Methods to improve Radiated Susceptibility

[ankitvirdi4]

Hello all,

My product is a type of a measuring instrument and it is susceptible to RF at about 85-100MHz (Loss of accuracy).

I tried filter caps , but I mostly think its the trace length.

I couldn't find any good articles related to solving RS test failures, if anyone can post the links to them or help me with what other ways can I improve its susceptibility at RFI. My circuit doesn't have any opamps, I am using a CT. Also, ferrite beads are kept as the last option, is there any other thing that I should explore?

 

[ravindragudi]

It would be helpful for others to suggest you solutions to the problem if you provide detailed description. However, I would suggest, adding up ferrite beads at the very entry of the signal on PCB will help reduce the problem. If the signal is a differential one, then use common mode chokes. If this doesn't help then you have to use other EMI RFI filters available off the shelf. Eg - MEM2012F50R0, MEM2012F75R0 from TDK, NFL21SP706X1C3D from Murata Electronics.
Another method could be to use a shielded cable.
If the issue is mainly in the trace length on PCB, then try moving these traces to inner layers covered by GND and PWR layer (if it is a multi layer board). If it is due to the wiring coming external to the PCB then the methods suggested above will be helpful.

 

[ankitvirdi4]

It would be helpful for others to suggest you solutions to the problem if you provide detailed description.

Thank you for your reply, It is an ammeter, I am using a CT and measuring voltage across the burden using differential ADC of the micro- controller.

The EUT is accurate when tested for RS at full scale, but when no current is applied, a pick up of about 30mA is seen at during 85-100MHz, amplitude of 10V/m , 80% AM at 1Khz. I am testing in a GTEM chamber and I understand there will be reflections causing unpredictable failures.

Also, above suggested EMI filters are expensive and hence not feasible, I am using a single side PCB.

 

[ravindragudi]

In that case you can definitely use a common mode choke, this will certainly help. Also try putting a up a small copper foil sheet where the sensitive traces are there on the PCB. Also try using X Y caps on the differential signals - careful with the ADC specs though.

 

[ankitvirdi4]

A common mode choke right after the burden? Or can use a SMD choke right before the ADC input?

Will take the above said trials and tell you what happened

 

[FvM]

My circuit doesn't have any opamps, I am using a CT.

Surely it's not the CT causing susceptibility problems. You missed to describe the involved signal processing. If e.g. the CT is directly interfacing a µC analog input, that's the point where you should focus the filtering attempts, as well as the (analog) power supplies.

 

[ankitvirdi4]

You missed to describe the involved signal processing. If e.g. the CT is directly interfacing a µC analog input, that's the point where you should focus the filtering attempts

Yes, the CT has a burden of 15 Ohms both ends which go the uC analog input decoupled using 47pF, I will try changing the decoupling cap to 0.01uF. Any other improvements?

 

[FvM]

A solid ground plane for the circuit, bypass capacitors at all µP supply pins directly connecting to the ground plane. RC or LC filter at the analog pins also directly connecting to the ground plane. Ferrite beads for all connections leaving the ground plane.

 

[marce]

If it is a single sided design then anything you do is compromised, the design needs at least two layers so you can get some ground plane.....
As you said radiated what sort of enclosure do you have...