Barrel connectors for coax cables: Does the rating make a difference?

From left to right: Cheap zinc alloy F-81, 3-GHz F-81, 4-GHz F-81, & F-61

After spending the day recrimping cable runs and replacing wall plates at a friend's house, I started wondering about my own setup. My coaxial wall plates were mostly 'builders grade' from Leviton, bought by the contractor to match the wall color. How much insertion loss (a bad thing) could I lose from them, and how much could I improve the return loss (a good thing)?

PPC loss testing

PPC claims that the F connectors in typical home networks are the most common failure point for DOCSIS signaling issues, with loss of continuity in the connectors causing RF ingress/egress, reflectivity, and impedance mismatches. What you might notice when unscrewing a wall plate is that an older RG-6 or RG-59 cable might feel very sticky, as grease oozes out between the outer jacket and shielding braids, making the crimped connector come loose. It's also possible that wear and tear to the wall plate's connector threadings are causing signal loss. Checking for continuity and keeping connectors tight is a great idea, but is just part of the picture.

Nowadays, you can find F-81 (F female to F female) barrel connectors, sold either loose or in keystone inserts, that are rated for various frequencies. But do these mean anything, other than the maximum frequency the manufacturer used when testing them? I figured I'd find out.

Inside an F-81 barrel connector

Sales pages for 3-GHz F-81 barrels often have a few specs in common. These barrels (sometimes with a blue dielectric, but not always) tend to be made of brass with nickel or gold plating, with a seizing pin of beryllium copper with silver plating, and polyethylene dielectric. The seizing pin should make contact with the cable's conductor in 2-4 points, which is why it's important that crimped connectors have the right amount of conductor sticking out. Cheaper barrels, rated for 1 GHz or (more likely) not bearing any published rating, may be made from poorly machined zinc alloy. The models sold by Show Me Cables have the following characteristics:

So while insertion loss is constant at different frequencies, the return loss (which lowers insertion loss) is constant.

To put some barrels through their paces, I connected 5 F-81 barrels to 5 F-61 (F male to F male) barrels between my Wi-Fi router's impedance converter, a 4-way splitter, and a Coaxifi wall plate antenna, figuring that quantity would surely be enough to create demonstrable insertion loss. Measuring Wi-Fi RSSI in dBm at a remote room, I took samples with my status quo connection as a baseline, a connection comprising 4-GHz barrels (green dielectric), another with 3-GHz barrels (blue dielectric), and another with 1-GHz barrels (clear dielectric). The results below are pretty interesting.

It turns out that the difference between 4-GHz and 3-GHz barrels is minute, amounting to 1 dB in additional insertion loss or presumably 0.20 dB per barrel. They may even be identical for insertion loss, if I were to test them dozens of times. But the difference between 4-GHz and 1-GHz barrels is substantial, amounting to 18.1 dB in additional insertion loss or 3.62 dB per barrel. But why?

Typical unrated barrel

There are several visible differences between cheap barrels and frequency-rated barrels. The cheap ones have zinc alloy housings, with tapered threadings around the dielectric, and usually a thinner nut in the middle to boot. The conductivity may be poorer than with brass, the continuity may be weaker since the dielectric between the cable connector and barrel may be farther, and it's going to be harder to hold the barrel in place if tightening connectors by hand.

Is that enough of a difference to replace all your wall plates? Maybe not. And to dispel a common misconception, the rating of a coaxial cable or a barrel does not represent any sort of 'cut off' frequency for signals. But DOCSIS 3.1, MoCA 2.x, Wi-Fi over coax, and other applications are all pushing the envelope when it comes to frequencies passed in a home coax network. The more splitters, barrels and connectors the signal has to traverse, the weaker it'll be, especially if the splice on a barrel has poor continuity. So if you're replacing old wall plates anyway, you may as well upgrade to 3-GHz or 4-GHz barrels.