate drive requires that the trigger pulse has a sufficiently fast rise and fall speed. To make the power MOSFET fully turned on, the voltage of the trigger pulse must be higher than the turn-on voltage of the power MOSFET. There are many types of MOSFET tubes, such as STW15NB50, IRF840, etc. In this design, STW15NB50 is selected, with a minimum turn-on time of 24 ns, a turn-off time of 15 ns, a drain-source voltage VDS of 500 V, and a peak pulse current of 58 A, which can meet the design requirements.

Figure 3 is a half-bridge power amplifier circuit, R1, R2 are bridge balance resistors; C1, C2 are bridge arm capacitors; D1, D2 are bridge switch absorption circuit elements. Its working principle is as follows: two anti-phase square wave excitation signals are connected to the bases of the two switch tubes respectively. When HO is high and LO is low, Q1 is turned on and Q2 is turned off. The current passes through Q1 to the primary of the transformer to charge capacitor C2. At the same time, the charge on C1 is discharged to Q1 and the primary of the transformer, thereby inducing a positive half-cycle pulse voltage at the secondary of the output transformer; when HO is low and LO is high, Q2 is triggered to turn on and Q1 is turned off. The current is charged through capacitor C1 and the primary of the transformer, and the charge of C2 is also discharged through the primary of the transformer, inducing a negative half-cycle pulse voltage at the secondary of the transformer, thereby forming a power amplifier waveform with a working frequency cycle. Since the power amplifier tube works in the saturation region or cut-off region of the volt-ampere characteristic curve, the collector power consumption is reduced to a minimum, thereby improving the energy conversion efficiency of the amplifier, which can reach more than 80%.

MAX4428, IRF series driver chips or amplifier circuits composed of triodes can be used to drive MOSFET tubes. However, the driving frequency of MAX4428 and some other integrated driver chips can generally only reach about 200 kHz, while this design uses triodes connected as shown in Figure 4, the driving circuit frequency can reach about 2 MHz, the output is free of noise and the cost is low, and it can successfully drive the MOS tube on/off.

In order to maximize the instantaneous power output, the impedance of the probe needs to be matched. Compensation impedance is added to the output end of the power amplifier to offset the inductive reactance and capacitive reactance of the entire circuit, maximize the transmitted power, and maximize the efficiency of converting electrical energy into acoustic energy. The matching circuit is shown in the dotted box in Figure 3. The half-bridge inverter output is coupled to the transducer through the capacitor after being coupled by the transmission line transformer. The transmission line transformer consists of a twisted pair and a magnetic ring. When the pulse train transmission frequency in the circuit is 1 MHz, the output impedance of the excitation source is 50 Ω; since the workpiece to be measured is also part of the transducer, the probe should be placed on the surface of the workpiece when measuring the probe impedance. If the measured load impedance is 500 Ω, the number of twisted pair turns should be about 10.

After tuning and matching, the transducer reaches resonance when driven by the electromagnetic ultrasonic power source. Figure 5 shows the collected excitation voltage waveform of the transducer. It can be seen that a pure sine wave with a frequency of 100 V is obtained, and its peak-to-peak value is close to 100 V when the external voltage is 100 V.

3. Stimulus Source Software Design

The software design mainly involves programming the microcontroller to achieve communication with the host computer, control the output of the CPLD, adjust the output frequency of the AD9850, etc. The program flow is shown in Figure 6.

4 Conclusion

The electromagnetic ultrasonic excitation power supply adopting DDS technology and single-chip control technology has a simple hardware structure and is easy to program and control. Compared with the traditional analog signal generator, it has high frequency accuracy and precisely controllable phase, which improves the flaw detection effect, facilitates the digital control and operation of the whole set of equipment, and reduces the size and weight of the equipment.