2
\$\begingroup\$

I am trying to sample the voltage drop across different components at high frequencies, and figured the best way to do it was with an instrumentation amp so there's not much additional load placed on the main circuit, and then feed that output into a peak detector to get a dc value. I found an amp that looked like it would work in LTspice, but the output of the amplifier (red) is not matching the voltage drop across resistor Rm (green). Changing the supply voltage, gain, and values of Rm/R4/C1 do not change the output of the amp.

Is the idea itself wrong or am I using the incorrect amp for this application?

RF circuit with instrumentation amp

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Your amp has a much lower bandwidth than your input signal. Try reducing the sine wave to 10KHz. You also might need to add a negative supply rail to your amplifier. \$\endgroup\$
    – Ste Kulov
    Commented Jun 12 at 21:46
  • \$\begingroup\$ There is no gain resistor between G1 and G2, \$\endgroup\$
    – RussellH
    Commented Jun 13 at 0:25
  • \$\begingroup\$ Some more info: I want to work in the MHz range, so lowering the frequency is not an option. Also for this particular amp, leaving G1 and G2 open is recommended on the datasheet to have unity gain. \$\endgroup\$
    – Peator
    Commented Jun 13 at 1:23
  • 1
    \$\begingroup\$ @Peator Then you need a different amp. I know you need higher frequencies. I only suggested the above exercise so you can see that the amp works when the frequency is much lower. \$\endgroup\$
    – Ste Kulov
    Commented Jun 13 at 2:17

2 Answers 2

Reset to default
4
\$\begingroup\$

You need to get the inputs within the common-mode voltage range of the amplifier. That means bipolar supplies. Input common mode range depends on supply and output voltage but it's never less than the negative supply rail and you're hitting it with -1.4V on the negative peaks. See the datasheet

enter image description here

Your op-amp has a gain-bandwidth of (typically) only 60kHz, which means that your error will be 30%-ish at 1/15 of your operating frequency and there won't be much left at 1MHz.

You've also made kind of a high-pass filter with the resistor and capacitor. The 10p has a reactance of about 16kΩ at 1MHz so you're going to get almost no differential input voltage at any frequency 1MHz or lower. At best (as frequency tends to infinity) the differential input voltage would be attenuated by 17%, so even if the inamp had infinite frequency response and bipolar supplies it will never give a gain close to 1 at any frequency.

\$\endgroup\$
1
\$\begingroup\$

As Spehro suggested you are violating the common mode range of the amplifier.

This can be avoided by providing a DC offset voltage.

enter image description here

The BW of the amplifier is 60Khz for a gain of 1 .Your input frequency is 10Mhz.

enter image description here

If you want to work with Mhz frequencies then you need to change the Amplifier

\$\endgroup\$

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