I'm trying to process a sinusoidal signal into a square wave at frequencies above 1 MHz. This is my first time doing this so it doesn't work properly. If anyone can help with an explanation, I would greatly appreciate it.

This is the circuit:

AD8611 circuit

AD8611 datasheet

According to datasheet this comparator should operate on frequencies upto 100 MHz.

At frequencies 1 MHz - 10 MHz everything seems to be fine.

The waveforms look like this:

1 MHz

10 MHz

The higher the frequency, the less the output on the comparator resembles a square wave:

enter image description here

This is how it should be, but I still want to ask for advice.

Am I doing everything right? How can I improve this circuit?

The comparator is designed for frequencies up to 100 MHz, but above 10 MHz the square wave begins to lose its shape.

The output of the comparator will be fed to the bipolar transistor.

The probe for the output is set to 10X, the probe for sine wave still in 1X poistion. Here are the udpdated screens:

1 MHz

10 MHz

50 MHz

Now it looks more realistic.

  • 1
    \$\begingroup\$ Is the op amp rated to have its inputs driven negative with respect to its negative power supply? That seems likely to cause problems to me. \$\endgroup\$
    – Hearth
    Jul 3, 2022 at 20:12
  • \$\begingroup\$ What does your layout look like? Probing technique? Also, what is the specified rise/fall time of your oscilloscope? And @Hearth indeed, it appears to have input clamp diodes, making this an illegal configuration >0.3Vpk. The input swing of several volts handily violates this, hence the squished bottom phase (which I assume is supposed to be an ordinary sine input). \$\endgroup\$ Jul 3, 2022 at 20:24
  • \$\begingroup\$ fig 23 of this : analog.com/media/en/technical-documentation/data-sheets/… seems ... saying that it can work for a "sine" (? what kind ...) input frequency of 50 MHz. \$\endgroup\$
    – Antonio51
    Jul 3, 2022 at 21:13
  • \$\begingroup\$ This is one of those times when a photograph of your setup would be helpful. Since you are not exeperienced at this stuff, I'd guess there are a number of things you're doing wrong, and a picture would help. But the first thing I'd check (while taking the picture) is your scope probe. For these frequencies you MUST use a 10x probe. \$\endgroup\$ Jul 3, 2022 at 22:13
  • \$\begingroup\$ @Hearth It seems that following datasheet (fig 23 & 24) that input can be "negative" ... \$\endgroup\$
    – Antonio51
    Jul 4, 2022 at 10:59

1 Answer 1


Since both waveforms are BW limited, the biggest problem is the 1:1 probe and perhaps the DSO BW.

Checklist for 100 MHz output

  1. Low ource impedance low (<=400 Ohms)

  2. Low pF probe (10:1 only, calibrated on square)

  3. Short probe ground (< 2cm or tip & ring only)

  4. Short traces from output and ground

  5. Decoupling cap <= 1cm from IC.

  6. DSO not limited by 20MHz filter and pref. Rated for >=-300 MHz.

  7. Expect prop delay and risetime to meet spec. If all other conditions satisfy datasheet.

  8. don't expect textbook squarewaves on your 1st design, that might require CML or ECL logic comparators to include at least 7th harmonic.

  • \$\begingroup\$ Indeed, I was confused why 10x probes give another set of data rather than 1x probes. Was interested why, but you indicated this is a MUST for certain reasons. My oscilloscope is far from being a professional tool. It is a Siglent SDS1104X-E. \$\endgroup\$
    – Way Run
    Jul 3, 2022 at 22:24
  • \$\begingroup\$ As for layout: all the components are located on a breadboard (yes, it is awful in terms of noise and reliability), but this is done just for studying... \$\endgroup\$
    – Way Run
    Jul 3, 2022 at 22:34
  • 1
    \$\begingroup\$ Also checked DSO: it is not limited by 20 MHz. but is rated to 100 MHz only (default mode in all measurements). Switching to 10X helped indeed. At least I can see now the difference in everything: more details in 1 MHz, stable and confident hills on 50 MHz. Thank you \$\endgroup\$
    – Way Run
    Jul 4, 2022 at 10:24
  • \$\begingroup\$ 10:1 probes reduce C and current when raising R which increases the slew rate , ground length due to 10 nH/cm also causes resonance so zero length is preferred by removing tip and gnd clip using only tip+ring at output+gnd. This is for square waves > 20 MHz. Probe must be calibrated. \$\endgroup\$ Jul 4, 2022 at 20:59

Your Answer

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

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