How to kill harmonic peaks generated by bias tee?

I inserted an Aeroflex 1-26 GHz bias tee into my circuit, and it caused terribly large DC noise at 34's Hz harmonic frequencies(as picture below).
How can I kill all this peaks?

Thank you!

You are obviously feeding a 34 Hz signal to your circuit. Do you think the bias-T (a passive circuit) is generating it? Did you check the power supply?

Um...not sure. But at least the noise shows up after bias tee is included, no matter the RF power is ON or OFF. I then substituted a 2-4 GHz narrow band bias tee for the noisy one, and the 34 Hz noise in DC was totally gone! However, it does not fit my working frequency. By doing this comparison, can I conclude the 34 HZ is generated by the Aeroflex broadband bias tee?

The DC port of bias tee is connected to a preamplifier for STM, should that be a cause?

If it's not the power supply, it may be a low frequent self-oscillation with feedback through the DC port. Bypassing the DC port with large capacitors (100 - 1000 μF) might help.

I vote for ground loops between you test equipment. Try an inside/outside DC block at your spectrum analyzer input port:


http://www.pasternack.com/inner-oute...-category.aspx