Cosmin and Shabaz are posting excellent blogs that demonstrate the capabilities and cool features of the Keysight N9322C basic spectrum analyzer and the TI CC11XL dev kit. See Cosmin use the TI CC11XL and the N9322C to demonstrate Doppler shift in this blog post and get a good overview of the N9322C and the CC11XL in this blog post. Shabaz provides a great tutorial on using the N9322C tracking generator to evaluate antennas in this blog post and a very straightforward demonstration of FM demodulation and analysis capabilities in the blog post.
To provide another perspective to readers, my focus will be on side-by-side comparison of the Tektronix MDO4000 SA and the Keysight N9322C SA.
It is informative to see how similar instruments portray information about nearly identical data. As users we want to trust the measurements our instruments provide us. Generally we can trust the measurements our instruments provide us as long as we use each instrument correctly, keep them calibrated, and understand the specifications and limitations of each instrument as set out by manufacturers.
I set up an MDO4104-3 beside a Keysight N9322C BSA and connected a basic antenna to the RF input on each instrument. The instruments were configured as follows:
| Instrument | Center Frequency | Span | RBW | Ref Level |
|---|---|---|---|---|
| Tektronix MDO4104-3 | 876.813200MHz | 100.000 kHz | 300 Hz | -50.0 dBm |
| Keysight N9322C | 876.813200MHz | 100.000 kHz | 300 Hz | -50.00 dBm |
The plan was to let each instrument examine a small 100 kHz slice of the spectrum around 868 MHz, picking up and displaying nearby transmissions. In conducting this experiment I learned several things:
- Something (or things) in my study transmits frequently at 867.81 MHz and at 867.78 MHz
- The transmissions are burst-y and seem to end with a period of what might be steady rate FSK? Perhaps someone with more experience looking at FSK transmissions can explain what is happening at the end of some of the captured transmissions.
- The Tektronix display is cleaner and easier to read. The Keysight display suffers from clutter and what to me are non-intuitive layout issues. I go deeper into the clutter and layout issues in my full review on the Road Test site.
- Keysight has superior PC interface support. It is possible to capture AVI video of the display at up to 15 fps. That is fast, but not fast enough to make this comparison work. The Tektronix PC interface does not allow video capture of the display - at least not that I could find.
- Under the specific conditions of this comparative test, the Keysight BSA is more responsive and has a better probability of intercept (POI).
- Both instruments have good noise floors. In this live test case the noise floor on the Tektronix was slightly better that on the Keysight, but the Keysight surpassed the Tektronix when the built in preamp was activated. Both instruments were set to average 16 traces. The photos below show the noise floor on the Keysight with and without the preamp turned on.
N9322C noise floor WITHOUT preamp - about -125 dB
N9322C noise floor WITH preamp - about -132 dB
Compare these with the photo below which shows the noise floor on the Tektronix.
MDO4104-3 noise floor - just shy of -130 dB
The video clip below shows the two instruments measuring identical slices of spectrum in real time with the same span, RBW and amplitude settings, detailed in the table above.
To create this video I used two HD cameras, one pointed at the N9322C screen, the other pointed at the MDO4104-3 screen. The cameras were started and stopped at the same time and I used clicks picked up by both cameras in the audio tracks to align the video in this dual screen view.
Now, to be fair, the RF signals being fed into the inputs of the two instruments are not identical. The signals presented at the RF inputs for analysis come from similar, but not identical, antennas separated in space by about one foot (25 cm). Although the radiated signals impressed on each antenna will have the same modulation and carrier frequency, signal amplitudes will likely be different due to location differences and antenna characteristic differences. This is not a calibrated, high rigor test. The objective was to compare, in real time, how the two instruments react to and display real life RF signals.
What I'm seeing in the video is that the Keysight N9322C is faster at analyzing spectrum under these conditions. I will estimate it is 3 times faster than the Tektronix MDO4104-3. One consequence of the faster analysis time is that the N9322C provides a more detailed and complete picture of the bursting transmissions within this slice of spectrum. Compare the two instrument displays in the video during a transmission burst and you will see that the N9322C is updating more frequently and providing more detail within each channel even though both instruments are set to the same span and RBW.
I think it is important to note that the video does not show that the Tektronix screen is larger than the Keysight screen and that alone is a nice plus for the Tektronix SA. Also, because the Tektronix SA is less featured than the Keysight SA, it took less time and fewer steps to set up the MDO4104-3 compared to the N9322C. However, in spite of the larger screen and simpler user interface, in this case I prefer the faster updates and better probability of intercept provided by the N9322C.
There are things the MDO4104-3 does better in my opinion, mostly thanks to the tight integration of powerful time domain tools with frequency domain tools. The N9322C is a dedicated spectrum analyzer and as such I expect it to have better frequency domain performance, but it too has some time domain analysis capability such as AM/FM/FSK/ASK demodulation. The N9322C can display time domain waveforms of demodulated signals, and eye diagrams, and bit streams, but the MDO4104-3 can correlate spectrum to independent 'scope channels, so . . both instruments have very appealing and remarkably useful features. Which is better really depends on the measurement needs arising from situation to situation.
