Ultrasonic wind sensor [CV7]

[ChrisseH]

Originally Posted by ronstory

I pulled the manual and brochure for the DX900+. It looks to calculate the leeward angle directly, since it has a two axis speed sensor, one for longitudal speed and another for lateral.

Depth Frequency: 360kHz
Depth Range: 60m
Transverse Speed Range: ±6knots
Longitudinal Speed Range: ±60knots
Accuracy of Transverse and Longitudinal Speed:
±0.1 knots for speed under 10 knots
±1% for speed above 10 knots
Display Resolution: 0.01 knots
Operating Temperature Range: -15°C to 55°C (5°F to 131°F)

It also has a three-axis accelerometer and a three-axis rate gyroscope.

The brochure also claims it more accurate and faster than the calculated drift. Check out the graph on the right hand side.

Airmar

This is an interesting sensor, but the challenge here is that both the longitudinal and transverse flow is dependent on where the sensor is installed, on the three dimensional shape of the hull near the sensor, the heel angle and on how the boat is "slipping" trough the water. So as the direction of the flow is changing so does the curvature of the hull which changes the flow rate due to the Bernoulli effect.

Calculating the angle from two perpendicular flow readings is thus not the problem, but the challenge is how to calibrate the flow readings so that they correspond to the true speed of the boat in longitudinal and transverse directions. The speed forward can be calibrated by motoring a known distance in both directions, but how can we calibrate the transverse speed? By towing the boat sideways? At different heel angles?

[Jman]

Originally Posted by a64pilot

See, I think I could calculate drift with an E6B, and once drift is known, then you either add or subtract that to “true wind” to get actual wind?
I know I’m butchering terms, sorry. For example apparent wind to me is relative wind

With boat speed and heading compared to GPS course over ground and speed over ground, you can of course get what you are referring to as drift. In most cases in a boat (all the manuals for calibration talk about doing it in a body of water with no current, pretty hard or impossible to find that if you are not in a lake) the majority of this comes from current.

Actually leeway (boat moving sideways (away from the wind, or in super tech tech boats with asymmetrical boards it is possible to move towards the wind)) could only be done with a sensor in the water that measures sideways motion.

Racing computers use the design’s predicted leeway for a given apparent wind speed and angle.

Since the title of this forum begins with the word Cruiser, I think it is pretty safe to ignore leeway in True Wind computed displays. Nearly all sailing is done using apparent wind anyway.

[ronstory]

OK, the more I've dug through the user manual of the DX900+, the more interesting it becomes. While they diagrams for the best location based on various boats hull shape so a sailboat it's pretty much the standard "on the port side, low and just ahead of the forward keel edge". They have other diagrams for non-sailboat hulls as well.

Regarding calibration, siince it a Bluetooth device... there is an app for that. From the manual in marketing-speak:

http://airmar.com/uploads/InstallGuide/17-626-01.pdf

About the CAST™ App
Download the CAST app by Airmar to view, analyze, and share sensor data and evaluate your sailing performance. Simply install the CAST app on your iOS or Android device to view available NMEA data over a Bluetooth® or network connection. Create custom gauges and page views to see only the data that is important to you. With the CAST app you can perform speed, temperature, and attitude calibrations without extra converters or cables. In addition to firmware updates, the app logs all of the NMEA sensor data over a customized time period for analysis. Smart devices with Bluetooth connectivity can be used to view the data, creating a mobile powerhouse.

Now the really interesting thing is there are "attitude" calibrations, which leads to believe and the 3 axis sensors will provide heel angle.

I'm being digging through looking docs but can't fine the N2K PGNs it supports. That would shed some addition light on it's capabilities beyond the marketing stuff.

[ronstory]

Oops! Missed the obvious. I made the mistake os assuming this was an ultrasonic sensor like the vaporware DST900 that never made it market. The DX900+ is electromagnetic sensor.

My bad, but reading is a powerful thing....

[a64pilot]

With COG and heading leeway can be calculated.
Whether it’s due to wind or water current would as you say need an in water sensor, but does the cause matter?

In truth as a cruiser, if initial course is say due N, then taking up a heading that maintains that course is all that is relevant?
Beyond that from a cruising perspective it’s just an interesting discussion, I know I’m not going to spend $$ to measure leeway.

[ronstory]

Yes, agree... it's a spendy unit. With a 3 year old DST800 in a bronze thru-hull in our boat that works great, the are two more barriers to purchase.

1. My DST800 does everything I need
2. The bronze thru-hull is incompatible with the DX900+ so I would need a haulout to replace it with plastic or stainless.

The one other advantage not mention is there is no paddle to get clogged but I'm not sure if it worth the extra $700+ for DX900+. Now if I was a serious racer, perhaps. But I have 44 year old boat that really solid (heavy by today's standards for a 35 footer) that is fun to sail. I'll let the J-boat folks beat me to start of happy hour... just as long as I don't miss 'last call'.

[seandepagnier]

Originally Posted by ChrisseH

...but how can we calibrate the transverse speed? By towing the boat sideways? At different heel angles?

The problem gets even worse with waves. I don't think water speed sensors even going forward are always accurate for it.

What about towing a rope with some drag, and measuring the force it pulls as well as the angle?

Originally Posted by a64pilot

In truth as a cruiser, if initial course is say due N, then taking up a heading that maintains that course is all that is relevant?

Without knowing true wind it is difficult to compare past performance and to quickly tune the boat. It is impossible to generate an accurate polar diagram for weather routing. The autopilot can also steer to true wind.

Beyond that from a cruising perspective it’s just an interesting discussion, I know I’m not going to spend $$ to measure leeway.

I think I can do it for a total of less than $30 using 3 ultrasonic cleaning transducers ($8 each) a microprocessor, some driver circuits and amplifiers.

Would sure be interesting to compare the results to dragging a line and measuring it which is also possible using a microprocessor, might be even cheaper.

[ChrisseH]

Originally Posted by boat_alexandra

What about towing a rope with some drag, and measuring the force it pulls as well as the angle?.

This traditional method using a rope has been used to determine the leeway coefficient in the Pendrick formula. Using a rope is certainly a valid way to measure leeway, but this should then be done at different speeds and preferably also at different heel angles. It is quite clear that the rope method does not work well when the speed is very low. An additional challenge is to accurately measure the angle between the rope and the longitudinal mid-line from the boat.

If this method would be used simultaneously with recordings from the DX900+ we could afterwards calculate suitable correction factors. The fact that DX900+ also can measure heel is of course an advantage, but the calculations must (as far as I understand) at least initially be done externally. But maybe Airmar could develop a calibration sequence utilizing this method?

[ronstory]

While accurate measurement it a lofty goal, I've become more appreciative of the fuzzy logic approach of just more and less.

Since I've now got a Boat Unit worth of cool wind instrument I seldom use (but they are cool), I would be content just knowing if my leeward drift was more or less depending on the trim and attitude. Most of my sailing tweaks is to just make my trim setting better than before, but rarely perfect. But occasionally I do get lucky.

For the serious racer generating polars with 3+ digits of accuracy and detailed logs of the race condition for later review, absolute calibration is key.

However, just not for me ... but the math sure would be fun.

Of course, YMMV.

[Dockhead]

Originally Posted by a64pilot

With COG and heading leeway can be calculated.
Whether it’s due to wind or water current would as you say need an in water sensor, but does the cause matter?

In truth as a cruiser, if initial course is say due N, then taking up a heading that maintains that course is all that is relevant?
Beyond that from a cruising perspective it’s just an interesting discussion, I know I’m not going to spend $$ to measure leeway.

The cause matters, because leeway is a crucial indicator of sailing performance whereas drift over ground is totally irrelevant to sailing performance. These days with GPS telling us directly what our COG is, we no longer care about leeway for navigational purposes, but we still care a lot to know how well we are sailing upwind.

Perhaps with enough experience you don't care as much, because it's (perhaps) a direct function of speed and heel for a given boat, and so you just KNOW when you've got it in the groove. Without knowing leeway, I think eventually most keen sailors figure out that, putting aside sail trim, speed is almost everything when trying to get upwind efficiently (and most of the rest is sailing as flat as you can without reefing a headsail). If you do anything to compromise speed, leeway gets you and ruins your VMG to windward. I do have an instrument which tells VMG to windward which is really useful, but for the reason Sean pointed out earlier in the thread, its accuracy is seriously compromised by its assumption that Heading = CTW -- because of this, it doesn't know True Wind exactly -- it doesn't know exactly from where the wind is blowing. I sometimes set a waypoint at a location which is dead upwind according to the AVERAGE true wind direction, and read off VMG to that waypoint -- in some ways more useful than VMG to windward.

Obviously "Drift" -- Heading vs. COG -- is irrelevant to all of this. We sail in the interface between air and water; the ground is irrelevant to sailing performance and thus all ground-referenced data is useless for this particular purpose.

[Dockhead]

Originally Posted by ronstory

While accurate measurement it a lofty goal, I've become more appreciative of the fuzzy logic approach of just more and less.

Since I've now got a Boat Unit worth of cool wind instrument I seldom use (but they are cool), I would be content just knowing if my leeward drift was more or less depending on the trim and attitude. Most of my sailing tweaks is to just make my trim setting better than before, but rarely perfect. But occasionally I do get lucky.

For the serious racer generating polars with 3+ digits of accuracy and detailed logs of the race condition for later review, absolute calibration is key.

However, just not for me ... but the math sure would be fun.

Of course, YMMV.

I think this is an entirely valid approach.

Most of us understand that less speed and more heel = more leeway and ruined VMG to windward. Sail flat and fast and you get upwind; don't and you don't. Most of us can feel the groove, pressure on the tiller, rudder angle -- the boat talks to us. We don't need hard data all that much.

On top of that, few cruisers actually do much sailing hard on the wind anyway. "Gentlemen don't go . . . " etc.

So yes, it takes a certain nerdy bent to care all that much about this, probably. I guess I'm one of those nerds . . .

[Dockhead]

Originally Posted by ronstory

Yes, agree... it's a spendy unit. With a 3 year old DST800 in a bronze thru-hull in our boat that works great, the are two more barriers to purchase.

1. My DST800 does everything I need
2. The bronze thru-hull is incompatible with the DX900+ so I would need a haulout to replace it with plastic or stainless.

The one other advantage not mention is there is no paddle to get clogged but I'm not sure if it worth the extra $700+ for DX900+. Now if I was a serious racer, perhaps. But I have 44 year old boat that really solid (heavy by today's standards for a 35 footer) that is fun to sail. I'll let the J-boat folks beat me to start of happy hour... just as long as I don't miss 'last call'.

If I may say a word in favor of non-mechanical speed logs --

I've been using an Airmar CS-4500 ultrasonic speed log for 5 years.

They are far more accurate and responsive than paddlewheel logs and don't change as they accumulate slime like paddlewheel logs do.

Boat speed is a whole level higher in the hierarchy of necessary data, than leeway. I agree that leeway data for most of us will be in the angels-dancing-on-the-head-of-a-pin category, but boat speed is fundamental, and accuracy directly affects how useful this data is.

The CS-4500 is expensive and has a number of flaws, but I'll personally never go back to paddlewheels. This was one of those things which I never ever regretted spending the money on.

[ronstory]

Originally Posted by Dockhead

I think this is an entirely valid approach.

Most of us understand that less speed and more heel = more leeway and ruined VMG to windward. Sail flat and fast and you get upwind; don't and you don't. Most of us can feel the groove, pressure on the tiller, rudder angle -- the boat talks to us. We don't need hard data all that much.

On top of that, few cruisers actually do much sailing hard on the wind anyway. "Gentlemen don't go . . . " etc.

So yes, it takes a certain nerdy bent to care all that much about this, probably. I guess I'm one of those nerds . . .

Yes, flat is fast... and more comfortable. That said, I prefer the term 'geek' to 'nerd'... but that is just a personal call.

[ronstory]

Originally Posted by Dockhead

If I may say a word in favor of non-mechanical speed logs --

I've been using an Airmar CS-4500 ultrasonic speed log for 5 years.

They are far more accurate and responsive than paddlewheel logs and don't change as they accumulate slime like paddlewheel logs do.

Boat speed is a whole level higher in the hierarchy of necessary data, than leeway. I agree that leeway data for most of us will be in the angels-dancing-on-the-head-of-a-pin category, but boat speed is fundamental, and accuracy directly affects how useful this data is.

The CS-4500 is expensive and has a number of flaws, but I'll personally never go back to paddlewheels. This was one of those things which I never ever regretted spending the money on.

OK, the ADOTHOAP reference made me actually full-on laugh out loud. We are going to have to make that an official CF acronym. ;^)

So the ultra sonic is really that much better? I haven't noticed a big diff in the paddle slime effect but the boat is dived regularly so maybe that is staving it off.

As for the CS4500... what does it connect to? it seems to just put out a paddle pulse stream?

[Dockhead]

Originally Posted by ronstory

OK, the ADOTHOAP reference made me actually full-on laugh out loud. We are going to have to make that an official CF acronym. ;^)

So the ultra sonic is really that much better? I haven't noticed a big diff in the paddle slime effect but the boat is dived regularly so maybe that is staving it off.

As for the CS4500... what does it connect to? it seems to just put out a paddle pulse stream?

Yes, the CS4500 puts out analogue pulses and can be connected to an old-fashioned stand alone speed log, maybe with SeaTalk output. For a modern N2K system you need TWO converter boxes, one which converts analogue to 0183, and another which converts 0183 to N2K (or else use the 0183 listener port on the plotter as I do)

But the difference to a paddlewheel is huge. Far more responsive, and linear. On Panbo you can see some graphs from tests comparing CS4500, paddlewheel, and GPS in still water, which shows this dramatically. Slime does not change the correction factor as on a paddlewheel. It's truly a big leap forward, despite the several drawbacks (silty water makes it go crazy, to name just one).