Keysight N7776C and N7745C: the swept-wavelength pair, and what actually matters when you buy or run them

What the N7776C actually is

The N7776C is Keysight's top-line tunable laser source. It is the direct descendant of the 81600B and the 81606A, which is worth knowing because a lot of labs still run those, and the software ecosystem carries forward. Same photonic application suite, same measurement concepts, better hardware.

The headline number is sub-picometer repeatability in both static and swept operation. That sounds like marketing until you are actually doing DWDM channel work, where 100 GHz spacing is 0.8 nm and 50 GHz spacing is 0.4 nm. A laser that wanders 5 pm between sweeps introduces systematic error into every single measurement that uses it as a reference. Sub-picometer means the source stops being a source of error.

The tuning resolution is 0.1 pm, and the fine tuning happens without overshoot, automatically, by coordinating the motor and the piezoelectric actuator. On older tunable lasers you had a separate fine-tuning control and you learned to work around the overshoot. On the N7776C that is handled. If you have spent time babying an older TLS through a wavelength optimization procedure, this is the thing that will make you happy the first week.

Output power is over +16 dBm across 1270 to 1360 nm — the O-band. That power level matters because it is enough to allow external modulation in BER testing. If you are doing receiver optical subassembly (ROSA) work for Ethernet, verifying spectral responsivity and sensitivity, this is the instrument that has the headroom to do it without an amplifier in the path.

Sweep speed goes up to 200 nm/s, and the N7776C does two-way sweeps. That is not a spec you care about until you are running production. Then it is the whole ballgame, because sweep time is test time is cost per device.

What the N7745C actually is

The N7745C is an eight-channel optical multiport power meter. It is the successor to the N7745A, one of the most common optical power meters in the field. Eight InGaAs sensor ports covering 1250 to 1650 nm.

Dynamic range up to 65 dB in swept and logging applications without ranging. This is the specification that matters most and that people misread. Without ranging means the instrument does not have to stop and change gain to cover that range, which means it does not introduce a discontinuity in the middle of your sweep. If you need more than 65 dB, the N7745C does fast power range switching, and you stitch two sweeps together to get past 80 dB of dynamic range. That is how you measure a filter with a deep stop band without losing pass band detail.

Memory is 1 million samples per port, with an additional 1 M/port buffer for continuous logging. Data transfer is up to three times faster than the previous generation — production-relevant, not lab-curiosity relevant.

Polarization dependence is under ±0.01 dB typical. On a multiport meter doing swept measurements, PDL in the instrument itself is a measurement error you cannot calibrate out easily, so a low number here is worth paying for.

Why they belong together

Keysight designed these to work as a swept-wavelength system, and the specs only make sense in that context. The N7776C sweeps. The N7745C logs eight channels simultaneously while it sweeps. Together, with the N7786C polarization synthesizer if you need it, you get swept-wavelength insertion loss, polarization dependent loss, and dispersion measurements across eight devices at once.

That is the actual use case: multiplexers, PON splitters, wavelength selective switches, ROADMs — anything with multiple ports where you need to characterize every port across wavelength, and where doing it serially with a single-channel meter would take eight times as long.

The dynamic range spec on the meter and the SSE spec on the laser are the two halves of the same measurement problem. Low source spontaneous emission on the laser means the noise floor of your measurement is not set by the laser's own broadband emission. High dynamic range on the meter means you can see into the stop band. If you pair a low-SSE laser with a mediocre meter, or a great meter with a noisy source, you get the worse of the two, and you paid for a spec you cannot use.

I have seen people buy an N7776C and pair it with an old single-channel power meter because they already had one. It works. It also throws away most of what they paid for.

Where these break, and what to watch for

The C-generation instruments are new enough that I have less repair history on them than I do on the 81600B and the N7745A. But the failure modes on optical test gear are consistent across generations.

On the tunable laser

Wavelength accuracy drift is the thing that gets you, and it gets you silently. Nothing errors out. The laser just slowly stops being where it says it is, and every measurement downstream inherits the error. This is why the calibration interval matters and why you should actually respect it. If you are seeing systematic offsets in your insertion loss measurements that do not make physical sense, verify the laser against a wavemeter before you go tearing apart your DUT fixture.

Output power drift is the second one. The laser diode ages. On a laser rated for +16 dBm in the O-band, losing a couple of dB over years of use is normal and correctable in calibration, but if you are running near the edge of your power budget for external modulation, it will bite you before you notice it.

Mode hops on a laser that is supposed to be mode-hop-free mean the cavity or the temperature control needs service. That is a bench repair, not a settings problem, and no amount of firmware fiddling fixes it.

On the multiport meter

Detector responsivity drift is the equivalent problem. The InGaAs detectors age, and the wavelength-dependent correction that makes the meter accurate across 1250 to 1650 nm is only as good as the last calibration.

Watch the connector interfaces. Those N774-I quad adapters take physical abuse, and a worn or contaminated interface on one channel out of eight gives you one channel that reads wrong while seven read fine — a genuinely annoying failure to diagnose because your instinct is to blame the DUT.

Channel-to-channel consistency is the thing to verify. On an eight-channel meter, the whole value proposition is that you can trust all eight equally. Run a reference source into each port and confirm they agree before you trust the data.

Buying these used

Both instruments are current-generation Keysight, which means used units exist but the market is thin, and prices reflect that.

For the N7776C, get the option numbers. The N777-C family has wavelength range options and output power options, and a laser that covers the wrong band for your application is worthless to you regardless of how good the price looks. Confirm the wavelength range in writing before you buy.

For the N7745C, confirm the channel count — there are configurations with fewer than eight ports — and confirm the connector interfaces are included. Two N774-I quad adapters for a full eight-channel setup. They are not expensive relative to the meter, but they are easy to omit from a listing and annoying to source separately.

For both, ask when they were last calibrated and whether there is a certificate with measured data. A pass/fail sticker tells you almost nothing. The certificate with actual measured values at each test point tells you how far the instrument has drifted and whether it is close to the edge of its spec. We issue that certificate — see our calibration page for what is included and what it costs.

And ask about the software. The photonic application suite drives these instruments for spectral measurements, and if the seller does not know whether the license is transferable, find out before money changes hands.

If you are running an older 81600B or N7745A and wondering whether the upgrade to the C-generation is worth it, send us an email and tell us what you are measuring. Sometimes it is, sometimes your existing setup recalibrated will do the job for a fraction of the cost. We will tell you which one we think it is.

Frequently asked questions