What would be the most useful type of near field probe for EMI measurements?

I've got some ideas for near field probe types that I'll be blogging about in the near future. But, I'd like to know what engineers have found to be the most useful near field probe when sniffing out emissions violating sources.

First to note, the correlation between near field and far field is very poor. Near field (reactive fields) fall of with the inverse square law. Far field components (radiation field) fall off inversely. At close range (less than 1/6 wavelength) either the magnetic field or the electric field will dominate, but those will bear little relationship to the far field, which will dominate at typical antenna distances.

Thus, if you place a probe close to a slot in the enclosure, you may get a strong contribution, but it may be insignificant at 10 meters. Alternately, if you place a probe close to a cable, you may get a very weak contribution, but you may get excessive emissions at 10 meters.

Having said that, I find the loop most useful - it will find currents traveling on a cable or across a slot in the enclosure or even on traces and decoupling caps on the circuit board - larger loops are more sensitive but less selective, switch to small loops to home in on particular circuit board elements.

The dipole antenna is best to find a hot chip - put it right over the die to find out which chip is most noisy. The dipole is also useful for sniffing out open connectors - once the connector is filled, a loop will work better.

It is possible to make up a probe using combination of loops and dipoles - I believe this is what is in HP's probe. They sort of work, but I would never accept their calibration tables.

To summarize, you'll generally get better response from loop antennas, but don't rely on the readings you get.

Bill Kimmel

I like one inch square unshielded loops. There is one case where what the loop picks up directly correlates to far field radiation. If local fields, say on the top of an IC, can induce a millivolt or so into such a loop, then any cable passing the same spot will pick up that much or more and if at the resonant frequency of the cable driven from that point, you have an emissions problem if the cable is not in a shielded enclosure. A millivolt into a resonant antenna can generate >10 microamps of current which directly correlates to emission that will likely be a problem. There are lots of articles on my site http://emcesd.com that deal with this.

Doug

You bring up calibration tables and correction factors. I agree that it is a stretch to have a far field correction factor. Would it be best to have factors that convert the measurement to magnetic flux density?

It's not obvious to me that knowledge of flux density will be useful. But you should see if Doug Smith has some answers - he is the guru in this field.

Are you reading this, Doug?

Bill Kimmel