0
\$\begingroup\$

I have tried many different op-amps in many configurations. But almost always I am getting a very unstable configurations:

https://youtube.com/watch?v=ybhDs-dw6j8

https://youtube.com/watch?v=O2C596rmJqg

The original signal is very stable.

Parameters I try to archive:

  1. Bandwidth - 50 MHz
  2. Slew Rate - 200-300 V/µs
  3. Single supply if possible

Is it actually achievable? I know that the stability is much better in the dual supply configurations.

Some pictures: enter image description here The actual test circuit. .1u & 6.8u caps on +Vc. I have tried many different resistors (from zero to 4.7k) on the feedback circuit and I am not getting any better results: enter image description here

With small resistors it is a bit better on changing signals but it is much worse with the level ones. Larger are better for level ones but worse for faster changing ones.

I took this from the IC datasheet: enter image description here

After many experiments this is the best waveform I have archived enter image description here

\$\endgroup\$
  • \$\begingroup\$ A 2 volt pk-pk sine at 50 MHz requires a slew rate more than 300 V/usec. \$\endgroup\$ – WhatRoughBeast May 28 '17 at 21:52
  • 2
    \$\begingroup\$ Sure it's actually achievable. Show us a schematic. And pictures would be better than youtube links. \$\endgroup\$ – uint128_t May 28 '17 at 21:56
  • \$\begingroup\$ for a 20 ns cycle triangle wave , a 2Vpp sine is fundamental to a 2.6Vpp triangle wave thus dV/dt= 2.6V /10ns = 260 V/us. Depending on your impedance specs for Vout, Zin, Zout, Zload, and Insertion loss, or gain, anything is possible ONLY AFTER well defined specs. (hint) but don't think about Op Amps \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 May 28 '17 at 22:24
  • \$\begingroup\$ unless it's an LMH6601MG or better.. but don't show schematic. ALWAYS show better specs FIRST. Your Youtube video shows complete disregard of EMI rejection on a 25kHz square wave and proper probing methods with huge SMPS noise pulses. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 May 28 '17 at 22:31
  • \$\begingroup\$ And could you please provide us with a picture of the physical circuit which produced those scope traces? I suspect it will be enlightening. \$\endgroup\$ – WhatRoughBeast May 28 '17 at 23:13
1
\$\begingroup\$

Use a Ground Plane. Bypass the OpAmp +VDD (and -VDD) to that plane within 1mm of the IC. Use Surface Mount Technology SMT Cs and Rs. Use no vias in the bypassing; each via is 1nanoHenry. Place your feedback resistors right at the Vin- (Virtual Gnd) pin of the Opamp, 1mm max distance. Assume 10pF capacitance on Vin-, and size the resistors to accomplish less than 10 degrees extra phaseshift; 1Kohm and 1pF is 1nS tau, 160MHz F3dB and 45 degree phaseshift at 160MHz, or 45/3 = 15 degrees at 480MHz; thus evaluate lack-of-oscillation with 430 or 240 Ohms Rfeedback.. Place a 100 ohm resistor, right at Vout pin, to isolate the OpAmp Feedback behavior from any downstream (next stage) capacitive loading; maybe you can reduce that 100 Ohm to 80 Ohm or 50 Ohm or 33 Ohm or 22 Ohm, or 15 Ohm, but be aware of how close you are to irksome overshoot and oscillation.

And do not use an opAmp ---- labled in the datasheet as stable from gains > 10, or gains > 2, or gains > 5, or gains > 100 ---- for unity gain performance.

Ohhhhh Do not use an older opamp, needing +- 8 volts VDD, in a 0/5v circuit

\$\endgroup\$
  • 1
    \$\begingroup\$ "Use Surface Mount Technology SMT Cs and Rs." Oh, puleez. I've done 50 MHz with through-hole parts. \$\endgroup\$ – WhatRoughBeast May 28 '17 at 23:02
  • \$\begingroup\$ Actually I forgot to mention that I feed ADC input with it. \$\endgroup\$ – P__J__ May 28 '17 at 23:03
  • \$\begingroup\$ In the circuit the CLC4007ISO14 is used. \$\endgroup\$ – P__J__ May 28 '17 at 23:52

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.