THS3491: Interference occurs even without input and when there is a 30 MHz signal. It seems to be caused by self-excitation.

Hi Xinglin, 

I think that the noise is coming from the power supplies. In the datasheet of the device its mentioned that you should use a minimum value of 6.8µF decoupling caps at the supply pins and place them as close as possible to the supply pins. I see in your schematic that you have used only 0.1µF decoupling caps in your design. 

Can you please try using 6.8µF decoupling cap or higher and check.

I have attached the screenshot from datasheet explain the same:

Regards,
Aditya

It should not be the power supply problem. the datasheet mentioned 6.8µF capacitor is in DC power supply section, this capacitor in the more distant area of the chip. the DC power in this circuit is a quite precision equipment. Decoupling is accomplished within the power supply itself.

When the coaxial cable at the input port of the circuit is removed, this interference will disappear. However, once the signal is connected, this phenomenon will occur after a while. Even if the input signal is removed or the coaxial cable is disconnected at the signal generator side after the phenomenon occurs, the interference will still exist. Only when the coaxial cable connection at one side of this circuit is disconnected will the interference disappear.

Hey, have you tried using a new coaxial cable? Because we have seen such issues in past when the coaxial cable went bad.

Regards,
Aditya

I have used different cables, but its not really effective

Hey, will look into it and get back to you.

Hi Xinglin,

1. Which EVM (Evaluation Module) are you using? Is it TI EVM or something else, can you please send a photo of the EVM board and the setup please.
2. What are the amplitude of the oscillations? Can you send the full picture of the oscilloscope please?

Regards
Aditya