Sadly my man in Lymington got back to be on the pilot remote control and said that whilst they’d have a look it wasn’t really their thing. They’ve given me the details of a guy in Croatia, so I’ll try him. But it’s got me thinking a bit about how to keep the boat’s instruments and pilot working long into the future without having to engage in a very expensive wholesale replacement.
The problem is one of history, and arguably corporate ownership. B&G are a venerable British marine electronics firms who produced the first electronic yacht instruments, starting in the 1950s. One of the best histories of marine electronics I’ve read is here, written by Nick White, the author of the truly awesome Expedition Navigation package (my only gripe with expedition is that it forces me to own a Windows PC – I’ve owned Macs since the ’80s).
B&G produced the first modern network style system and at some point in the 80s or early 90’s adopted a CAN bus technology that enables multiple devices to exchange data across a single network cable run around the boat, which they called Fastnet. It was a closed proprietary solution which allowed you to add whatever (B&G) elements you needed and have them all interoperate and display any of the data at any of the displays. In its day it was revolutionary. The modern NMEA2000 (N2K) Industry standard communications system is actually also a CAN implementation with a different messaging standard (which is irritatingly a closed standard owned by an industry body, rather than being open and public – though extensively reverse engineered these days). N2K achieves the same interoperability but across manufacturers allowing you to mix and match equipment, at least to some degree.
Trouper’s H2000 system uses this old B&G technology to connect its three processing units to 12 displays, and five fundamental measurement transducers. The fundamental measurements are depth, speed through the water, rudder angle, wind speed and direction, heading, trim, heel (and associated accelerations). The main processor combines wind and boat speed data to produce a calibrated and corrected set of both true and apparent wind values (the difference between the wind you’d feel stationary in the water, and the wind you feel as you move over the water), together with a calibrated depth (ours is set to report depth under the keel – some hold near religious views that this should be set to depth of water). The performance processor computes the boat’s performance data based upon an uploaded model of the boat’s performance (effectively the expected boat speed for different wind strengths and angles) known as the boat’s ‘polars’ to allow you to monitor your performance. It makes sail trim much easier to set as the system removes the effect of every gust and shift, so you can see if a trim change was effective. The performance processor also provides a serial comms interface to a computer (and Expedition in my case) that allows the computer to add data to the network as well as read its contents. The processor and also provides an NMEA0183 Industry standard serial interface for connecting a GPS. Finally the autopilot controller connects the heading, orientation, and ruder sensors to the network but also controls a motor unit that can move the rudder. It uses some quite clever algorithms keep the boat’s course straight – using acceleration data to create corrective inputs before the boat’s heading has changed, just as a good helmsman does. It will also sail to both true and apparent wind angles, using wind data from the network, which can be really helpful on longer passages under sail, or sailing upwind. The pilot was the first of the generation of systems to use this gyro data, and heavily used in the Open 60s of the late 90’s and early 2000s.
The remote control that has sort of died – given that it only works with the battery cover off, it’s not exactly useful on deck – was built by B&G at Ellen McArthur’s request, and she was often pictured clutching one when doing her single handed round the world record on her trimaran. It’s really handy – not only does it do pilot remote control functions but it allows you to set what data appears on the mast displays and alter system calibration and configuration settings.
B&G ultimately ended up as part of Simrad, in turn part of Navico, which is part of Brunswick Corp, a US corporate behemoth. On the way it stepped away from the firm’s long history of incremental upgrade and backwards compatibility, and gave up on its UK manufacturing site. H3000 had replaced H2000 and everything interoperated happily using Fastnet and the displays fitted in the same holes in the boat’s deck. H5000 moved to the NMEA2000 (N2k) bus standard, supplemented by some ‘private’ proprietary messages for performance data, but dropped all interoperation with Fastnet devices and even eschewed direct connection of sensors to the processor. Even the visually similar displays required a new larger hole. If you wanted to move to H5000 you had to throw almost everything away and start again. Each of those dozen displays I mentioned is in the region of £1000.
I didn’t go that route and instead have bought up spares from ebay and now have a small stock of spare parts. Indeed everything except the pilot remote control… Many of B&G’s service team, who were based in Romsey in Hampshire, ended up at a new firm called Tinley Electronics in Lymington (my nice man). Tinley have grown into A&T Instruments and initially targeted mainly the superyacht market, but essentially people with large legacy B&G installations. They continue to repair B&G displays and processors though now also make their own displays and processor that support the Fastnet protocol and fit in the holes. Their stuff is not cheap, but it will allow me to keep Trouper’s instruments working.
After a while B&G realised the error of their ways and released the H5000 Fastnet interface. This is an irritating device as it is so close to being what everyone needs, but not quite, presumably in the (erroneous) belief that it might cannibalise sales of new kit). It is essentially a one way converter from Fastnet to N2k. You can now connect N2K displays, including the nice H5000 colour graphic display at Trouper’s helm (guess how I know that the cut out is different – I ended up buying a 3d printer to fix that problem…). The problem is it doesn’t translate all the data – none of the performance or autopilot data, for example – and it is one way. Except that hardware is not really one way as it allows changes to display brightness in both directions. So it could pass much more data from N2k to Fastnet. This would be ever so handy as it would be a clean way to integrate my modern N2K GPS with the Fastnet instruments, and it could allow you to connect N2K autopilot controllers, complete with remote controls. But it doesn’t.
A&T’s public literature is a little opaque about quite how much translation their processor does between N2K and Fastnet. It supports both but I’m guessing that given that they’ve reverse engineered and reimplemented B&G’s work there will be agreements in place, or at least a little caution, about what they say in public. As far as I’m aware it doesn’t translate pilot control messages, and at £7k ex VAT it would be a somewhat extreme fix to a broken remote even if it did.
Boats attract technologists and it should be no surprise that there is a large community of people building open source software solutions for sailing boats – from navigation applications through to autopilots and instrument systems. They have reimplemented N2k in software libraries that make it simple to use and implement on small hardware devices like Arduinos. There is a powerful Arduino with two CAN interfaces. There is some public documentation of the Fastnet protocol from a few people reverse engineering it. So I got to wondering: I’ve got all the bits to set up a test rig (all those spares), why not reimplement the Fastnet to N2k gateway with support for all the missing bits? It might be possible to get hold of some Fastnet documentation but if not reverse engineering it should be fairly doable, given what already exists. It won’t be a quick project, so I’ll get it touch with the guy in Croatia and see about sourcing a spare remote control. After I’d started drafting this a RemoteVision I’m trying to negotiate a sensible price for a spare on eBay. I’ll also have a conversation with B&G Support – if nothing else it will be worth checking that they aren’t willing to do a software update for the gateway…