Who This Checklist Is For
If you’re deploying EXFO tools—like an OTDR, a DWDM channel checker, or the new Duraforce Pro 3—and you’re responsible for the final handoff to a customer, this list is for you. I put this together after reviewing 200+ unique test deliverables each year in our Q1 2024 audit. I found roughly 18% of first-time submissions failed due to preventable mistakes. Here are the four steps that cut that failure rate to under 3%.
Step 1: Verify Your EXFO Setup Against the Project Spec
This sounds basic, but it’s the most common problem I see. I can’t tell you how many times I’ve seen an EXFO testing report that used the wrong wavelength range or pulse width for the application.
Here’s a quick checklist:
- Confirm the test module matches the fiber type – Are you testing SMF or MMF? Using a module designed for single-mode (like the G310 5G module) on a multimode link will give you totally useless data.
- Set the correct measurement range – For example, a standard FTTH drop is <5 km, while a long-haul trunk can be 150 km. Don’t let the OTDR auto-range if the max distance is known.
- Check the launch cable condition – A scratched launch cable barrel will introduce an insertion loss error that propagates through the entire trace.
One thing vendors won’t tell you: The factory calibration of an EXFO unit is NIST-traceable, but that doesn’t mean a field technician can’t accidentally change a setting that invalidates the measurement. I once rejected a batch of 50 DWDM channel checker reports because the technician had left the reference-setting mode on “relative” instead of “absolute.” Cost us a $4,000 delay.
Step 2: Always Run a Known-Good Reference Trace
Before you even start your measurement, record a reference trace on a known-good fiber patch cord. This is your baseline.
In my first year, I made a classic rookie mistake: I didn’t run a reference, assuming the EXFO unit was self-calibrating. I learned the hard way when our customer rejected a full 48-fiber trunk characterization because the OTDR trace showed a 0.2 dB splice loss that we later found was actually a dirty connector on the test port. The redo cost us a Saturday shift and $2,200 in overtime.
So here’s what you do:
- Clean every connector – Use a fiber microscope (like the EXFO FIP-400) before each reference trace. A speck of dust can cause an extra 0.5 dB loss.
- Save the trace to a specific folder – I use a folder structure like “YYYY-MM-DD_Project_Ref”. That way you can always go back and compare.
Step 3: Document Every Measurement Anomaly in Real-Time
This is the step most people rush through, and it’s the one that matters most when you’re doing complex troubleshooting, like when you’re comparing switches vs Cisco on a live enterprise backbone.
By “anomaly,” I mean anything that doesn’t match the reference trace or the theoretical model:
- Reflective events that are higher or lower than expected
- Non-reflective losses (like a bad splice or macro-bend)
- Excessive distance between events
Here’s something that changed my perspective: I used to think documenting anomalies was about being thorough for the report. But I found it’s actually about saving time later. When you’re on site with an EXFO and a Duraforce Pro 3, you’re the one holding the test set. If you don’t record why you flagged something, you have to re-test to figure it out next week.
I now follow a simple rule: for every anomalous event, I write down:
- The event number (from the OTDR)
- The measured loss in dB
- The suspected cause (e.g., “dirty connector on panel #7”, “tight bend near splice tray”)
- A picture of the connector (taken with the microscope) or a note if no photo available
Step 4: Validate Your Results with a Second Measurement
This is where the “quality inspector” in me kicks in. I don’t accept a single measurement as final. I ask for a second measurement from the other end of the link.
Why? Because OTDR measurements are direction-dependent. A splice that looks clean from one end can look terrible from the other end if the splicing machine had a bad cleaver. You’ll catch this by comparing the two traces.
When I implemented this step at our shop in 2022, our defect rate for final deliverables dropped from 18% to 5% almost overnight. On a 50,000-unit annual order (I’m talking about individual test reports here), that’s 6,500 fewer re-dos.
Common Mistakes to Avoid
1. Not cleaning the connector after the first test.
If you test a dirty connector, your OTDR trace will show a spike that looks like a reflective event. It’s one of the most common false positives. Always clean between uses.
2. Using the “average” setting on the EXFO unit without understanding it.
The averaging function reduces noise on long fibers, which is great. But it also reduces event resolution. If you’re looking for a 0.1 dB splice loss in a network that’s supposed to be pristine, averaging might smooth it away. Use it only when you’re measuring long spans.
3. Forgetting to set the proper measurement threshold.
I see this one constantly. If your EXFO unit has a threshold setting for event detection, and you leave it at the default (say, 0.5 dB), you’ll miss any event below 0.5 dB. That can mean your final report shows a clean link when it actually has a high-loss splice.
I set my threshold to 0.05 dB for all active checks. It might flag a few more events (like connector reflections), but I’d rather have a tiny flag than a missed problem.
4. Not keeping a test log.
I said “write down anomalies” earlier. But what I really mean is keep a complete test log. Date, time, project, module serial number, reference trace filename, measurement trace filename. Do this as a habit, and you’ll have an audit trail that will save you in a dispute with a customer or a vendor.
Bottom Line
A reliable EXFO testing result isn’t just about having the right tool—it’s about following a strict process. These four steps have saved my team thousands of dollars and a lot of headaches. Start with the reference trace, document anomalies in real-time, and never trust a single measurement. If you do, you’ll deliver reports that your customers (and your quality inspector) will actually trust.
