Experience with Airmar DX900+?

[Dockhead]

The Airmar DX900+ has been out for a few years by now -- has really no one got any experience with it?


I had a CS4500 ultrasonic speed log for a number of years, and it finally failed last year. I couldn't source a DX900+ in Europe so went with the UST800 which is ultrasonic so presumably the same old technology as the CS4500.


Well, it's a bit of a disappointment. It seems less accurate than the CS4500. I calibrated it a couple of time but found it inconsistent and apparently off by 0.5 knots one way or the other.


So I think again about the DX900+. The early ones were plagued with trouble. Have Airmar got it sorted out now? Anyone using this?

[s/v Jedi]

Originally Posted by Dockhead

The Airmar DX900+ has been out for a few years by now -- has really no one got any experience with it?


I had a CS4500 ultrasonic speed log for a number of years, and it finally failed last year. I couldn't source a DX900+ in Europe so went with the UST800 which is ultrasonic so presumably the same old technology as the CS4500.


Well, it's a bit of a disappointment. It seems less accurate than the CS4500. I calibrated it a couple of time but found it inconsistent and apparently off by 0.5 knots one way or the other.


So I think again about the DX900+. The early ones were plagued with trouble. Have Airmar got it sorted out now? Anyone using this?

No experience but I noticed that the specs for UST800/850 don’t even mention speed accuracy (but they do temperature accuracy so it’s not that they “forgot” to mention it) and for the DX900+ they list it as 0.1knot

[Dockhead]

Originally Posted by s/v Jedi

No experience but I noticed that the specs for UST800/850 don’t even mention speed accuracy (but they do temperature accuracy so it’s not that they “forgot” to mention it) and for the DX900+ they list it as 0.1knot

I noticed that also. Nor is the accuracy of the CS4500 mentioned in the specs (in the blurb it's "as much as 0.1 knots".

[s/v Jedi]

Originally Posted by Dockhead

I noticed that also. Nor is the accuracy of the CS4500 mentioned in the specs (in the blurb it's "as much as 0.1 knots".

On the other hand, for my dst800 they don’t list it either and I think it’s pretty accurate. Like for many it has begun transmitting silly temperature data but looking at performance and pricing, I have decided to go with the new dst810 for now. I can buy 3 for the same price as a dx900+ I think.

[Jenta]

I hope you're wrong about the performance of the UST800 and it is either your installation, calibration or a bad unit.
Hate to wish ANY of that on anyone, but I'm installing my Airmar UDST800P in a few weeks so...

I had just assumed the .1 knot resolution was more or less the accuracy, but knew that was an assumption which we should all know what happens when you assume of course.
But I certainly was thinking it would be better than +/- .5 knots.

Still, my paddle wheel it is replacing never seemed to work well even when freshly cleaned, which I used to do a lot before giving up. I know calibration was largely to blame but I could not calibrate it properly with the electronics set up the way they were and this time around, I've got all new NMEA 2000 compatible gear from autopilot to wind vane and even if it IS only capable of +/- .5 knot accuracy it will still be an improvement.

(Though a bit disappointing of course)

[thunderhoof]

I was able to buy a CS4500 recently. I wanted to adapt it to my older Nexus system that has a pulse train speed input rather than the NEMA data input now supported by most systems. After installing a 2:1 divider on the pulse train I am now within the calibration range of the Nexus system.
I am calibrating the knotmeter using several methods: timed runs over fixed distance, upwind in average wind conditions, motoring in the mooring basin at fixed rpms and comparison to the gps with and against current. I find if I do this for a few outings I can get the calibration to about 1 count of calibration (about 0.6 %).
Now you all have thrown me a curve. Although the precision of my configuration is 10,000 pulses/ nm (0.01%) I have assumed the transducer is stable to better than 0.5% over a boatspeed range of 3 to 8 knots in waters that are between 55 and 75 deg. F.
Does anyone have any information that disputes or confirms these assumptions? If the stability of the transducer is several percent it is a waste of time trying to calibrate it more accurately than that.
Yes, I need accurate speed to determine my best trim and to spot if I have picked up kelp or other nasty stuff.

[Dockhead]

Originally Posted by thunderhoof

I was able to buy a CS4500 recently. I wanted to adapt it to my older Nexus system that has a pulse train speed input rather than the NEMA data input now supported by most systems. After installing a 2:1 divider on the pulse train I am now within the calibration range of the Nexus system.
I am calibrating the knotmeter using several methods: timed runs over fixed distance, upwind in average wind conditions, motoring in the mooring basin at fixed rpms and comparison to the gps with and against current. I find if I do this for a few outings I can get the calibration to about 1 count of calibration (about 0.6 %).
Now you all have thrown me a curve. Although the precision of my configuration is 10,000 pulses/ nm (0.01%) I have assumed the transducer is stable to better than 0.5% over a boatspeed range of 3 to 8 knots in waters that are between 55 and 75 deg. F.
Does anyone have any information that disputes or confirms these assumptions? If the stability of the transducer is several percent it is a waste of time trying to calibrate it more accurately than that.
Yes, I need accurate speed to determine my best trim and to spot if I have picked up kelp or other nasty stuff.

I wouldn't assume that your transducer will give you anywhere near that kind of accuracy with regard to your real boat speed. I think it's fundamentally impossible to measure boat speed very accurately by any of these means.


One major issue is the disturbance of the water flow by the boat passing through it and the boundary layer, which moves with the boat to some extent and not to the same extent all along. The ultrasonic transducers are supposed to be better because they are measuring water some distance from the hull, but I don't think it's enough to overcome this effect.


My CS4500 was the best speed log I ever had, but it was nothing like as accurate as you are assuming. At best within a few 10ths of a knot, and these devices need clean clear water -- any silt in the water and they go crazy.

[GordMay]

Originally Posted by Jenta

...
I had just assumed the .1 knot resolution was more or less the accuracy ...

Accuracy is how close a reported measurement is to the true value being measured.

Resolution is the smallest change that can be measured.

Precision describes the reproducibility of the measurement.

Sensitivity is an absolute quantity, the smallest absolute amount of change that can be detected by a measurement.

[pjShap]

I've been using the Airmar DX900+ for about 3 years now. I found calibration, last tried when I first installed the unit 3 years ago, impossible. There's a phone app that connects to the sensor's processor box via Bluetooth. The app has an auto-calibrate option, which I presume uses the phone's GPS as the calibrated speed. After the auto-calibration I saw seemingly random and totally ridiculous readings, like 1500 knots. I'm really not that fast. However, I set it back to default values, and the readings seem reasonably accurate ever since. I say reasonably accurate because I have not made the effort to test in smooth water with zero current, or run back and forth on a measured distance, etc. I have a steel boat, and after the initial purchase, when Airmar had insisted this was a great transducer for a steel boat, tech support told me they no longer recommend it for steel hulls. I rarely clean the sensor, and it continues to work. In Maine this summer, with colder water and less growth, it appeared reasonably accurate for two months without cleaning. In warmer waters it benefits from cleaning every few weeks.

[Jenta]

Originally Posted by GordMay

Accuracy is how close a reported measurement is to the true value being measured.

Resolution is the smallest change that can be measured.

Precision describes the reproducibility of the measurement.

Sensitivity is an absolute quantity, the smallest absolute amount of change that can be detected by a measurement.

Thank you Gord for the accurate definitions of these terms at precisely the right time to bring resolution to this sensitive issue.

I was just assuming, albeit knowingly optimistically - that since there was no accuracy given in the specs that hopefully the accuracy was +/- the resolution.

Regardless, I really just want to be able to have a more accurate read between true and apparent wind and have been using GPS speed on my B&G Triton since it can do that calculation while all my old Raymarine gear can't and will only show that the apparent and true wind speed were the same because my old speed transducer was always having fits.

So if the Airmar UDST800 is +/- .5 knots, that will still be an upgrade for me!

[Dockhead]

Originally Posted by Jenta

Thank you Gord for the accurate definitions of these terms at precisely the right time to bring resolution to this sensitive issue.

I was just assuming, albeit knowingly optimistically - that since there was no accuracy given in the specs that hopefully the accuracy was +/- the resolution.

Regardless, I really just want to be able to have a more accurate read between true and apparent wind and have been using GPS speed on my B&G Triton since it can do that calculation while all my old Raymarine gear can't and will only show that the apparent and true wind speed were the same because my old speed transducer was always having fits.

So if the Airmar UDST800 is +/- .5 knots, that will still be an upgrade for me!

Usable true wind data is actually what I'm after myself. I posted about it in other threads. True wind data is extremely useful for different purposes, from proper lay lines to autopilot performance downwind.



With boat speed off by half a knot, the true wind data won't be much good except in very strong conditions.



The other big obstacle to good true wind data is measuring the wind itself. Upwash, downwind acceleration, wind shear, heeling error, mechanical effects of the boat bouncing around -- all these make it really hard to get even decent apparent wind data.



Did you know, incidentally, that really big budget racing teams often back calculate apparent wind from true? After they've done all their sophisticated corrections and calibration on the true wind, the apparent wind as calculated is better than what is measured.

[pjShap]

Originally Posted by Dockhead

Did you know, incidentally, that really big budget racing teams often back calculate apparent wind from true? After they've done all their sophisticated corrections and calibration on the true wind, the apparent wind as calculated is better than what is measured.

I don't think that's possible. The original apparent wind measurement was the basis for the calculations. Therefore the calculation cannot be better than that original measurement.

[Dockhead]

Originally Posted by pjShap

I don't think that's possible. The original apparent wind measurement was the basis for the calculations. Therefore the calculation cannot be better than that original measurement.

Sure it can be better -- you're applying different corrections. Whether it actually IS better is controversial -- Ockham for example think it's not. See: http://www.ockam.com/2013/06/02/wind-shear/

Read up on it if you don't believe me, but back-calculated apparent wind really a thing. AFAIK, one whole generation of B&G racing instruments could NOT display apparent wind other than back-calculated from true.

[pjShap]

Originally Posted by Dockhead

Sure it can be better -- you're applying different corrections. Whether it actually IS better is controversial -- Ockham for example think it's not. See: Wind Shear and Gradient - Ockam Sailing Instruments

Read up on it if you don't believe me, but back-calculated apparent wind really a thing. AFAIK, one whole generation of B&G racing instruments could NOT display apparent wind other than back-calculated from true.

What would be the source of those better corrections? All you know is apparent wind speed and direction, speed and heading through the water (for true wind calculation), GPS SOG and COG (for ground wind calculation). None of those can correct an inaccurate apparent wind speed or direction. There is no additional information about the wind.

[Dockhead]

Originally Posted by pjShap

What would be the source of those better corrections? All you know is apparent wind speed and direction, speed and heading through the water (for true wind calculation), GPS SOG and COG (for ground wind calculation). None of those can correct an inaccurate apparent wind speed or direction. There is no additional information about the wind.

Sure there is. Data from the rate gyros, just to name one example, corrects the apparent wind measurements for motion at the masthead.



Here is the process laid out in the SailMon calibration manual, including back-calculation of apparent wind from true:





I learned a lot about all this by reading this and bet you will too: https://sailmon.com/files/Manuals/Ca...nce-Rev1.4.pdf