Electronic Design

  
Reprints     Printer-Friendly    Email this Article    RSS        Font Size     What's This?


[Ideas For Design]
Extend RMS-To-DC Converter's Range

Contributing Author  |   ED Online ID #4701  |   September 5, 2000


Update by Barrie Gilbert and Dana Whitlow
In the original 1990 Idea For Design, gain switching was used to extend the dynamic range of ac signal measurement by providing autoranging for an rms-to-dc converter. When the operating range of the AD736 was exceeded, a comparator with hysteresis was used to switch in an attenuator at the input of the rms-to-dc converter and apply a compensating gain to its output.

With high-frequency rms detectors, extending the dynamic range is even more critical. This is because the currently available dynamic range is only about 30 dB at best for such detectors. Generally, this is insufficient for handling the combination of high crest factor and wide measurement range of complex high-frequency waveforms, such as the digital communications signals commonly employed today.

The simple circuit employed in 1990 can't be implemented now because the comparator delays and switching transients are too slow to handle broadband signals. But, the availability of accurate, high-frequency, variable-gain amplifiers (VGAs) and high-speed rms detectors allows alternative approaches. These enable users to avoid making step changes in gain. Through this updated circuit, a novel means for wide-range, true-rms measurement of complicated high frequency waveforms is presented (see the figure).

In this scheme, a VGA inserted into the signal path ahead of the rms detector becomes part of an AGC loop. It maintains the output of the rms detector at a constant value when the difference relative to a set-point is integrated to form a control voltage. That voltage is applied to the VGA, whose gain is an accurate exponential function of the control voltage, resulting in the control voltage being logarithmically related to the input. A prototype circuit conformed to the desired logarithmic response law within 1 dB over an 80-dB range of input-signal amplitude.

The high-frequency rms detector, AD8361, is burdened only with the crest factor of the signal waveform, because AGC action maintains its rms input at a fixed level. The loop's error integrator is differential, with one input driven by the output of the rms detector, and the other driven by a reference voltage that sets the detector output.

Both halves of a dual VGA, called AD605, are used in cascade. This provides a measurement range of 80 dB, even for a signal with a high crest factor, such as random noise. Each VGA is set for a slope of 20 dB/V by 2.5 V applied to its VREF pin, resulting in a maximum-control range of 40 dB per stage as the control voltage varies from 0.5 to 2.5 V. The maximum-gain range of IC1a is −14 to 34 dB with a jumper between its OUT and FB pins. The gain range of IC1b is 0 to 48 dB, by the absence of this jumper. Only the central 40-dB portion of each VGA's full-gain range is used.

Noise makes a growing contribution to the measurement as the VGA's gain increases in response to smaller input-signal levels. A three-pole low-pass filter is included in the signal path to prevent the rms detector from responding to broadband noise outside the desired passband. In this example, that's a few kHz to about 2 MHz.

The filter is split into two sections. The first section, IC2a, is located between the VGAs. It provides an active complex pole pair. Following IC1b, the second section, containing RFP and CFP, furnishes the third pole. For IF applications, a bandpass filter would be more appropriate. A passive filter would be suitable too. The cascade of two VGAs performs well up to about 10 MHz. As it has a higher gain, IC1b primarily limits the bandwidth. The AD8361 is essentially flat until around 1 GHz and has useful response to about 2.5 GHz. Therefore, measurements at higher frequencies will become possible as faster VGAs become available. These, however, will also need to provide a "linear-in-dB" gain function.

Though not obvious from examining the schematic, it should be noted that an attenuation of about 6 dB is inherent in the passive portion of the low-pass filter. This is due to the rms detector's 225-Ω input impedance working against RFP. In addition to the differing gain ranges chosen for the two VGAs, this ensures that the signal swings throughout the signal path remain within the linear limits of the active devices. When making changes, one must pay close attention to the distribution of gains and losses through the signal path.

The averaging time constant of the detector over which the "mean" is measured is set by CFIL, which is on pin 6 of IC3. The constant should be long enough to suppress ripple resulting from modulation on the signal. But, this filter adds a pole to the AGC loop which must be accounted for when choosing the integration time constant provided by IC2b and surrounding components. This is in order to ensure loop stability. Additional low-pass filtering of the output is supplied by RFIL2 and CFIL2 at the circuit's output. With the values shown, the overall measurement time constant is about 3 ms, and there's about 1 dB p-p of equivalent ripple on the circuit's output when the input signal is wideband random noise (flat from near dc to 2 MHz).

Located on pin 7 of IC3, RPD sinks current from the error integrator's summing junction when the output of the VGA cascade is below loop equilibrium. This drives the integrator's output more positive to increase gain. When the VGA's output is too high, the AD8361 output sources current to the summing junction. That drives the integrator's output lower, which results in reduced gain.

This circuit was designed for 5-V single-supply polarity operation. That guided the choice of the particular VGA (AD605) and op amp (AD8032) used for the circuit. The 2.5-V reference input to the AD605 needs to be stable and predictable because the output slope expressed in voltage per dB is proportional to that reference.


<-- prev. page     1 [2]     next page -->

Reprints   Printer-Friendly  Email this Article  RSS    Font Size   What's This?


  • Network-On-Chip Tools Arrive for The Masses
  • Tackling System Design Challenges Through Early Verification
  • ESL Tools Take Center Stage As Designers Move Up
  • Parasitic Extraction Tool Targets Next-Generation Custom ICs
  • Synopsys Jumps Into ESL-Synthesis Pool
  • Verify Control Systems Before Committing To Hardware
  • You're Using How Many FPGAs?
  • Tool Up For The FPGA Blitz
    1) Build A Smart Battery Charger Using A Single-Transistor Circuit
    (187 views today)
    2) Hot Hands For Some Cool Rock: Motion Sensing Meets Audio Engineering
    (166 views today)
    3) Science Fiction Meets Science Fact In Today's Robot Research
    (96 views today)
    4) GPS-Derived Grandmaster Clock Delivers Ultra-Precise Time And Frequency Sync
    (89 views today)
    5) What's All This Transimpedance Amplifier Stuff, Anyhow? (Part 1)
    (86 views today)
    ALL TOP 20



    POST YOUR COMMENTS HERE
    Name:

    Email:
    Your Comments:

    Enter the text from the image below


    Please refresh the page if you have trouble reading this text.

    Search Electronic Design
         
      
     
    Email Newsletter
    Sponsored By:
    Electronic Design UPDATE provides readers with late-breaking news, opinions from industry experts, and timely technology stories. It's a unique opportunity to get your product message in front of engineers, engineering managers, and corporate managers while they're reading about critical information online.

    Enter Email to Subscribe
      

    Electronic Design Europe Electronic Design China EEPN Power Electronics Auto Electronics Microwaves & RF
    Mobile Dev & Design Schematics Find Power Products Military Electronics EE Events Related Resources