- Introduction
- Goals for the Project
- InqWind – Sensing the Wind – Electronic design and experimenting.
- InqWind – The No-Brainer Sensor – AS5600 – Finding a better sensor for an Anemometer.
- InqWind – Salt and Sweating the Details – First cut for the speed sensor.
- InqWind – Plastic Bearings and AS5600 Experiments – Evaluating the test Mule.
- TBD InqWind – Through the Grinder – Rudimentary software described and testing.
- TBD – The Final Solution
Introduction
I have a small sailboat. I do not have a yacht. Living far from coasts, I’ve mostly sailed on lakes with an occasional outing to a bay or to parts of the Intercostal waterway. Even with these glorious jaunts, I’ve been with groups of other sailors on perfect sunny days. In other words, I’ve never really needed any expensive navigation equipment. Give me a simple paper chart, printed for basically free off a NOAA or Corp of Engineers website add a few feet of old VHS video tape tied to my shrouds and masthead and I’m good to go! I was once given a $30 hand anemometer as a Christmas gift and although pretty cool, I quickly discovered quantifying the wind speed really didn’t add to my sailing prowess. I also discover how hard it is to find wind direction and speed with this thing when nestled at the helm behind these big, cloth wind diverters. I must have looked pretty silly licking my finger and sticking it up in the air!
Now, I’d like to do some trips and navigation becomes a real necessity. Doing a perusal of equipment and its really quite easy to spend $10K on a chart plotter system. Even getting something to tell you wind speed and true direction can easily set you back a Boat Buck. I recently found OpenMarine and watched a slew of YouTubes on its power and flexibility. Apparently is has gotten so good that even people with million dollar yachts are using it in preference to established proprietary systems. Using, commodity electronics (Raspberry Pi’s and Arduino micro controllers) a system can be put together for next to nothing. Well… at least for someone that plays with this stuff in their free time. In my case, the hardware is basically free, gathering dust in part bins. One can simply install the software on an SD card, load the operating system, hook up to the Internet and download free charts for many of the world’s coastlines and you’re on the way.
With the selfless devotion of another sailor/software developer, there is even an arguably better auto pilot than the big players in the PyPilot project. However… when I turned my attention to simply getting the wind speed and direction into this system, is when the sticker shock really hit. Sure… there are tons of commodity and DIY Internet anemometers but most are geared for weather and are only good for N or NE or at best NNE type directions. As soon as you want anything with resolution better than 15°, the prices soar. Even on the OpenMarine forum, I found only a handful of initiatives, but none gave a completed system. I did get a great tip on the forum for the open-boat-projects.org that holds some great promise on a couple of DIY anemometers. You might try one of those for your own DIY project.
Now… InqOnThat.com is more about DIY and building things with electronics and software. Sailing has yet to be mentioned anywhere else on the site. For me… with my bent, I want to build my own from my own designs. I don’t think anything is necessarily wrong with those other DIY anemometers, I just like to think I can come up with a better and cheaper mouse trap.
This is being written before I have anything really working. I’ve only done a few tests to convince myself that I may be able to pull this off. There is much to be tried, researched, studied and tested before a final solution. If you also have similar interests in the building of one, you might find my musings of value. If you are just interested in the final solution, well… you’ll have to wait a while. This series of posts will be about this journey. I will keep a contents linking the posts and in the eventual, final solution, I’ll have links to all the hardware, software and build instructions. Along the way, I’ll also include details about the hardware and software in case this project is stillborn or if you want to branch off differently than my progression.
Goals for the Project
This is the initial wish list. It will likely change as reality sets in, but we have to start with a goal. We want some apple pie and eat it too.
- Unit Price $20 – I put it this way, because I typically buy in lots of 5 or 10 to get a better price and to account for my tendency of letting the smoke out of parts. If someone else wants to make only one, it will cost more.
- 3D Printed – The price will include the plastic used, but you’d have to have this pretty expensive piece of equipment just to achieve it.
- OpenMarine Integration – Of course it has to work with the Open Source system to show nice virtual gauges and to facilitate PyPilot’s navigation skills.
- WiFi Communications – Connecting to the OpenMarine system via WiFi to eliminate at least one set of wires.
- Low Power Consumption – By this I mean averaging in the milli-watt range. I’ll take advantage of the microcontroller’s sleep modes to reduce power consumption greatly. This presents some challenges since communications are typically rather slow to establish and the most common methods of measuring wind speed uses counters that have to be powered continuously. But I have a few ideas worth exploring.
- Small Angle Resolution – Not being a sailor used to using this type of equipment, I don’t quite understand why I need better than 1° resolution, but I see the $1K units touting smaller resolutions like 0.1°. I’ll shoot for this.
- Small Speed Resolution – Again, not seeing the point, but I’d like to shoot for sub knot speed resolution. This is challenging because of the power goals being closer to the top of the list and because of the non-linearity of typical sensing methods.
- High Send Rate – The weather stations only do about one sample per 15 seconds. Even I can understand this is near worthless. I almost forgot to put this item in. This one is easy. I think the user will decide on a slower send rate that can be accomplished. Does anyone really want 100 samples a second flying around their network. But hey… maybe a racer does. No sense restricting it if it’s technically possible.
- De-bouncing Resolution – I’ve seen units touting tiny resolutions, but if it’s bouncing around +/- a couple of degrees or couple of knots, I hardly see the point. I’ll explore some smoothing solutions. The challenge here is to not introduce too much latency.
- Weather Proofing – Not exactly sure what this implies yet. The electronics can be pretty easily sealed away, but somewhere there needs to be free spinning bearings. Putting seals and gaskets that can tolerate years of sea spray would add friction that can’t be accounted for. Does it need to survive 40 knots… 70 knots… hurricanes? For my usage… no. I won’t be doing IP ratings, as I have no intentions of “going into production“, but I will attempt to use logical solutions for keeping the elements out and I will test it for at least highway speeds in rain. This might be a destructive test.
Some Baselines
There are all kinds of methods for tackling this project and many people have strong opinions about their own product that makes it clearly superior. This is not meant to be an exhaustive search for the best solution. It is simply using my assumptions, what equipment I have on-hand and what I’ve experimented with in the past. Case in point, I see a lot of modern work being done with ultrasonic wind sensors. Most appear to work on the principle of Doppler effects, but I have seen a couple using the wind running through a whistle or microphone to, I guess, measure frequency and/or amplitude variations and mapping that to a speed and direction. These all seemed really cool to experiment with because there were no moving parts to corrode or lose lubrication over time. But to me, they seemed like they were only one errant bug away from being knocked out of commission and a tiny inlet can get just as salted up as a bearing. Besides, I attempted a few experiments with commodity level ultrasonic sensors and found that they couldn’t handle really slow wind speeds that I’m cursed with.
To wrap this up, I will be exploring the use of Hall-effect sensors for both direction and speed and the journey begins here.
Inq