I've been handling field test equipment orders and deployments for about six years now. In my first year (2018), I made the classic mistake of assuming a new hire knew the difference between a PIM test and a simple cable sweep. The result? A $3,200 order of certified jumpers failed acceptance because we tested with the wrong tool. Straight to rework.
That was just one of three major screw-ups that led to a checklist I now refuse to deploy without. Here's the thing: most field test failures aren't technical. They're procedural. And they're almost always preventable.
This guide is for the engineer or team lead who is about to deploy an Anritsu Site Master (the MS2720T, Site Master S331L, or any of the handheld spectrum analyzers) for a critical RF survey, PIM test, or cable and antenna analysis. I'll walk you through the three checks I do before every field day. Each one has a story.
The mistake: In September 2021, I sent a technician to validate a newly installed DAS antenna system. He connected his Site Master directly to a port that still had +24 dBm of composite power from the BTS. The front-end of the spectrum analyzer was rated for +20 dBm max. The unit didn't blow immediately, but the noise floor shifted permanently. The calibration was shot. The repair cost $890 plus a one-week delay.
The cure: Before connecting any cable, do a two-second power check. Look at the band plan. If the site is active (which it usually will be for a sweep), verify you're connecting to a passive antenna port or a coupled port with at least 30 dB of isolation. I now include a DC/power block in every kit for the MS2720T. Not ideal, but necessary. This is step one on my checklist, written in sharpie on the inside of the hard case: "Confirm power is OFF or at coupled port."
The mistake: The vendor failure in March 2023 changed how I think about span settings. We were troubleshooting intermittent interference on an LTE band (Band 13, 700 MHz). The technician ran a sweep from 600 MHz to 800 MHz. The spectrum display looked clean. He logged it as "no interference present." The customer called two days later with a dropped call report. When I looked at the data, I saw a classic span error: there was a narrow spur at 737 MHz that only appeared when the resolution bandwidth was set to 30 kHz or lower. The technician had left the RBW on auto (1 MHz), which completely masked the signal.
The cure: For any interference hunt or spectrum survey, I use a two-step span rule. First, I set a wide span (the full band + guard bands) for a 10-second max hold to catch anything obvious. Then, I set a narrow span (1 MHz to 10 MHz) centered on the carrier frequency, manually set the RBW to 30 kHz, and watch the trace for at least 30 seconds. I didn't fully understand this until that $3,200 failure. The checklist item reads: "Wide scan (full band) then narrow scan (1 MHz) with RBW = 30 kHz."
The mistake: In Q1 2024, a field tech was doing PIM testing on a newly installed tower sector. He was using a PIM Master (MW82119A) and an Anritsu MS9710B for the return loss measurement. He was getting inconsistent results—PIM values fluctuating by 10 dB between sweeps. He spent hours re-torquing every connector. The problem? He was using a cheap 20-year-old adapter (N-male to 7-16 DIN) that had a bent center pin. It wasn't visibly bent—it was bent about 0.5 mm. The PIM test was actually measuring the connector's poor contact, not the antenna system's quality.
The cure: I now inspect every adapter with a magnifier before the first measurement. Yes, really. And I have a strict rule: never use an adapter that has been dropped. We've caught 47 potential failures using this rule in the past 18 months. The checklist item: "Inspect all adapters visually. Open Short Load (OSL) calibration on the same port you will test. Do not calibrate on a different bench."
To be fair, the easy path is to blame the adapter. But the real issue was a lack of a pre-check. I get why people skip it—time pressure is real. But a 5-minute adapter check beats a full day of troubleshooting.
Here's the exact checklist I tape inside the lid of my Anritsu hard case. It's saved us an estimated $8,000 in potential rework over the past two years.
When I compared our Q1 and Q2 results side by side—same vendor, same test equipment, different inspection discipline—I finally understood why the details matter so much. The checklist isn't about being slow. It's about being consistent.
Misconception: "A handheld spectrum analyzer is just a less accurate lab unit."
I used to think this, too. The reality is that a well-calibrated Site Master in the field is often more reliable for deployment validation because it's measuring the system under real thermal and environmental conditions, not a perfect 23°C lab bench. The trade-off is that you must be more disciplined with your procedure. The instrument won't save you from a bad connection.
Misconception: "Rush jobs don't need the full checklist."
This was true maybe 15 years ago when towers had simple analog systems. Today, a multi-band, multi-technology site has many more failure points. The assumption is that rush orders cost more because of overtime. The reality is they cost more because skipping a procedural step (like the RBW check) leads to a rework truck roll. I once saved a three-hour overtime drive by spending 3 minutes on the checklist. It was a good trade.
One more thing: The MS2720T is a fantastic tool. It's also a complex one. If you're training a new hire, don't let them touch the unit until they can recite the power limit (the unit's spec is +20 dBm max input; always verify this before connecting to any live port). As of January 2025, the current pricing for the MS2720T and a basic PIM Master is approximately $18,000-$25,000 (based on distributor quotes; verify current pricing at anritsu.com). A repair after a simple power overload is about $1,500 (based on a September 2024 repair invoice I personally handled). It's cheaper to spend 30 seconds on the power check.