0
\$\begingroup\$

For a resonant switching power application I need to measure the amplitude of a differential signal that is a sine wave with frequency between 1 and 2 MHz, and amplitude up to about 2 volts. There is also a common mode voltage of about 50 volts to deal with.

I've looked into high voltage difference amplifiers like LT1990, but it seems like the bandwidth is always too low. They cannot handle signals in the MHz range.

It seems like maybe some sort of current mirror circuit might be in order. Maybe someone can point me to an appropriate candidate.

Additional information:

I am trying to build a current monitoring circuit for the resonant tank on the primary side of an LLC converter. The purpose is to monitor the secondary side current by monitoring the peak-to-peak voltage across the resonant DC link capacitor. The LLC converter is driven by a microprocessor with an ADC.

Some of the comments led me to the idea that maybe all that is needed is a blocking capacitor and a half-wave rectifier. Then the peak current could be measured by the ADC, which is probably adequate enough for the purpose.

\$\endgroup\$
6
  • \$\begingroup\$ Can you not just bring the voltage down? Use a voltage divider on each line to ground. Bring down the common mode to a couple of volts. Or use a common mode choke. Or use a passive filter to remove the DC bias. There are a lot of options. \$\endgroup\$
    – user110971
    Dec 24, 2019 at 13:20
  • 3
    \$\begingroup\$ Can you capacitively-couple that differential signal to remove the DC bias, then deal with it? Break this problem down into steps. \$\endgroup\$
    – rdtsc
    Dec 24, 2019 at 13:20
  • 1
    \$\begingroup\$ To get more useful guidance you are going to need to provide further details of the system. Where does the signal cone from, and what does it mean? What is the cause and purpose of the DC offset, and how does it vary? What noise sources may be present? What fidelity of recovery is needed, on what timescale? \$\endgroup\$ Dec 24, 2019 at 13:50
  • \$\begingroup\$ It can be measured in many ways, what output is required ? analog ? digital? pk, avg? pk-pk? rms, frequency, compressed? linear , but it should be measured with an impedance <= 1kOhm. pls update spec \$\endgroup\$ Dec 24, 2019 at 15:45
  • \$\begingroup\$ The added specs are incomplete for servo control stability, aFAIK \$\endgroup\$ Dec 25, 2019 at 3:34

1 Answer 1

2
\$\begingroup\$

The proper way to measure ripple in a 1MHz SMPS is to use a 100nF ~ 1uF low ESR DC blocking cap and 50 Ohm coax with 50 Ohm termination . 0.1uF*50R =...30kHz HPF

This may reject low frequency noise and pass the ripple you are trying to measure.

The causes of Ripple are mainly the ESR of the caps, chokes and switches with some possibilities of resonance and harmonics stimulated by std step load tests like 20% to 100%, 100~20% or just steady state.

The reasons for the measurement should be shared.

There are many ways to detect the peak or pk-pk voltage with 500 MHz GBW Op Amps and diodes but with 2V ripple , you don't need any gain just a fast diode and unity gain precision peak detector.

The load < 1K is critical to prevent measurement errors from stray capacitance and inductance and for cable distance , must use 50 ohm terminated coax.

You could also compress it and use an RF power detector (cheap ) LTC5507.

You can also get within 50~75 mV of 2Vpk-pk and 100kHzBW @ 10kOhm using a Schottky diode. By reducing Cap load of 100pF to near 0, improves slightly.

enter image description here

specs: source = 50Vdc + Slider ripple 0 to 2Vp 1MHz @ 0.1 OHm ESR
Ac couple with 100nF = 2 Ohm @ 1MHz = 150kHz HPF rectify with 1N5712 model into 10k + 100 pf = 200kHz LPF

SIMulation

The other way to measure it is 0.1UF coupled to a 50 Ohm Spectrum Analyzer but with diode clamps to protect the input.

\$\endgroup\$
3
  • \$\begingroup\$ This is a very through answer. However it seems that I didn't communicate the actual issue properly. I will revise my question in order to make it clear that I am trying to build a current monitoring circuit for the resonant tank on the primary side of an LLC converter. The purpose is to monitor the secondary side current by monitoring the peak-to-peak voltage across the resonant DC link capacitor. \$\endgroup\$ Dec 24, 2019 at 20:38
  • \$\begingroup\$ It always gets easier when you have design specs with tolerances and purpose in a question. Monitor both or monitor 2nd to control primary \$\endgroup\$ Dec 24, 2019 at 20:40
  • \$\begingroup\$ The key to my answer is the impedance of AC load and breakpoint for HPF detector and LPF. I suggest using these, but you may want to change the BW or phase margin. \$\endgroup\$ Dec 25, 2019 at 3:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.