My mind is alert and my body is jumpy. I can’t sleep. The boat slews on her side and I watch the apparent wind instrument rise. I listen to the hull resonate through the surf. Each time I feel a flicker of fear. Will the autopilot hold?

‘No pilot’ is the most horrifying sentence to a single-handed sailor. It marks an abrupt end to a competitive entry in a race and signifies the beginning of a long and torturous ride home. ‘No pilot’ was the prospect that I faced yesterday. With the potentially disastrous outcome of either a broach round up and/or crash gybe if the pilot gave up, I had struggled to sleep. I had become sensitive to every wave and gust and its impact on the boat. Not without warrant. Three or four times in the pitch black of night I had needed to spring up the companion way steps to dive for the tiller, as the autopilot ‘off course’ alarm rang out its warning tone. By morning my blood pressure must have reached an all-time high.

Taking the autopilot remote with me, I later climbed back beyond the bunk to the area beneath the cockpit, next to the fuel tank. There the two course computers are situated. I worked out that using the remote to go into ‘standby,’ I could in the shortest time possible, flick the switch. Flicking the switch engaged the other pilot and with it, I hoped my problems would be turned off. Unfortunately, the outcome was that I merely engaged new ones.

I have just returned to my computer. As I was writing this, I had to abandon it in a fit of panic. The demon ‘off course’ bared his teeth once again. The situation was salvaged in time, but by a quivering wreck of a skipper. My nerves will be shot to pieces after 1,500 miles more of this.

The autopilot is jacked up on its highest level of response. Yesterday, the control head was freezing up – a wind shift alarm would go off, but would not allow me to acknowledge or adjust course in response. In that moment of freeze, the response would jump back to the lowest response 1, which is far below the default factory setting of 5. Response level 1 is no way near high enough to respond to the slightest condition change. Slippage under load suggests a hardware malfunction and certainly the first pilot suffered a grinding bearing; was loud, over-heating and power hungry. But the second was running quietly and smoothly, inferring instead a glitch in the wiring; a course computer default or a problem with the NMEA information-feed.

Yesterday was alas a Sunday, but Geoff Sargent, former head of technical services at Raymarine kindly returned my call. We talked through the wiring. One possible cause we discussed was the control head unit in the cockpit could have water infiltration. I unplugged it and until a squall came through just now, was hopeful, although not convinced, that therein lay the problem. My current options are now more limited. As a mechanical solution, I could disembowel the drive unit and look for sheered pins and steal parts from the other pilot as replacements. On the electrical side, I could hard wire an NMEA feed from the Silva FDX server to the autopilot. In doing so, I would maintain my wind pilot feature but loose all other instruments on board; Wind, Speed, Depth and GPS.

Fortunately I have a chart, some pencils, a handheld GPS and batteries! For now, I will try running with the pilot on a compass course only. The sea is rough, with a 10-12ft swell. If I can avoid hove-to with the toolbox and two autopilots in pieces, I will. Abandoning the wind pilot function alone is far from ideal. Variations in wind are lost and sail trim must be catered for manually. With a spinnaker up, it would be an accident waiting to happen.