0
\$\begingroup\$

I was trying to use the LTC6269-10 in non inverting configuration in order to filter the high frequencies around 400 MHz, but when i plot the bode diagram in LTSpice it shows a resonance around that frequency. Also, if i try to change the values of the resistors in the circuit the resonance tend to shift and increase in value. What could it be?

These are the two circuits that i have tested: Circuits

And this is the bode diagram:

Bode diagram

(The green one is the circuit on the left, the red one is the circuit on the right)

\$\endgroup\$
1
  • 1
    \$\begingroup\$ You may also want to model resistors with parasitics for this problem \$\endgroup\$
    – Voltage Spike
    Jun 28 '17 at 23:00
2
\$\begingroup\$

Aha (!) the infamous gain peaking problem seen in the non-inverting op-amp circuit!

This is one of those things you learn to watch out for. Remember that the gain of this type of amplifier configuration is: -

\$1+\dfrac{R_5}{R_6}\$ (for the right hand diagram)

But that formula hides the fact that the reactance of leakage capacitors needs to be taken into account. The op-amp has a common mode input capacitance of 0.45 pF and this will progressively shunt R6 as the operating frequency rises.

In other words, gain will start to rise with frequency and the gain will have increased by 3 dB when Xc = R6. This happens when: -

Frequency = \$\dfrac{1}{2\pi R_6C_{IN}}\$

So, input capcitance is 0.45 pF and R6 is 3000 ohms hence F = 117 MHz. At this frequency the circuit gain is rising and keeps rising at 6 dB per octave until it becomes limited by the open-loop gain of the op-amp. The data sheet isn't very good at pictorially showing this so I have taken a liberty with what they do show and extended it: -

enter image description here

The original extract from the DS is on the left and my mangled version is on the right. Hopefully this should explain why you see a peak at about 250 MHz. The bright red line is what I predict your gain profile will be until it hits the darker red line (what I expect the open-loop gain to be if the graph is extended).

When you chose lower value resistors (your left diagram), that peak got pushed out to 400 MHz and this generally ties in with what you should expect. It's never an exact science of course but this generally is what you are seeing.

Generally the best fix is to lower your resistors so that the peak is pushed out way beyond the open-loop gain of the op-amp. I don't see any problem in making R6 = 300 ohms (ten times lower) and making R5 a 2k7 (ten times lower too).

Remember that when you build this there will be stray capacitance that can make the problem worse so keeping the resistor values low is going to be sensible. Note that the DS also states this: -

Since LTC6268-10 is a decompensated op amp with gain-of-10 stable, it requires that CIN/CF ≥ 10.

This means that adding a capacitor across the feedback resistor has to be done with care.

\$\endgroup\$
2
\$\begingroup\$

Put a taste of capacitance across the feedback resistor to compensate for the input capacitance. Maybe 0.1~0.5pF for R1=9K. And read the datasheet carefully.

Lower value feedback resistors may be useful.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ This opamp has 0.5pF common-mode input capacitance. With 27k feedback resistor, the -3 dB cutoff is at about 11 MHz, and the phase shift towards 70-80 deg comes at 20X of that, that's why the ~200 MHz resonance. The results will be more surprising if the PCB will have no voids under In+ and In- pins ... This is probably not the best opamp for 400MHz operations. \$\endgroup\$ Jun 28 '17 at 23:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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