Panbo

Airmar H2183 compass, best in class?

... written for Panbo by Ben Ellison and posted on Feb 26, 2009
Airmar_H2183_test_video2.jpg

Airmar's new H2183 heading sensor is motion compensated every which way; i.e. it's a 3-axis solid-state compass integrated with a 3-axis accelerometer and a 3-axis rate gyro. That's news both in terms of real world compass precision and the masthead performance of Airmar's PB200 Weather Station, whose gyro only senses yaw. In fact, while Airmar is pleased with the how well the PB200 seems to work on Dan Corcoran's sailboat, it may not promote the Weather Station to sailors until it can produce a model that incorporates the 3-axis gyro. But the H2183 is shipping now, for about $700 retail, and Airmar claims that its accuracy in "dynamic conditions" --- i.e. on a boat in a seaway --- is "best in class." And it's backing up the claim with the side-by-side testing video snapped above. While the "competitor" compass is not clearly identified, it's almost undoubtedly Maretron's well regarded SC200. The performance gauntlet is slapped down! And if Maretron or Simrad or whoever has a rebuttal, Panbo is ready to run it.

While I can not personally attest to H2183 performance, I've seen the unit and was quite impressed with the mounting system. The bracket can be fixed to any vertical bulkhead regardless of orientation, because the compass itself can be spun any which way in the bracket so it's aimed at the bow. In fact, you only need to ease up on two thumb nuts to fine tune it. That's dang clever, as is the flexible way Airmar incorporates both NMEA 0183 and 2000 data in one cable that can terminate to multiple plug or combiner/splitter box options. I'm getting personal experience with that data/cable scheme because it's also employed on the PB200 --- which I'll discuss soon --- as well as Airmar's new GH2183 GPS/Heading sensor, which is cousin to both.

Airmar_h2183_mounting_bracket.jpg

Posted by Ben on February 26, 2009 8:42 PM

Comments

Posted by: Alex at February 27, 2009 1:35 PM | Reply

The PB200 already has a 3-axis accelerometer and a 1-axis rate gyro as well as an onboard GPS. For purposes of correcting the True Wind Speed and Direction, would going to a 3-axis gyro really help? It could be that the H2183 is meant as a stand alone unit and that the PB300 is way out there.

Posted by: Olsonist at February 27, 2009 4:17 PM | Reply

It lloks impressive but for normal navigation (ie: connected to autopilot) why is pitch and roll so important?

Posted by: John at February 27, 2009 6:53 PM | Reply

Ben,

Looking at the video, the test "boat" is rotating in only one dimension, so how is it measuring angle of rotation in two?

Thanks,
John

Posted by: John at February 27, 2009 8:56 PM | Reply

John,

Unless the compass is perfectly aligned with the fore and aft rocking, then it will be picking up two dimensions.

Posted by: Anonymous at February 27, 2009 10:54 PM | Reply

Impressive demo, but... The pitching is way too fast for any boat I'd want to be on. It's good that Airmar does well in these conditions, but in more normal conditions the difference might not be so obvious. It would be nice to try this yourself. Does mounting it closer to the CG help, does it still work well when mounted on a much higher mast?

After looking at pictures of heading sensors on the internet, my wild guess is that the competitor is a Simrad. The Maretron SSC200 is black. All the more reason to repeat the test in the lab.

Posted by: norse at February 28, 2009 2:23 AM | Reply

I changed the screen capture to show how both sensors are detecting changes in pitch, roll, and heading, one a lot more than the other. I agree that the rocking horse test bed creates pretty extreme movement, but then again a seaway produces motion in all directions too, even when one axis seems predominant, especially high up.

Posted by: Ben at February 28, 2009 10:56 AM | Reply

First, is this product shipping in production quantities, or is pre-release PR hype?

Second, marketing claims are cheap to make and essentially meaningless these days. I hate to sound so jaded, but to quote The Who - "Won't get Fooled Again".

When some objective tester runs this through it's paces in real world conditions against the real competition that is in the market place, I'll pay attention. Until then, it's just a press release.

Posted by: Russ at February 28, 2009 4:29 PM | Reply

I am writing to comment about the mounting system Ben wrote about.

I have only installed, not evaluated the H2183 (boats on the hard), ... but the first thing I noticed was the great mounting system.

Ben was very verbose and right on about the mounting system ... considering there is just not much to say ... it goes from out of the box to installed very painlessly. Same with the N2K connection. When I plugged it into my Raymarine e-series, it just worked, making my first N2K install seem like child’s play.

Comparatively just relocating the flux-gate compass included with my autopilot, in an effort to get better accuracy, took 10x the install effort mostly due to the efforts to pull apart sections of my boat to just find a compatible surface that meet the installation manuals requirements.

Posted by: Dan Corcoran (b393capt) at March 1, 2009 12:24 PM | Reply

Ben: can you provide any more information on your suggestion of an updated Airmar PB200 more specifically designed for a sailboat masthead (i.e. with a 3-axis rate gyro)? In particular, is such a product actually in the works or is it more speculative?

The rocking horse demo is impressive, but such motion is not at all extreme or inappropriate compared to the motion of a sailboat masthead in a seaway. (As anyone who's had to climb a mast at sea can attest.)

Posted by: JCS at March 2, 2009 3:10 PM | Reply

Well, I spoke directly with Airmar's product/marketing people (even taped the conversation with my magic Pulse pen), and I was careful when I wrote that they "MAY not promote the Weather Station to sailors until it can produce a model that incorporates the 3-axis gyro."

So they were not definite about the change, and it sounds like they're still working with beta testers like Dan and others to fully understand the PB200's real world performance. And doing their own testing. There were various Weather Stations and compasses on board the Airmar boat I went out on last summer, all hooked up to the same comparative data analysis software you see in the video.

My impression (unlike that of certain cynical Panbots I know ;-) is that Airmar tests the hell out of products, while taking their sweet time to actually get them to market. It is, after all, an engineer founded and managed company (like Maretron, and Garmin, and ???).

Hence the SB(sailboat)200 might be the PB200 with a Windex on top, or might have the 3-axis rate gyro added, or who knows what... and I wouldn't venture a guess as to when, or even if, it will appear.

Posted by: Ben at March 2, 2009 4:15 PM | Reply

This is Dan ... (from Dan's Sailboat). The primary focus of my evaluation has been on the wind measurements, and the effectiveness of the motion correction software on my sailboat mast on different points of sail at different wind speeds and when downwind with asymetric sails. When I spotted the improvments of the PB200 over the PB100 I saw the application in sailing I described in the PB200 article I wrote here in February ... and was excited to get results better than I imagined.

When it comes to the compass however ... if Airmar had not reminded me there was a digital compass on the PB200 a few days before my boat was hauled for the winter, I wouldn't have even evaluated it beyond my original test to see if it was good enough for my radar MARPA feature to hold targets (it was). In that additional last day of testing my crew and I compared it to a Raymarine gyro and a Weems and Plath hand compass, on a flat sea. After seeing the results I won't ever trust my flux gate again for DR navigation practice ... but in a sense it isn't a fair test, the PB200 is up there free from all iron objects, stereo speakers, etc. etc. ... and my poor flux gate is subject to a 350+lb Yanmar engine four feet away, battery cables throwing feeding my DC panel 4-5 feet away in another direction throwing off magnetic fields that differ each time the fridge cycles, and all the accessories in my galley four feet the other way. Place a handheld compass on the same spot and the results are +/- 20 degrees depending on my point of sail. The calibration procedure improves upon that ... but even though I get a very high level of success reported at the end of the calibtration procedure ... the compensation made from the calibration procedure isn't so good as bared out in my last day test results.

If our thoughts about where a compass should be on a sailboat was inverted (as Ben suggested) ... I missed that entirely.

So I ask ... is there a real interest in this community in knowing how well the digital compass performs on a swaying 50 foot high sailboat mast? Does a wildly accurate compass heading matter to anyone, or is the COG from your GPS the primary focus (be honest here) or the great circle route calculated by the chart plotter what matters ... and a gyro compensated flux-gate compass otherwise good enough? If you want me to do some more digging ... tell me how you would use a more accurate compass heading on your boat so I can give this some more thought.

Posted by: Dan Corcoran (b393capt) at March 4, 2009 12:46 AM | Reply

Dan, Accurate, responsive, and fast heading data is very important to boats, for radar overlay, MARPA, autopilot performance, Class A AIS reporting, and probably other things I've left out.

It can even effect chart plotting if the plotter itself is fast enough. When I tried NavNet 3D last summer, its amazing performance at speed was slightly less so when we switched from using Furuno's satellite compass to its flux-gate. The onscreen boat didn't keep up with the real boat quite as well.

I don't know for sure, but believe that smart autopilots use heading input to help steer a straight line even when they're steering to a wind angle. If that's true, could it be a factor you didn't realize in your Airmar PB200 testing?

Posted by: Ben at March 4, 2009 9:12 AM | Reply

Here is some background info on why what Airmar is doing is important for us. In North America and Europe, one degree of tilt of a magnetic compass will cause 3 to 5 degrees error in compass direction. Greater tilts are common on a sailboat, so the compass errors could be huge if they were not corrected. Traditional compasses are gimballed to avoid this but they still have errors during acceleration. An electronic compass which is not gimballed needs to be compensated mathematically, which requires that the tilt (pitch and roll) be known.

Posted by: norse at March 5, 2009 12:10 AM | Reply

@Ben:

Yes, autopilots generally always steer to the heading input. The wind angle will slowly change the 'desired' heading, often accompanied by an indication thereof (for example a beep.)

The ones that I am familiar with also remember the orginal heading when you enable the pilot and go into an error mode when the 'desired' heading changes more than a number of degrees from the original heading.

Posted by: Kees at March 5, 2009 7:40 AM | Reply

Thanks, Kees, but what I meant was the internal intelligence of the pilot, its ongoing efforts to analyze what the boat is actually doing and adjust its rudder commands so the boat steers as straight a path as possible.

I don't think it matters if the pilot's goal is a specific apparent wind angle, a waypoint, or a course; regardless, the pilot still needs good, fast heading info to make good rudder decisions.

Posted by: Ben at March 5, 2009 8:07 AM | Reply

Yup, you're abolutely right. I did not intend to disagree so I my apologies if it came across that way, I only wanted to elaborate on how apparent wind mode works.

Fast accurate heading is instrumental to a pilot because that means it can trust the yawing info it is receiving.

Posted by: Kees at March 5, 2009 9:07 AM | Reply

I am about as sure as I can be, that the autopilot was not using the PB200. Raymarine technical support was clear that if another heading source is on the network, the autopilot will ignore it in favor of it's own fluxgate compass.

As I go to add compass evaluation to my testing, I am wondering how to capture the multiple compasses all at once with some software. It would seem the flux gate I will have to pick off with NMEA-0183 output, for the PB200 and H2183 will I be better served with them wired as 0183 or 2000 ?

Posted by: Dan Corcoran (b393capt) at March 6, 2009 7:18 AM | Reply

I apologize, I answered incorrectly just above.

When going through my PB200 test notes, planning more testing this Spring, I found that I documented taking the step to disconnect the existing fluxgate and substitute the PB200 heading output directly into the NMEA-0183 input of the autopilot during the testing I described.

Therefore, the autopilot was in fact receiving heading info from the PB200, in addition to, the wind angle to steer by the wind with ... which provided extraordinary results.

Posted by: Dan Corcoran (b393capt) at April 17, 2009 12:08 PM | Reply

I was wondering if anyone can address the following questions:

What is the difference between compass products called "rate compensated" and the Airmar, H2183? Is the rate compensated feature as accurate or more accurate then the design of the H2183?

A fluxgate compass like the one sold by ComNav needs a separate add-on unit called a "Heading Rate Stabilizer" to achieve heading information that ComNav says eliminates "Northern turning error", (whatever that may be). How does northern turning error come into play with the H2183?

Thanks for any help on this

Posted by: Richard C at June 26, 2009 10:41 AM | Reply

Richard, I believe that Northern turning error is symptomatic of gimbaled compasses. If the heading sensor gets pitch and roll from accelerometers or other tilt sensors, it shouldn't be an issue.

Posted by: bill at July 30, 2009 10:11 AM | Reply

Hi everyone.
This is Matti writing on 65d30m North, 025d11,1m East. I am tackling with just same problems as You all are.

AIRMAR H2183.
I have purchashed Airmar H2183 with Garmin GI10 display. Transducer itself is mounted on 2,5 meters height from standard bridge in an 24 meter long WWII harbour tug, in magnetically cleanest possible place.
Airmar H2183 is put to deliver heading information 10 times per second and connected to Garmin GI 10 display (compass rose mode).
Airmar calibration has been done with Airmar Weathercaster software, which seems to be the only possibility. I have not (yet) got any answer from Airmar, if there is possibilities to do "tabular calibration" to "feed in" previously steered true or magnetical courses. So, the only way to calibrate Airmar H2183 seems to be turning circles on calm seas.

GARMIN GPS18
As a gps I use Garmin GPS 18-5Hz, which is fast enough to deliver heading information 5 times per second.

SEACLEAR
I use a free program SeaClearII from Sping. Charts are from Finnish Maritime Administration.
In SeaClear I have connected Garmin GPS18 for true heading (only when moving) and Airmar H2183 for Compass heading.


NAVIGATIONAL TESTING
I have started to test Airmar H2183 + Garmin GI10 + SeaClear and Garmin GPS-18-5Hz + Seaclear.

1st TASK
was to make a deviation table. It shows a strong quarter circle error (D-coefficient) which the automated calibration routine inside Airmar WeatherCaster and H2183 could not (yet) correct.
This tells me that I have to seek more properly, magnetically cleaner place above deck. This work will be continued next summer when sea is not anymore frozen.


WHY DO I WANT A MAGNETIC COMPASS HEADING

I want heading information because it simply is the most important parameter of a moving boat or ship.

I also want to connect magnetic heading information to a radar an to a properly (= compass card like) heading display (which is Anshutz Steering Repeater Compass display 133-560).


WHAT ADD ON�s DO I WANT OF MY NAVIGATION SOFTWARE is big compass rose display with two needles, COG-heading line (from gps when moving) and Compass (keel) heading line

Regards Matti

Posted by: Matthew at October 28, 2009 5:16 AM | Reply

Garmin is selling this too. They call it a "High Performance Marine Heading Sensor" (no Garmin model name, it's just an accessory) and say it is "based on Airmar H2183". I'm not sure what the value added is (it costs more).

Posted by: norse at December 8, 2009 4:21 PM | Reply

How about ...

http://www.panbo.com/archives/2009/11/airmar_h2183_improves_situational_awareness_feature_on_raymarine.html

Plus more if the Garmin autopilot uses both the heading and the gyro outputs of the H2183

Posted by: Dan Corcoran (b393capt) at December 9, 2009 6:28 AM | Reply

I need to replace my faulty Raymarine flux gate compass for my autopilot, which has never been great even after a number of calibrations, and has now gone totally wonky.
Can anyone recommend whether the Airmar H2183 or the Maretron SSC200 is significantly better than the other (besides the mounting system, which I believe the Airmar wins hands down)?

Thanks
Evan

Posted by: Evan at September 14, 2010 12:50 PM | Reply

I too am looking at these alternatives: Airmar H2183 vs Maretron SSC 200. I want to output heading data on N2K to my NavNet 3D (MFD12) and simultaneously output HDG directly to an autopilot that only takes 0183.

I have learnt that the Airmar H2183 is about the same cost as the Maretron SSC 200, but ships without cables. A special combination cable and junction box need to be separately purchased to supply N2K & 0183 simultaneously...all a bit complicated compared to the SSC 200 which has separate N2k & 0183 ports and ships with an 0183 cable and takes std N2K cabling, without any need for junction boxes, etc.

Actually, my bigger question relates to the rate gyro features of these sensors. What is the right trade-off between getting close to the vessel's centre of gravity (cog)vs avoiding magnetic fields? The Maretron SCC 200 & the Airmar H2183 both have sophisticated rate gyros, but their instruction manuals still urge locating close to the vessel's cog. In my stabilizer-equipped motor cruiser, the closer I move to the c.o.g. the more ferrous metal & magnetic fields I have to deal with. If I go "up" to a locker under the flybridge helm, I'm further away from the c.o.g. but completely free of magnetic interference. I would have thought I'd be better doing the latter and letting the rate gyro features do their work....what do others think?. .

Posted by: Anonymous in reply to Evan at September 12, 2011 6:15 AM | Reply

I am having excellent results on my sailboat placing the H2183 far forward (80% of he boats weight is behind the compass) and starboard of the center of gravity.

I use the feature of connecting the 0183 direct to the autopilot, and the N2K to my instruments. I didn't use a junction box, just terminal blocks. You need to choose choose a power source, either thru N2K or direct. I choose direct so that my N2K and instruments can be powered down and I can still run the autopilot.

Posted by: Dan Corcoran (b393capt) at September 12, 2011 9:01 AM | Reply

Hi Evan

Go for the Maretron, it is a fantastic sensor, and setup for magnetic deviation on board is simple (power on the unit and perform 4 complete turns of the boat within the first 8 mins) Additionally because there are no moving parts inside it can be mounted next to ferrous metals (as long as they are stationary, not like a tool bag or dive gear) We always use this sensor via N2K to NN3D and 0183 to a Navpilot.

Note: can only be powered via N2K side.

Posted by: Nick F at September 12, 2011 4:21 PM | Reply

This is Matthew writing from 65d30m North, 025d11,1m East.

It is good to see that others on lower latitudes are working with those transducers as Maretron SSC200 and Airmar H2183, and others.


After few years of my last writing I have found that:

a) On higher latitudes the inclination angle of earths magnetic field makes all transducers almost useless in bad weather and exellent in good weather. This is based on the fact that on my home waters the inclination is 76 degrees, so the boat has only 14 degrees to pitch or roll before the magnetic flux vector comes from "zenith" of the transducer and after that the heading changes it direction of amoun 180 degrees.

This phenomena has not been taken into account on Airmar H2183 according my sea trials. Airmar does not deny that.

On calm waters both of them work exellent.

On speeds more than 10 knots Northernly Turning Error has no effect on them only as long as the boat itself does not heel. If the boat heels during tacking, both of those transducers became useles because of "zenithing".

On acceleration and deceleration Airmar is good if and only if heading remained constant.


b) Installation of transducer to the cog or as hig as possible

ba) has effect on magnetic cleaness of environment of transducer and hence on the calibration.

bb) has no other effect on the transducer, as long as the total acceleration of transduder does not exceed transducers maximum values, perhaps 1g (= 9.81m/s2) (which Airmar does not tell). This will not be easily exceeded by sail, but by fast patrol boats it is the main limiting factor.

If You install the transducer in the cog, the angular acceleration remains the same as in mast top but trajectional acceleration ceases nearby zero and the only component making acceleration left is the heaving of the boat body itself. This is the reason why manufactures suggest placing nearby cog (no need for transducers of so wide dynamics).

Comparison of solid state transducers with transducers with moving parts inside (= gimballed), like all traditional magnetic compasses, fluxgate transducers (VDO ADIS 360, Ritchie DH100, Raymarine Autopilots (Autohelm 4000 in my own boat) etc.) shows that they both have same weakness, acceleration, but for different reasons.


On compassing the main aspect is to maintain horisontal level for measuring magnetic field components.

On lower speeds and latitudes and calm seas traditional compass, gimballed fluxgate compass and solid state compass with electronic levelling installed, in magnetically clean place as close cog as possible, are all good.

When seas go high, latitudes go hig and velocities go high, it means more accelerations which make moving parts swing, and far away off of the cog parts accelerate and decelerate with hig velocities.

Perhaps the 9 degree of freedom- transducers are not good enough for reliable compassing. I suspect - and recommend, that particularly on high latitudes we need a solid state transducer which has 3 orthogonal magnetic field measuring grids, placed at 60 degrees of each other, instead of only one, just to avoid the "zenithing" problem. 3-axis acceleration and 3-axis velocity measurements are a must to maintain virtual level of compass transducer.

I did not answere much, nor were there particular questions addressed to me.
But anyway, heading transducers are so interesting ... I will continue tests ...

Posted by: Matthew at May 7, 2012 5:07 PM | Reply

Thank you very much Matthew for this insight.
I am in the process of positioning an H2183 in the boat and this comes at just the right time.
My current plan, using your advice, is in the cabin just under the cabin roof near the mast step. The old compass used to be on the cabin floor by the mast support just inches away from the keel bolts which I am not so excited about. This is for a 45' GRP sailboat with deck stepped mast.
I have had serious problems with autopilot behavior in rough seas previously and have also replaced the autopilot computer to a Simrad AC12 this winter.

One question: Have you tried a GPS compass in these situations or what performance would you expect?

Posted by: Henning at May 9, 2012 12:15 PM | Reply

Hello Henning, and all others.

Answer to You question about GPS compass: Yes, I have tried a GPS Compass, Furuno SC50, and it was good, but still too expensive.

Posted by: Matthew in reply to Henning at August 14, 2012 12:42 PM | Reply

Leave a comment