Fi

Despite the fact that I had to drive down to Bastrop on Friday to pick up a SIM card because none of the Best Buys here in Austin had them, the transition from Cricket to Google Fi on Saturday morning was rather painless. All told, took about 10 minutes to swap the SIM, activate the Google Fi app on my Nexus 5X phone, and 20 minutes to completely transition my phone number.

Screenshots of phone before transition from Cricket (at 11:52am) to Google Fi (at 12:05pm)

My Cricket subscription is now automatically cancelled, just a scant 3 days before the next month’s payment was due. It’s been a good run, but Cricket didn’t offer a tethering plan or international roaming for my bring-your-own-phone device. They refused to flip that tethering switch, saying “We don’t support that phone model.” Yet their parent company, AT&T, would’ve had no trouble doing the same if I was their direct customer. I signed up with Cricket over 12 years ago because I “didn’t want to be locked into a contract”, and now I’ve finally left, as if I had been in contract the whole time. I fear change, I guess.

Domestically, Google Fi leases service from a handful of mobile carriers (many mobile brands do this). The SIM contains the base minimum to connect to the T-Mobile network. The Fi app has an option to “Tune Your SIM Card for Best Coverage”, meaning it needs to consult Google’s systems to get the provisioning info for the other carriers so it can write them to the SIM. So sure, fine, yeah, I can be offline for a few minutes while it scans the mobile bands to test connectivity with each carrier. Luckily I had nobody calling me.

The marketing literature says Fi’s data plan has built-in VPN. The fine print is that it’s only activated if you turn on “Automatically connect to high-quality public WiFi networks”, in which case it will tunnel your data plan through a VPN on whatever public WiFi your phone touches. This means the Google Fi app pulls down a database of the SSIDs and passwords of every WiFi hotspot Google’s learned about. That’s kinda troublesome from a trust standpoint, because you’re at the mercy of whatever crypto is used in this VPN, but that’s the nature of the beast if you’re dealing with Google.

Also, if you’re on WiFi, the plan defaults to WiFi calling. The call quality is fine as long as the WiFi isn’t saturated at whatever hotel, mall, arena, airport, or cafe you’re visiting, otherwise you’ll start buffering and having drops in your conversation. And if you leave that WiFi hotspot, prepare to have your audio disrupted for a few seconds while the phone transitions to GSM calling. So heads up.

Otherwise, the plan costs are about on par with Cricket (based on my usual monthly data usage), and they offer international roaming, which will prove essential here shortly.

More news later if something blows up.

Magnetic California

It’s not going to win any beauty contests, but with my slapped-together Small Magnetic Loop Antenna (small MLA), I just worked two stations in the San Francisco Bay area with 20 Watts on FT8! (11pm CDT, 20-meters, antenna in living room.) By comparison, it takes me over 40 Watts to do the same on my 20m dipoles hanging outside. Huh.

Also, I heard a station in Hawaii and another further out in the Pacific, though they couldn’t hear me to log a contact. That’s a first for me.

Small Magnetic Loop Antenna, totally slapped together. Tripod, PVC pipe, non-conductive broom handle, pipe clamps, Velcro straps, and zip-ties. Primary (gamma match) loop: 1ft diameter, squished for SWR adjustment; RG57/U. Secondary (resonant) loop: 4ft diameter, RG8.
20~145pF air variable capacitor at top. Knob tunes the secondary loop for the target frequency (sweep knob until static on radio sharply increases in volume). Braid and center conductor of RG6 cable soldered together and clamped.
Primary loop (gamma match); RG-59/U, 1ft diameter. Braid and center conductor soldered together, then soldered across SO-239.
Close-up of bottom construction of gamma match.

The antenna matches with low-enough SWR on 20m (still have to use a tuner), and the capacitor isn’t high enough in value to get a good match on the low end of 40m (my band of interest), but it works better than expected. It’s a difficult, complex thing to fidget with, and is very fickle and easy to disturb because everything’s floppy. But…it works.

A nice feature of small MLAs (“small” meaning the outer loop is less than 1/10 wavelength of your intended frequency) is that they are selective and have strong nulls perpendicular to the loop plane. This allows you to rotate it to find and cancel nearby sources of radio noise. I’ve identified a few noises around the apartment complex, and want to take the rig outside (after some rework) to see if I can do better location of the noises.

I’m glad I put so little work into it; the cost of failure is lower this way. I was frustrated with it Sunday night when I built it, but tonight, even after sunset (with California in gray-zone), I managed to log a few contacts.

Research and more reworks are in the future. Potential improvements:

  • better construction with SO-239/PL259 connectors on weatherized capacitor box
  • use RG8/U (50ohm) for primary instead of RG6 (75ohm)
  • make primary closer to 1/5 size of secondary as recommended
  • use higher value capacitor (200pF?) with free 360° motion
  • design a remote tuning gearbox.

So let me say this: MLAs are interesting.