Archive for December, 2014


-July 19, 2013 in EDN

Questions on EMC pre-compliance testing for radiated emissions

Questions on EMC pre-compliance testing for radiated emissions

Thanks for all the great questions presented following my recent EMC webinar, sponsored byRohde & Schwarz and hosted by UBM TechOnline. If you missed the webinar, you may go here to download a copy of the slides and listen to the webinar “on-demand”. As I mentioned in the previous two postings, I’ve grouped them by topic and will be answering them all the best I can. Be advised that for many questions pertaining to EMC, the best answer is, “it depends”, so there may not be one answer for all cases. I’ll try to include my assumptions in the answers. The questions have been edited for clarity.

This posting will address questions on Pre-Compliance Testing for Radiated Emissions.

Q. Most folks don’t have access to a semi-anechoic chamber. Any recommendations for the room used for radiated pre-compliance tests?

A. True…most semi-anechoic chambers are very expensive. I’ve been very successful using conference rooms (or better) basements to set up a simple 3m measurement range for confirming pass/fail data. Of course, I like to give myself about 6 dB of margin, just to lower the risk of failing in a real chamber.

Here is one major issue: Ambient signals (that is, signals from external sources like AM & FM broadcast, television, two-way radios, military transmitters and mobile phone and wireless communications) can mask the harmonics from the product, unless you can test in a shielded room. Making the measurement in a rural area or basement can help. Knowing where the major emissions lie is important, because you can then narrow down the span to include just that harmonic. If the harmonic in question is still being masked by an ambient signal, often, you can narrow down the resolution bandwidth (assuming the harmonic is a narrow band “CW” signal) and make an approximate amplitude measurement.

Another issue is that very strong ambient signals, say from a nearby FM broadcast or television station, can overload the front end of the EMI receiver (to some degree, depending on whether it has preselection filters) or spectrum analyzer. Because spectrum analyzers have very wideband front end circuitry, a signal that overloads the analyzer can affect (compress) the other measured signal amplitudes.

You’ll also need to know the system losses and gains, so you can calculate the voltage at the antenna terminals. This would include coax loss, attenuator losses (if used), preamp gain (if used), and the antenna factor (provided by the antenna manufacturer).

Q. The equipment you show is for pre-compliance or to resolve a problem. For EMC full compliance, which are the advices you give in order to comply with the tests?

A. You really have two choices; assemble the equipment and semi-anechoic chamber required to perform your own radiated emissions testing (can cost $500k to $3M, depending on 3m or 10m chamber), or have a third-party test facility run the compliance testing for you. Unless your company is willing to fund such a project, my advice is to rely on an outside test lab.

Q. How to setup pre-compliance test lab, what are minimum possible equipments required with limited budget?

A. If your budget is limited, the minimum equipment required is a good spectrum analyzer ($2 to $20k), an EMI antenna (or antennas) that can tune from at least 30 to 1000 MHz (however, the U.S. FCC can require testing up to 6 GHz, or more), a turntable to set up the product with its attached cables, and an antenna tower that can be adjusted in height and can turn the antenna from horizontal to vertical polarization. You may also need a preamplifier to boost the signal from the antenna. It’s possible to make a DIY turntable and tower to save a little money. The total cost might run $10k to $50k. Note, also, that the frequency range required also depends on your product family and test standard used. For example, some ITE and medical products require measurement all the way to 30 GHz, which could double those costs.

Q. If you use a current loop probe to measure the dBuA on a spectrum analyzer and calculate the current for differential-mode currents, do you still need to input the L & s of the formula you mentioned to calculate E field?

A. Yes, you do. An estimate works OK. The main thing is to capture the length of the circuit trace or cable.

Q. CM and DM calculations – where did the constant come from? How was it derived?

A. These are derived in Dr. Clayton Paul’s book, Introduction to Electromagnetic Compatibility (2nd ed.) starting on page 509 (Differential-Mode Current Emission Model) through page 516 (Common-Mode Current Emission Model). Assuming we start with the Hertzian dipole model, the DM and CM constants are comprised of the intrinsic impedance of free space, phase constant and a factor of 2*PI or 4*PI.

Q. Why is conducted emission’s maximum frequency 152 MHz?

A. I don’t recall mentioning 152 MHz, however, generally speaking, common-mode emissions start to decrease around 200 to 300 MHz as a dominant factor in harmonics. That’s not saying that CM currents don’t exist above that frequency – just that they aren’t the dominant source.

Q. Could you use that scope connected to an antenna for wide band radiated emissions?

A. Interesting question, as most designers have access to oscilloscopes, rather than spectrum analyzers. As well, most digitizing scopes also have FFT functions. The answer is it depends a lot on the sensitivity of the scope in question. Most scopes can’t measure accurately below a few mV and the signal levels at the antenna terminals might be in the uV region. So, unless your scope can measure down to the uV level, it probably won’t work well. Some of the newer scopes from Rhode & Schwarz and LeCroy have sensitivities in the uV, however, so they might be worth a try. However, you’ll miss out on the convenience of dedicated spectrum analyzer controls, of which there are many. If cost is an issue, you might check out the Rigol DSA815 analyzer that I reviewed recently.

Q. How accurate are FFT readings from a oscilloscope using regular oscilloscope probes compared to near-field probes?

A. Hard to say, as most near-field probes aren’t really calibrated. Now, if the question is really, “Can I perform troubleshooting with either near-field probes or scope probes”, the answer is “yes”. The other part of your question relates to the accuracy of the FFT data on a scope. I would have to say, most of the newer scopes will be fairly accurate, however input sensitivity and internally generated noise in the front end will come in to play and mask very small signals. It’s important that the scope have the ability to be adjusted to a resolution bandwidth of 100 to 120 kHz (for the 30 to 1000 MHz band) in order to compare accurately with spectrum analyzer measurements. But consider the discussion above, as well.

I have written extensively on how to perform pre-compliance testing for radiated emissions and would invite you all to read some of the postings here on The EMC Blog, as well as the links to some of my articles in other magazines.

Feel free to add additional questions related to PC boards. I’ll be posting additional questions asked during this webinar in later blogs. If you missed the webinar, you may go here to download a copy of the slides and listen to the webinar “on-demand”.


BRL Test is your EMC EMI Headquarters

December 3, 2014 | by Lisa Winter

Researchers Closer To Using Light Instead of Wires In Computers

Researchers Closer To Using Light Instead of Wires In Computers (photo credit Vuckovic Lab)

Light is much more efficient at transmitting data than electricity can through wires, but getting it to work reliably in a computer has been somewhat problematic. A team of engineers have just announced a new “optical link” device made out of silicon that is able to bend light at right angles, which is an important advancement toward replacing electric wires in computers with optics. The research was led by Jelena Vuckovic of Stanford University. The paper was published in the journal Scientific Reports.

“Light can carry more data than a wire, and it takes less energy to transmit photons than electrons,” Vuckovic said in a press release.

The current paper builds off of the lab’s previous work, in which Vuckovic’s team developed an algorithm that allowed for necessary optical devices to be developed automatically. It also allowed them to design the nanostructures necessary to manipulate light for optical data transmission.

It was that algorithm that allowed the team to build the optical link: a very small piece of silicon with nanoscale vertical etchings. The eight-micron-long link acts like a prism, breaking down beams of light based on wavelength. The etchings are shaped so they direct the light at 90 degree angles in opposite directions, forming a T. The ability to manipulate the light in this manner is a significant step forward in optical data transmission.

The link is made out of silicon because its index of refraction (an indicator of how quickly light travels through a certain material) is 3.5. This is much slower than infrared light moves through water (1.3) or air (just about 1). The spaces between the etched lines allow the researchers to precisely manipulate how the light will be reflected and transmitted as the light passes between air and silicon.

“We wanted to be able to let the software design the structure of a particular size given only the desired inputs and outputs for the device,” Vuckovic explained. “For many years, nanophotonics researchers made structures using simple geometries and regular shapes. The structures you see produced by this algorithm are nothing like what anyone has done before.”

Though the link is an impressive device now, the algorithm first approached it as an ordinary piece of silicon. The etched lines were added and tweaked as necessary, based on the success predicted by the computational model’s use of convex optimization. Though the computer needs to run several hundred trials to make sure the optical link is calibrated correctly, it only takes about 15 minutes to do.

“There’s no way to analytically design these kinds of devices,” lead author Alexander Piggott added, speaking to the superlative advantage the algorithm brings.

In addition to the T-shaped beams of light generated by the link in this study, the algorithm makes it possible to have endless ways to manipulate light, which could be very useful in optic data transmission as the field progresses.