Difference Amplifiers Enable Low-Loss, High-Performance Full-Wave Rectifier

The following is a guest blog from Analog Devices. The authors are Chau Tran and Fotjana Bida. Their biographies are at the end of this article.

A full-wave bridge rectifier converts an AC signal to a full-wave DC signal. Typically, a bridge formed by four diodes achieves full-wave rectification. Figure 1 shows four diodes arranged in series pairs, with two diodes conducting current during each half cycle. At any given time, two diodes are forward biased, while the other two are reverse biased, effectively eliminating them from the circuit.

The result is a DC output, where the current flowing through the load is the same during both half cycles. A smoothing capacitor can be added to the output if the rectifier is to be used as a DC power supply. The main advantage of this bridge circuit is that it does not require a special center-tapped transformer, thereby reducing its size and cost.

This classic circuit has many disadvantages, however. The current flowing through the load is unidirectional, so the DC voltage developed across the load should have an average value of

Equation 1

In reality, however, during each half cycle, the current flows through two diodes, so the output voltage amplitude is two diode drops less than the input amplitude.

For example, with a 5-V peak input, the peak output will be about 3.8 V. The ripple frequency will be twice the supply frequency; for example, with a 60-Hz supply, the ripple frequency will be 120 Hz. In addition, the circuit suffers from crossover distortion and temperature drift.

Figure 1

Classic bridge rectifier.

Classic bridge rectifier.

The circuit shown in Figure 2 improves the performance of the classic four-diode bridge by employing two low-cost, high-performance difference amplifiers and two low-cost diodes to eliminate the loss at the output. This approach achieves better precision, lower cost, and lower power consumption than conventional techniques.

In this circuit, VIN is a sine wave. During a positive half cycle, diode D1 conducts. Both amplifiers A1 and A2 act as inverters. The result is a positive voltage at VOUT with amplitude exactly the same as that of the input. During a negative half cycle, diode D2 conducts. Amplifier A1 now has a gain of –2/3, while A2 has a gain of +3/2. The net gain of –1 provides a positive voltage at VOUT with amplitude opposite that of the input. The combination gives a loss-free full-wave rectifier. Signals as large as ±10 V can be handled at frequencies as high as 10 kHz.

Figure 2

A simple full-wave rectifier.

A simple full-wave rectifier.

This design has several inherent performance advantages, including cost, crossover distortion, gain error, and noise. The gain accuracy of the rectified output is determined by the 10 kΩ resistors. Precisely matched, these laser-wafer-trimmed resistors guarantee gain error of less than 0.02%. The circuit's noise gain is only 2, resulting in lower noise, offset, and drift.

Figure 3

Performance of the simple full-wave rectifier.

Performance of the simple full-wave rectifier.

The circuit's performance is demonstrated in Figure 3 with a 20-V p-p input signal at 1 kHz. The fact that the output overlaps the positive cycle of the input means that the exceptional input and output characteristics of the difference amplifiers create a negligible loss.

Unlike the classic circuit, the characteristics of the two diodes in the new circuit have no effect on the output voltage. Therefore, the performance over temperature is better.

A precision full-wave rectifier built with two difference amplifiers and two diodes offers several advantages over traditional designs. Specially, the output voltage shows no loss as compared to the input voltage. The difference amplifier solution has no crossover recovery problem and is optimized for low drift over a wide temperature range.

Chau Tran works in the Linear Products Group at Analog Devices in Wilmington, Mass. He joined the company in 1984. In 1990, he graduated with an MSEE degree from Tufts University. He holds more than 10 patents and has authored more than 10 technical articles.

Fotjana Bida is a product development engineer for the Linear Products Group at Analog Devices in Wilmington, Mass. She joined the company in 2007 and has worked in design as well as in product development of precision signal-processing components. She holds a MSEE from Northeastern University.

14 comments on “Difference Amplifiers Enable Low-Loss, High-Performance Full-Wave Rectifier

  1. RedDerek
    March 7, 2014

    I suspect there is some phasing issue, but at the low frequencies talked about, it would be negligable. I remember seeing this type of circuit some time ago, but cannot remember. Good to see it again and nice simple explaination.

  2. Steve Taranovich
    March 7, 2014

    I'm glad you like it @RedDerek—-there are lot's more to come like this tutorial!

  3. Davidled
    March 7, 2014

    When capacitors and inductors affect the AC circuit, phase could be generated in the voltage & current. When I look the circuit, there is no inductor. We do not expect any phase in the circuit.

  4. RedDerek
    March 7, 2014

    @DaeJ – I agree in general. However, there are inherient delays in opamps as well. Thus, my comment about some phasing issues. It is not much though.

    I have one circuit that takes the input sine wave as differential and divides down to a reasonable working voltage for the circuit. There is filtering that adds in much delay. Then I added an all-pass filter to get the waveform back into phase with the original.

  5. kendallcp
    March 8, 2014

    … but it's not really fair to compare it to a regular diode bridge and then ding the bridge for its disadvantages.  Add some opamps and the bridge becomes a much more workable solution.  Putting it in the feedback path of an op-amp eliminates the diode voltages (and their mismatches) and you can 'slap a cap' straight across it to give you the envelope.  That in turn can be directly sensed differentially or turned back into single ended form with a diff amp (manufacturers of diff amps take note!).  Did this in the 80s and it worked fine.  Helps to keep a lid on amplifier slewing as well.  That is always something to keep an eye on with rectifier circuits, which tend to underperform for low level high frequency signals since there's often an unmonitored opamp output thrashing about madly by a diode drop or two just to keep up.

    There are better fits for single supply operation too.  It's pretty straightforward to produce an accurate full-wave rectifier circuit that needs no negative rail but still works on a bipolar signal.  The circuit shown here needs split supplies to cope with an AC signal.

  6. Netcrawl
    March 9, 2014

    great article @Steve! thanks for that,  you clearly explained everything, rectification systems are primarily designed for conversting AC input signals into DC voltage signals, they're probably the most commonly used stuff in power electronics and transmission. 

  7. Davidled
    March 9, 2014

    A simple full-wave rectifier has D1 and D2 which is partially for rectifier circuit. Diode D1 and D2 acts like voltage clamping.  Output performance without D1 and D2 could be different while changing AC voltage wave.  My concern is that there would be a noise or output signal distortion when either class full rectifier circuit or a simple full rectifier circuit is integrated with other circuit, digital logic, or high frequency input signal.

  8. samicksha
    March 9, 2014

    I guess most modern phasers are a part of a digital signal processor, often trying to emulate analog phasers.

  9. chirshadblog
    March 10, 2014

    @samicksha: Yes but is there a  big huge difference in quality ? 

  10. samicksha
    March 11, 2014

    @chirshadblog: What exactly you refer as quality here, is it performance related issue or quality of hardware.

  11. chirshadblog
    March 11, 2014

    @samicksha: Actually both since quality has links to both the areas. So both parties do play a major role.  

  12. etnapowers
    March 11, 2014

    The figure 1 refers to an ideal no losses bridge, for diodes having a conduction threshold (˜1V) that is comparable to the input AC voltage amplitude, the output waveform is a littel bit different, so for this rectifier some low losses diodes are needed, expecially when the input AC voltage amplitude is low.

  13. etnapowers
    March 11, 2014

    The diodes D1 and D2 of the circuit in Figure 2 are turned on/off alternatively , so I think that a diodes matching really accurate is very important to avoid a cross conduction due to a lack of matching between the turn-on and turn-off times of the diodes.

  14. etnapowers
    March 11, 2014

    Looking at the circuit in figure 2, the full wave rectifier I see that the positive terminal of  A1 is polarized at GND DC , but the positive terminal of A2 is at a non zero voltage DC value. I wonder if this difference in the polarization of A1 and A2 influences the gain loop.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.