It's alive!!!

 "It" being our Iridium GO, a little satellite communication gizmo that we can use to download weather reports and wind maps ("GRIBs") while far offshore (at 2400 baud - really!). We've been trying, without much success, to get it working ever since the boat was launched back in June. It turns out that Garcia used a totally inappropriate kind of coaxial cable (RG-223 instead of LMR-400) between the external antenna (mounted on the arch at the stern of the boat) and the transceiver itself (mounted at the nav station). The result was a nearly complete loss of signal. 

 (Hard-core geeks can take a look at the Times Microwave cable calculator to see why.  The Iridium up/downlink frequencies for "end user" stations are at about 1625 MHz. Pick RG-223 and LMR-400 and look at the cable efficiencies (or, if you are REALLY geeky, at the dB losses.))

I've ordered a length of the correct cable (LMR-400-UF, a more flexible variant of LMR-400) and the necessary connectors (male TNC), and hopefully by the end of this week (or maybe the next...) we will have a working system.

Meanwhile, to test the software setup, I just put the unit itself (which has mediocre built-in antenna) out in the cockpit and connected to it with WiFi from my iPad. Amazingly, it works!

Downloading the GRIB files

And - the graphical wind display!

 

This is very good news, because it means that even if we have problems with the external antenna, we should be able to get weather information while far offshore (such as when sailing between Europe and the Caribbean).

Further update (March 19, 2021)

The saga of our Iridium GO continued for about another ten days. 

Katy and I carefully measured the actual distance from the antenna to the location we were thinking of using for the GO (above the door to the aft head). When you account for all the snaking around (up and down and through the arch, the aft compartment, up and down bulkheads, etc.), the cable run would have been at least 15 meters. The absolute max length for LMR-400-UF (a very high quality, purpose-appropriate cable) would be 12.8 meters (to keep the loss to no more than 3dB, the value Iridium specifies).

So - we had to find another location. We settled on the bulkhead over the access hatch to the aft watertight compartment. By routing the cable down through the hull in the space just forward of the starboard helm, the cable run could be kept to 10 meters. We discussed this with Vincent Mauger of Grand Large Services, and he concurred that this should work, and would not adversely affect the integrity of the aft watertight compartment. Power for the GO ( all 8 watts...) could be gotten from one of the SFSP Scheiber blocks in the engine compartment.

After several days of diligent work by GLS, it's all working.

Five bars!

 The takeaways from this saga are:

• The cable loss specification must be respected. Use LMR-400-UF or an equivalent, and keep the cable run as short as possible, and definitely not more than 12 meters.
• The short adapter cable that patches between the TNC connector on the LMR-400-UF and the crappy little TS-9 connector on the GO is fragile. Our first problem with the GO was a broken adapter cable. The mounting plate that Iridium supplies with the external antenna kit subjects the cable to very small radius bends, and it's easy to pinch the cable under the plate. Hence, we routed the cable entirely outside the plate (see above).
• Don't discount the possibility that the GO itself has a problem. Our original GO seemed to have a problem using an external antenna at all; perhaps the TS-9 connector failed internally? In any case, Grand Large Services replaced the unit without any questions. Bravo!
• Even with everything working well, the Iridium service is slow. The claimed data rate is 2400 baud (not very fast), but it's actually much worse than that. Even with 5 bars, the connection is unstable, and the most data I've received in one pop is about 30Kbytes before re-connection (which takes at least 30 seconds) Luckily, the transfer protocol used by PredictWind accommodates restarting, so the data already received is not lost.