Monday, May 13, 2024

Wind Instruments for Rotating Masts



I sailed my Farrier trimaran for almost 20 years before I decided I really needed a masthead wind instrument. Telltales and a Windex work fine but I wanted to accomplish two things: first, learn to sail m boat better, and second, learn what were my optimum tacking/jibing apparent wind angles (AWA) to achieve the best VMG to windward. After studying a lot of solutions online, and learning from others' experience (particularly on the F-boats email list and Sailing Anarchy forums), I found solutions that should work for almost any small boat, and this post is an attempt to summarize them.

The problem: stock wind instruments cannot compensate for mast rotation. The manufacturers assume the instrument will always be pointed in the same direction as the boat is headed. On a small trimaran with a rotating mast, this is usually not the case.

There are two ways to correct wind readings from the masthead: using compasses and using a rotation sensor. I have tried both on my boat. This post will explain how to achieve each, depending on your configuration and goals. The idea is that your masthead sensor detects apparent wind relative to the direction the mast is pointing, so you have to figure out the difference between the direction the mast is pointing and the direction the boat is pointing and add or subtract that from the AWA measured from the masthead, then re-inject that data on your NMEA2000 network for your displays. Note that apparent wind speed (AWS) is directionless and unaffected by rotation. Also, true wind angle and speed (TWA and TWS) must be re-calculated as well, using the corrected data.

Here are some of the questions you need to answer to determine the best solution for you:

1. What kind/manufacturer of instruments and displays do you use?
2. Do you already have instruments installed on your boat?
3. Is your goal to learn how to sail your boat better, or do you want to develop polars so you can predict your speed in various conditions?

Here's a summary of options and solutions:

1. If you have B&G instruments, and your instruments already include an electronic compass like the Precision 9, then your simplest option is to purchase another compass like the Garmin Steadycast, mount it on the mast, and use a Shipmodul Miniplex to correct the masthead readings and re-transmit. I won't go into detail on this solution because the people at Shipmodul are willing to assist in configuration, and my understanding is that they are very helpful.

2. If you don't have B&G, or if your B&G does not include an electronic compass, you can still do the same by adding two compasses. As of this writing, the Precision 9 sells for around $700, the Steadycast for $160, and the Shipmodul for around $400, so it's not the least expensive option, nor is it the most accurate. The B&G claims +/- 2 degree accuracy and the Garmin +/- 3 degrees, so in theory your reading could be off by 5 degrees. Under most conditions this probably won't occur, but B&G recommend calibrating your compass every time you sail. 5 degrees of error is probably fine for club racing while you're watching your AWA and VMG, but it's probably not as effective for calculating polars.

3. If you have B&G (no compass), Raymarine, or Garmin, you can install a Honeywell position sensor at the base of the mast and correct for mast rotation using free software running on a device like a Sailor HAT ESP32. The Sailor HAT is a very small, low power computer. It's approximately as powerful as my mid-1990s IBM Thinkpad, but its primary use more generally is in embedded systems. This is the most accurate method of rotation correction currently available that I'm aware of. The Honeywell sensor usually sells for around $500, and the SH-ESP32 complete setup is less than $100. I'm happy to supply systems and code for anyone who doesn't want to take on this project themselves. The Honeywell sensor has an accuracy of 0.06 degrees.

4. Another option is an ultrasonic sensor from French company LCJ Capteurs. They can provide a complete system with masthead wind sensor and mast angle correction device. This currently sells for $1450 USD, but I'm not sure if that includes the Honeywell rotation sensor. If I were starting from scratch I would definitely buy their masthead sensor; it claims much higher accuracy than conventional sensors and I don't think it has any moving parts.

4. If you have Garmin displays, the solution is more complex. Garmin displays have no ability to select a data source. (Strictly speaking they do, but will always revert to a Garmin source if one is available). So if your Garmin/Nexus masthead wind instrument is transmitting non-corrected AWA on your N2K network, and another device is transmitting corrected AWA, the Garmin displays will ignore the correct device, even if you've configured it as the preferred source. This means that Garmin installations need to isolate the wind instrument (which is really the GND10 that translates from Nexus protocol to N2K) on a separate network, and use a device to both correct the readings and transmit corrected data on the primary N2K network. This is the system I have on my boat, and it works well.

My code is available on Github and I can provide hardware with the software installed. My system currently has multiple functions. In addition to mast angle correction with the Honeywell sensor, I'm also measuring mast angle with two electronic compasses (one I built myself, and one using Pypilot software on a Raspberry Pi). The system hasn't been running long enough for me to determine if the compass solution is sufficiently accurate for polars...more data required. My device is also translating NMEA0183 GPS position from my ICOM VHF radio, which gives me VMG to waypoint, etc, and it's also transmitting commands in the old Seatalk format to my Autohelm ST2000+ tiller pilot, courtesy of code from Nauti-Control. Finally, it serves a small web site for configuration and data display, from which I've included some screenshots below. The device could also easily transmit data to a tablet-based chartplotter (I use Navionics), but since I had a working system for that already, I decided not to change it, although it can be enabled as a backup if the primary AIS bridge to the network fails.

I'm using SignalK on the Raspberry Pi to log data (to an InfluxDB database), as well as to observe all kinds of interesting properties of my boat while I'm sailing. SignalK can calculate VMG to wind, VMG over ground, set and drift of current, leeway angle...the list goes on and on. SignalK is free, and supported by an amazing community online. It works well with a free chartplotter called Openplotter.

The web interface to the ESP32 shows a lot of data, but it's not important for sailing. What's important is what you can't see: the N2K sentences that are transmitted on the boat's network to the displays, with corrected AWA, TWA, and TWS.

Calibration: