"When Was This Cable Ready for Service?" An RF Quality Inspector's Field Guide

I ask that question at every site audit. “When was this cable ready for service?”

The answers I get are... telling. Some engineers read the date off the installation ticket. Others quote the day the radio was powered on. A few say “shop tested it before we brought it out” (which, honestly, means nothing once the cable has been coiled, transported, and manhandled through a roof hatch).

The engineers I trust pull out a test report. Not a ticket, not a memory, not a guess. A report with sweep data, distance-to-fault plots, limit lines, and a date. That's the only real answer to the question.

“Ready for Service” Has a Specific Meaning

In my line of work, I'm the quality inspector who reviews every cabling deliverable before it goes live. Roughly 40 sites a year, 200+ individual cable paths. I review them because the cost of a bad one is never just the cable. I've rejected 18% of first-time deliveries in 2024, mostly for connector issues that wouldn't have shown up in a basic go/no-go check.

“Ready for service” means the return loss, VSWR, and insertion loss of every path have been measured against the system requirement—across the actual operating frequency range, not just a band you feel comfortable with. It means the connectors are mechanically verified, and the baseline data is stored so a technician six months from now (who has never seen this site) can compare and spot drift before it becomes an outage.

Basically, it's an evidence-based statement, not a visual inspection and a nod of the head. The problem is that most field testing doesn't get that far.

What Actually Goes Wrong With “Good” Installations

Let me walk you through the kinds of defects I see that would never show up in a basic continuity test.

DC Continuity Tells You Almost Nothing

The crushed dielectric story is the one that changed how I think about field testing. We accepted a feeder run at a site where continuity was fine—the multimeter said the inner conductor was connected end to end, the outer shield was connected, everything looked great. The link came up. All lights green.

Except the cable had a crushed section in the middle—the kind of damage you get from standing on a jumper or closing a hatch on it. At certain frequencies, that crush created a reflection big enough to bounce signal around like a ping-pong ball. The radio could still connect, but the system was effectively losing power at the exact frequency we needed it most. And the damage was completely invisible from the outside.

I didn't fully understand why VSWR sweeps were non-negotiable until that site. A sweep caught it in minutes. The return loss peak doesn't lie.

Cable Faults Are Frequency-Dependent

Another thing that surprised me when I moved into acceptance testing: a cable can pass with flying colors at one frequency and fail badly at another. A run that sweeps beautifully at 900 MHz may be completely unacceptable at 3.5 GHz. This is why “we pre-tested it at the shop” is only useful if the shop tested it at the actual operating frequency of the network, with the actual connectors terminated.

Never expected that to be such a recurring problem. Turns out it happens more than you'd think—especially with jumpers where the impedance match falls apart at higher frequencies.

The Connector Is Where Cable Installations Go to Die

Look at any batch of cable acceptance failures and you'll see a pattern. The cable itself is rarely the villain. It's the connector termination that fails—a center pin that sits a fraction of a millimeter too low, a nut torqued with a wrench that was reading 30% high, moisture hiding behind a connector boot that wasn't sealed properly.

We had one run where the connectors were “within spec” on paper... until we noticed the torque wrench in the kit wasn't calibrated. It was reading about 30% high. On paper, everything was “tight.” In reality, three of the six connections were loose enough to flex during thermal cycling, letting moisture creep in. The return loss dropped by 8 dB over four weeks, and the sector started generating false handover alarms. I still kick myself for not catching the wrench issue on the first audit.

You Can't Troubleshoot What You Never Documented

Even when a cable is tested, the results often live in a notebook or a half-remembered conversation. If there's no baseline, then six months from now when the return loss jumps, you don't know whether that's a new failure or a condition that was there since day one.

You can't compare, you can't prove, and you can't claim warranty credit from a vendor. You just have a problem site and a shrug.

The Real Cost of Skipping Acceptance Testing

Let me put some numbers on this. In 2023, we had a sector fail three times in six months. “Intermittent high VSWR.” The crew went out, re-seated the connectors, ran a quick sweep, and it passed. A few weeks later, it failed again.

After the third truck roll, we finally did a distance-to-fault (DTF) measurement with a Site Master S331P and found the problem in about ten minutes. A corroded pin inside a connector, exposed to moisture through a tiny breach in the boot. The fix was a $20 replacement connector. The three truck rolls cost north of $9,000, plus the downtime for a site that was supposed to be covering a busy intersection. The math here is honestly embarrassing.

That's the real price of “good enough.” Not the test equipment. Not the 30 minutes it takes to sweep a run properly. The $9,000 in repeat visits. The $22,000 rework on a larger install with a damaged feeder. The customer complaints that you can never directly trace to a marginal cable but that hang over your head anyway. The cost of not measuring is always higher than the cost of measuring.

How I Test a Cable Before Calling It Ready

Here's the sequence I've settled on after years of doing this. It catches the failures that actually happen in the field, and it doesn't waste anyone's time.

1. Visual inspection first. Check the run for tight bends, crushed sections, damaged connectors, pulled boots. If it looks wrong, no sweep will fix it.
2. Verify the connector mechanically. Torque with a calibrated wrench—and check that the wrench itself is calibrated. Check center pin depth and socket contact. In my experience, this alone catches about half of all acceptance failures.
3. Calibrate the test setup properly. This is not the step to rush. Use quality reference loads—the Anritsu 22N50 precision termination is my go-to—and do a proper open/short/load calibration before the measurement. A poor calibration gives you a confident false reading, which is worse than no reading at all.
4. Run a DTF measurement. The distance-to-fault function on the Anritsu S331P is the closest thing to X-ray vision we have in this industry. It shows you where the fault is, not just that it exists, which turns a head-scratching investigation into a straightforward fix.
5. Sweep the full operating band. Check VSWR and return loss across the actual frequency range the system will use. Don't test at 900 MHz if the radio runs at 3.5 GHz. Common industry acceptance for new feeders is return loss of 20 dB or better (about 1.22:1 VSWR), but the system requirement should always win.
6. Verify insertion loss. The signal has to arrive with enough power. A long cable run with high loss will strand customers even if the VSWR is perfect.
7. Document on-site, with a baseline. I type my notes and capture measurements on a DuraForce Pro 3, which has survived more rain, drops, and dust than any field tablet I've used. The point isn't the device—it's that documentation happens at the site, at the moment of measurement, not reconstructed from memory later. GPS-tag the results. Photograph the test setup. Store the DTF trace.

Seven steps. Thirty to forty minutes per path, once you've done it a few times. And now the question “when was this cable ready for service?” has an actual answer—a date, a signature, and a file with data behind it.

The fundamentals of RF transmission haven't changed much in this industry, but the expectations have. A decade ago, a quick continuity check might have passed. In 2025, with spectrum getting more complex and networks carrying more critical traffic, “probably fine” doesn't cut it anymore.

Bottom Line

If you can't answer the question “when was this cable ready for service?” with a documented test report, then the honest answer is: it isn't ready yet. It's just connected.

The good news is that the fix is straightforward. Get the right test equipment, use the right calibration reference—whether that's the Anritsu 22N50 or something equivalent—sweep the full operating band, and record the baseline. The tools have been around for years. The discipline is the missing piece.

The next time you're standing on a roof with a connector in your hand and the light is getting low, ask yourself: will I be able to prove that this cable was ready for service when it went live? If the answer is no, you're not done yet.