When I started comparing the Anritsu MW82119A and the Anritsu MG3694A, I expected the decision to be about price tags. It wasn't. After forecasting $180,000 in test equipment spending for our engineering group across 2024–2025, I found that the real differentiators are how your technicians work, what you'll be measuring, and how often you can keep both boxes calibrated. If you're about to buy either unit, this article will save you the spreadsheet work.
The MW82119A is a handheld PIM (passive intermodulation) analyzer that runs on battery. It's built for field technicians who need to test RF connectors, cables, and antennas at a cell site. The MG3694A is a benchtop signal generator that can produce frequencies from a few MHz up to 40 GHz, depending on configuration. One is a field repair tool; the other is a lab instrument. Both carry the Anritsu name, but you couldn't choose a more different pair of tools if you tried.
I'm focusing on total cost of ownership (TCO), because that's how I've managed procurement for the last six years. Unit price matters, but calibration, carrying cases, training, and idle technician time eat budgets quietly. Let's dig in.
The MW82119A weighs around 3 kg and runs for hours on an internal battery. I can hand it to a technician at 8 a.m. and get a full day of PIM tests without a power inverter. The MG3694A, on the other hand, is a heavy benchtop unit. It's not designed for tower climbs. If someone tells you they're planning to use the MG3694A for site sweeps, they're either doing something unusual or about to hate their life.
In Q2 2024, we did a side-by-side trial with both units. Our field lead spent 10 minutes setting up the MG3694A on a cart near a base station, then realized he couldn't run it without AC power. We switched to the MW82119A and completed the entire sweep in 20 minutes. That's the difference between a dedicated portable tool and a lab instrument trying to do field work.
As a cost controller, I have to respect the MG3694A's performance. It's a true signal generator. Need a clean, unmodulated signal at 20 GHz? The MG3694A does it with phase noise the field instruments can't match. That's why our R&D group, not the field team, originally requested it. For testing RF components, radar modules, or even medical devices that use wireless links (think of a Bluetooth-based blood pressure monitor), you need the MG3694A's precision.
But that precision comes with a cost beyond the purchase price. The MG3694A needs periodic calibration in a controlled environment. If you drop it or move it between labs frequently, you'll need recalibration sooner. The MW82119A also needs calibration, but its built-in self-test functions make it easier to verify in the field before each test. In our experience, that self-check catches about 12% of connection issues before they produce bad data.
Over six years of tracking invoices, I've learned to calculate TCO with five line items: purchase price, calibration, repairs, training, and the cost of wrong measurements. Let me walk through each with real numbers from our group.
The MW82119A is typically priced lower than a fully configured MG3694A, but that doesn't tell you everything. We also bought a power sensor and a set of RF cables for the MG3694A. Those accessories pushed the initial invoice up by 25%.
We sent the MG3694A to Anritsu's lab once a year for calibration, at a cost that averaged $850 including shipping. The MW82119A also goes to the lab, but because our techs perform field alignment checks, we've been able to stretch the factory calibration interval to 18 months without issues. That's a real saving.
The MG3694A's warm-up time is 30 minutes in a cold truck. That's idle labor. The MW82119A is ready in under a minute. Multiply that 30 minutes by 50 site visits a year and you lose 25 hours of billable technician time. At $75 an hour, that's $1,875 of wasted spend just waiting for a signal generator to wake up. The 'blood pressure' of the network — checking whether the RF path is healthy — is exactly what the MW82119A is built for. Don't ignore the warm-up cost.
It's tempting to think that a signal generator can do everything a test analyzer does. After all, both produce RF signals. But the MG3694A doesn't measure PIM. It doesn't have a built-in return loss bridge. You'd need to buy add-ons and still end up with a setup that's worse than the MW82119A for that job. The reverse is also true: using the MW82119A as a lab-grade signal generator would be a mistake. That's oversimplifying two very different tools.
When I compare product pages, like when our security team showed me the vSRX product page for virtual firewalls, I see a similar pattern. Software promises to replace hardware, but for physical RF testing, you still need a physical instrument. A virtual signal generator can't deliver -40 dBc harmonics from an SMA port.
You're a field service organization, a tower contractor, or a small team that needs one box for RF quality checks. It's portable, the battery lasts a day, and the training cost is low. If your work revolves around PIM testing on cell sites, this is the cost-effective choice.
You run an R&D lab, you're building communication modules, or you need a dependable source of RF signals for automated test benches. It's a serious investment, but the MG3694A is a workhorse. Just make sure your lab has a controlled environment and a calibration budget that includes annual recalibration.
You're a larger telecom group that supports both field maintenance and product development. The two instruments complement each other. You'll have to maintain separate calibration schedules and spare kits, but the flexibility to respond to any test request is worth it. Last year, the field team borrowed the MG3694A twice for special emitter testing, and the lab used the MW82119A to demonstrate a PIM issue on a prototype connector.
The prevention-over-cure principle applies here. The MW82119A's 5-minute self-test prevents a 5-day rework cycle when a site's coverage drops. The MG3694A's annual calibration prevents a month of bad measurements in the lab. Neither instrument is 'expensive' when you factor in the cost of being wrong.
To make the numbers concrete: over the last six years, our group spent $180,000 on test equipment and calibration. We cut unplanned rework spend by 17% simply by switching to field-level PIM testing before we paid a tower crew to climb. I'd argue that's the best ROI we've gotten from any Anritsu purchase.
If you're still on the fence, put together a TCO sheet that includes warm-up time, calibration, and worst-case rework. That spreadsheet will lead you to the same decision it led us to: you need the tool that matches the job, not the one with the prettier product page. And maybe save a link to this article for the next time someone in your group suggests 'one box' can do everything.