Real-time answers to technical questions

Frequently Asked Questions

1. How can I correlate the scan results to PCB design?
EMxpert supports two types of file formats: Gerber file format RS274x and HPGL. On rare occasions, a Gerber file may not be in a pure RS274x format and cannot be read by our Client. This file can often be read by a third party Gerber reader such as GerbView and converted to an RS274x format that can be used. GerbView can be obtained at http://www.softwarecompanions.com.
2. Does the system require annual calibration?
No, it does not. The scanner elements are passive and will not require calibration. If you believe that one or more of the probes have been damaged, the unit can be returned to EMSCAN for verification and repair.
3. What is the frequency and amplitude accuracy of the EMxpert measurements?
Frequency Accuracy: Marked Peaks - Typically accurate to 1 KHz
A calibrated spectrum analyzer or EMC analyzer normally has a frequency error of between 1% and 2% of span. During the process of marking peaks, EMxpert conducts several rescans of identified peaks after the initial scan is completed. The final accuracy is 1-2% of the final scan bandwidth shown on the spectrum analyzer display.
Some supported models of spectrum analyzer provide higher accuracy frequency measurements using an internal counter. For these spectrum analyzers the final accuracy is the specified accuracy of the counter.

Amplitude Accuracy: Spectral or spatial scans: approximately +/- 3dB
This is a combination of spectrum analyzer accuracy and probe factor responses from cell to cell across all frequencies from scanner to scanner.
4. How do I achieve higher frequency accuracy?
Frequency accuracy is compromised by the spectrum analyzer frequency measurement accuracy of 1% to 2% of span, unless the spectrum analyzer is equipped with a higher accuracy frequency counter. Therefore, by decreasing the span, the user can increase the frequency accuracy. Since the EMxpert final rescan for peak detection is set at a span of 5 x RBW, reducing the RBW has the effect of improving the frequency accuracy.
5. How do I achieve higher amplitude accuracy?
Decreasing the resolution bandwidth reduces the noise in the amplitude measurement for spectral and spatial scanning. For best results in a spatial scan, use an RBW of no greater than 100 KHz.
6. I have a spectrum analyzer that is not on the list of supported analyzers. Can it be supported?
In order to accomplish the real-time scanning that is fundamental to EMSCAN, the EMxpert only requires gated sweep option on the spectrum analyzer for real-time spatial scan.

If your spectrum analyzer is not on the supported spectrum analyzer list, please contact your local representative or EMSCAN at info at EMSCAN dot com to discuss the development of a custom driver.
7. How do I make sure that I see all the peaks that might cause me problems in the far-field?
There are instances where components of the DUT are some distance away from EMSCAN's Patented Scanner, and signals from these devices may be mixed with the noise floor. We recommend that a low noise amplifier be placed in series between the scanner and the spectrum analyzer to differentiate the signal from the noise floor. We also recommend that you reduce the RBW to improve the signal to noise ratio.

It is important to recognize that some of the far-field emissions may originate within the DUT, and then couple onto the connectors and cables (which may act as a high gain antenna). The signals, when still on the DUT, can be of low amplitude, while their impact on far-field measurements (when conducted out and radiated), could be up to 100,000 times as large (for further details ref. to Common Mode).
8. What are the Minimum System Requirements?
CPU 333MHz or faster
128 MB RAM, 256MB recommended
8MB Video or more
Windows NT/2000/XP/Vista/Windows 7 OS
10/100 Network Card
CD ROM Drive
9. What is the maximum power level that EMSCAN's Patented Scanner can receive without being damaged?
The surface of the scanner can handle thousands of volts. The scanner is rated for a radiated load of 10W or 40 dB. If a transmission line carrying RF power of more that 10 W / 40dBm is placed close to the scanner probe for extended periods, it is likely to damage the scanner.
10. How sensitive is EMSCAN's Patented Scanner?
The dynamic range (sensitivity at the low end) is limited by the Spectrum Analyzer Noise floor. So this is a function of the spectrum analyzer type, and of course the RBW selected as well as the use of an LNA.

To demonstrate the sensitivity of EMSCAN's Patented Scanner, we tested 50 ohm micro strip line using spectrum analyzer model HP8563E with a 40 dB (NF=2.9) gain LNA and set the RBW to 1kHz.

Minimum sensitivity of EMSCAN's Patented Scanner:
1 MHz = -44 dBm
100 MHz = -85 dBm
1 GHz = -100 dBm
2 GHz = -96 dBm
3 GHz = -85 dBm
4 GHz = -65 dBm
11. I added a low noise amplifier (or an attenuator) between the scanner and the spectrum analyzer and forgot to enter the value in the Amplitude Adjustment menu. Can I correct this without rescanning?
If you forget to set the amplitude adjustment, you can re-enter the Amplitude Adjustment area and modify the processed scans without rescanning.
12. How can I find the frequency of an intermittent or random signal?
Run a continuous spectral or spatial scan with peak hold. It will replace the peak amplitude with each successive higher measurement.
13. Why does the peak marking process take longer than the estimated time to complete?
EMxpert builds a peak list initially during the spectral scan. Then it goes back to each and every peak and zooms in to reduce the span and RBW to have accurate peak.
14. I am concerned about mutual interference. What feature should I use?
Mutual interference is important when you integrate your PCB with others into enclosed systems. Limit lines enable you to set the rules of allowable magnetic field strength and specify acceptable ranges.
15. I am doing Before and After testing. What feature should I use?
By setting a limit line that shows the current performance, it can easily be determined if the design improves or gets worse.
16. Can I create Limit Lines for FCC requirements?
EMxpert is a near-field pre compliance test system and does not provide far-field measurements. Thus you can't create limit lines that reflect FCC requirements.
17. I want to do a Pass or Fail test. What feature should I use?
Use the "Margin Table" feature of spectral and spatial scripting module. The Margin Table shows whether you pass or fail i.e. shows the variations that you encounter with your own designs and third party designs.
18. Why is the result of the spectral scan different than the image displayed on the spectrum analyzer?
EMxpert measures DUT's emission on the surface of the scanner. There is an insertion loss between that point and the RF input into the SA. There are different insertion losses for each and every cell because the path is slightly different. We also calibrate to account for the response across a wide band of frequencies. The scan results on the client viewer and SA are not the same because the SA is not compensated for this insertion loss.

You should see consistency in terms of likely frequencies that are exhibiting or characterize as peaks but you will not see the same levels. If you turn off the probe compensation feature, you will see the same pattern.
19. How long is a typical spectral scan?
45 seconds for a 4" x 4" PCB with a span of 100 MHz and an RBW of 100KHz. Scanning area, span and RBW are user selectable.
20. How long does a typical spatial scan take?
Less than one second if using the full scan plane with time gate sweep feature.
21. Why do the spatial scan results change when I rotate the DUT on EMSCAN's Patented Scanner?
The probes for the scanner have been custom designed by EMSCAN to provide sensitivity over a large angular region. A scale model of the probes is shown below. It effectively combines two separate loop probes into an 'X Loop'.
The X loop is much better than a traditional loop in terms of response versus field orientation as shown in the figure below. There is still however a blind spot. It was chosen to orient this blind spot at 45° and further to rotate every other row by 90° so that consecutive rows do not have the same blind spot.
The effect of this setup is to provide a cumulative sensitivity as is shown in the figure below. This is a small 5x5 array taken from the larger 1218 probe scanner.
The effect of this different sensitivity on scan results sometimes shows up when a device is rotated. If the device has only vertical field components, the scan results will be as is shown below in the top right. If this device is rotated 90°, then results are as expected and as shown in the bottom right.
However, if the device has a diagonal field then every other row will be in a dead spot and the resultant measured field will be very low. The scan results are shown in the top right. Now if this device is rotated 90° then dead spots will have switched to the alternate row. Although the high level results are as expected the presence of these low values distracts the eye. Furthermore, since the dead spots do not rotate they may give the impression that the scan results are completely different.
In certain situations the presence of dead spots can create misleading results. If one focuses on looking for patterns and continuation of signals these dead spots can usually be ignored. The display setting which does interpolation makes this much easier by smoothing out transitions. An example of this is shown below.
22. I am designing multiple layer boards. How can EMxpert help to diagnose EM problems?
Currently, there are no measurement techniques on the market that can exactly see what is going on inside a multi-layer board. A chamber for example, will only deal with far-field and will not differentiate between what is coming from the edge, the most external layer, the internal layers, and the interaction between the layers. It is the most blind of all EMI test solutions to help design a better PCB with EMI reduction at the source. Automated single probe, handheld probe, and EMxpert can differentiate at least between the emission from the edge (where EM from the internal layers will most likely escape and emit) and from the surface (EM for the most external PCB layers).

If one uses a pre-amplifier, the scanner or probe could pick weaker signals from the internal layers leaking through the external layers. A comparison of the spectral and spatial analysis with or without amplification could then help point the source (frequency may help specify a component, and the location will do as well). The EMxpert and the automated single probe could handle that; the handheld probe will not work.

Clearly, the EMxpert and the automated single probe are the only adequate tools for multi-layer PCB EMI measurements. However, the fact that the EMxpert is much faster, can also detect hot spots and current loops, and can track intermittent events through peak hold with continuous scanning, make it the best solution overall.

If you can't find answers to your questions, contact us