Laptop interface to TackTick NMEA

[tspeer]

Originally Posted by Tim B

...Anyway, on a more serious note, have you done any serial programming on macs?

No, and I hope not to. I aim to spend my time re-designing the boat, not writing tools to acquire the data to develop the tools to redesign the boat. If there's software out there that can do the job, I'll use it.

If so, could you pass me a read routine and I'll try to get something into the OpenPilot NMEA library so that you can use at least the basic NMEA test app. Please note that I don't have a mac, but I can e-mail you the code off-list if you like.

I'm also running XP under Parallels, and if I have to, I can boot up in XP. By all means shoot me an email. You can email me at tspeer dot com. Literally.

[Tim B]

Tom,

I Should have attached this screenshot on my last post. Please note that this is only the test-program. Support for the Windows serial system is not great, but it will be improved (eventually). I'll e-mail you the source off-list.

Cheers,

Tim B.

[tspeer]

Progress! I was able to read the NMEA data with the TacticTool software today, both with the computer booted into XP, and running XP in a Parallels window. What I didn't know before was I needed to click on the USB controller button on the Parallels window and select "Apple Wireless". Then I was able to add the MiniPlex41BT as a Bluetooth device and read it just fine.

The boatyard tech says there's a problem where the wires go into the GPS antenna, so the antenna may have to be replaced. I'd never have found that on my own. I may have GPS data going by the end of tomorrow....

[yipster]

well done! depending on use you may be in for some more pleasant study's

[tspeer]

Success! It turned out the wires had separated from the board inside the GPS unit. Once the GPS antenna was replaced, everything works as intended. I think the architecture I posted earlier is about the simplest and cheapest full-featured system available. And since most of it is wireless, it's very flexible and easy to install.

What software is that you're showing, Yipster?

[yipster]

click the link under the pic above and you find transas witch looks pretty much the same as i had it running under windows 95 in dos.
good program but there are more, some free. funny, well whats funny, that you encountered similar product problems installing.
i prefer the over a basic laptop simple and cheap systems full-featuring things myself better than the screw on your dash ready systems.

[Tim B]

Tom,

Did you have any luck with the Serial sample code I posted?

I'd like to know how well it runs under emulation on a Mac. If that works, then I can try to get the mapping widget to compile (under Windows, against GDAL), and give you a reasonable piece of Nav software to try.

I have done the satellite view as well now, but there are still a few (minor) bugs. I think these are in the NMEA routine though.

However, It's hard to tell whether it was the code, or the GPS, as the usb connection on the back of the GPS was not good.

I have also implemented safe/warn/danger markings on the gauges, and added the signals for their full functionality. There are also some rectangular gauges now, but they are only very simple.

Hopefully I'll get some more work done on it tonight :-)

Cheers All,

Tim B.

[tspeer]

Originally Posted by Tim B

Did you have any luck with the Serial sample code I posted?...

I haven't had a chance to try it, yet. I've had the boat at the local marina the past week, and I've spend all my free time working on it instead of the computer.

The GPS is working, now. However, the problem I'm having is extraneous wind data messages. Here's a sample of a few frames, taken while sitting at the dock:

$IIHDG,219,,,17,E*1E
$IIMTW,+08.0,C*30
$IIMWV,000,R,00.0,N,A*13
$IIMWV,079,T,06.0,N,A*1D
$IIRMC,021543,A,4724.104,N,12219.839,W,00.0,268,030408,17,E,A*11
$IIVHW,,,219,M,00.0,N,,*6E
$IIVLW,00129,N,000.0,N*59
$IIVWR,000,R,00.0,N,,,,*61
$IIDPT,004.2,-1.0,*46
$IIGLL,4724.104,N,12219.839,W,021543,A,A*48
$IIHDG,219,,,17,E*1E
$IIMTW,+08.0,C*30
$IIMWV,082,R,06.5,N,A*1A
$IIMWV,000,T,00.0,N,A*15
$IIRMC,021544,A,4724.104,N,12219.839,W,00.0,269,030408,17,E,A*17
$IIVHW,,,219,M,00.0,N,,*6E
$IIVLW,00129,N,000.0,N*59
$IIVWR,082,R,06.5,N,,,,*68
$IIDPT,004.2,-1.0,*46
$IIGLL,4724.104,N,12219.839,W,021544,A,A*4F
$IIHDG,219,,,17,E*1E
$IIMTW,+08.0,C*30
$IIMWV,085,R,06.5,N,A*1D
$IIMWV,082,T,06.5,N,A*1C
$IIRMC,021545,A,4724.104,N,12219.839,W,00.1,271,030408,17,E,A*1E
$IIVHW,,,219,M,00.0,N,,*6E
$IIVLW,00129,N,000.0,N*59
$IIVWR,085,R,06.5,N,,,,*6F
$IIDPT,004.2,-1.0,*46
$IIGLL,4724.104,N,12219.839,W,021546,A,A*4D
$IIHDG,219,,,17,E*1E
$IIMTW,+08.0,C*30
$IIMWV,000,R,00.0,N,A*13
$IIMWV,085,T,06.5,N,A*1B
$IIRMC,021546,A,4724.104,N,12219.839,W,00.1,269,030408,17,E,A*14
$IIVHW,,,219,M,00.0,N,,*6E
$IIVLW,00129,N,000.0,N*59
$IIVWR,000,R,00.0,N,,,,*61
$IIDPT,004.2,-1.0,*46
$IIGLL,4724.104,N,12219.839,W,021547,A,A*4C
$IIHDG,219,,,17,E*1E
$IIMTW,+08.0,C*30
$IIMWV,000,R,00.0,N,A*13
$IIMWV,000,T,00.0,N,A*15
$IIRMC,021547,A,4724.104,N,12219.839,W,00.0,270,030408,17,E,A*1C
$IIVHW,,,219,M,00.0,N,,*6E
$IIVLW,00129,N,000.0,N*59
$IIVWR,083,R,06.0,N,,,,*6C
$IIDPT,004.2,-1.0,*46
$IIGLL,4724.104,N,12219.839,W,021548,A,A*43


Notice the multiple messages like
$IIMWV,000,R,00.0,N,A*13
$IIMWV,079,T,06.0,N,A*1D
and
$IIMWV,082,R,06.5,N,A*1A
$IIMWV,000,T,00.0,N,A*15
and
$IIMWV,000,R,00.0,N,A*13
$IIMWV,000,T,00.0,N,A*15

The messages with 0 velocity at 0 degrees make the data look like it's filled with drop-outs. However, the signal strength from the mast-head unit is good and the instruments aren't flickering (or providing biased readings) like you'd think they would if they were responding to the null data, too. So far, I don't see any pattern to it that would allow me to easily eliminate the bad data, except for just throwing out the messages based on their containing null data. I don't have anything else putting out NMEA data, so it's not like another unit is stepping on the sensor data.

I'm ferrying the boat back to her home marina on Sunday, and I'll take data under sail, then. After that, I'll be doing trival stuff like taxes and work, so it will be a little while before I get back to important stuff like analyzing data.

[RHough]

Originally Posted by tspeer

The GPS is working, now. However, the problem I'm having is extraneous wind data messages. Here's a sample of a few frames, taken while sitting at the dock:

Notice the multiple messages like
$IIMWV,000,R,00.0,N,A*13
$IIMWV,079,T,06.0,N,A*1D
and
$IIMWV,082,R,06.5,N,A*1A
$IIMWV,000,T,00.0,N,A*15
and
$IIMWV,000,R,00.0,N,A*13
$IIMWV,000,T,00.0,N,A*15

The messages with 0 velocity at 0 degrees make the data look like it's filled with drop-outs. However, the signal strength from the mast-head unit is good and the instruments aren't flickering (or providing biased readings) like you'd think they would if they were responding to the null data, too. So far, I don't see any pattern to it that would allow me to easily eliminate the bad data, except for just throwing out the messages based on their containing null data. I don't have anything else putting out NMEA data, so it's not like another unit is stepping on the sensor data.

I'm ferrying the boat back to her home marina on Sunday, and I'll take data under sail, then. After that, I'll be doing trival stuff like taxes and work, so it will be a little while before I get back to important stuff like analyzing data.

The VWR sentence shows the same error.
I don't see a VPW sentence in the data.

The HDG sentence has 219M with 17E so 236 True.
The MWV sentence should read 318 T or about 82 R
The VWR sentence shows either 0 Right or about 82 Right
The HDG sentence is missing deviation information.

I'm surprised that something in the system is not flashing a calibration error.

Have you gone through the calibration process on the compass and wind instrument? The MWV sentence with the T qualifier should read the True Wind Direction 0-360 based on the HDG sentence and with the R qualifier it should read 0-360 from the bow.

With the VWR sentence reading 82 Right, 219 + 17E + 82R = 318 True

Sounds like the compass has not been swung (no deviation) and the wind instrument has may not be aligned.

Once the HDG sentence is complete and the wind instrument and compass are talking, the extra data should go away. The 0 -82 reading might be a 'feature' to trip a calibration needed message.

As I've said, I don't know the Tactick system, but the data you posted is very similar to what I've seen before prior to the calibration process. I would hope that you get the VPW sentence after everything is happy.

[tspeer]

Originally Posted by RHough

...
Have you gone through the calibration process on the compass and wind instrument? ...Sounds like the compass has not been swung (no deviation) and the wind instrument may not be aligned....

All true. I've not made any attempts at calibration, yet. I wanted to get everything talking so I could ensure the calibrations I entered were mutually consistent. I'm not sure what I can use as an accurate reference, except GPS. Hence my urgency at getting GPS into the mix.

One test I intend to do is to pick a day with calm winds (not too difficult in Puget Sound!) at slack tide. Then I'm going to motor in a big, square figure 8, all the time recording data. The straight legs on cardinal headings, plus looking up the local magnetic variation, should get the compass calibrated against the GPS true course. It may also help to identify mounting bias and scale factor for the wind instruments, at least for the low speed range (<10 kt). A similar maneuver with wind will further calibrate the wind instruments at different apparent wind angles with bare pole. The assumption for all of this will be that leeway is zero and current is zero, so the GPS velocity is down the centerline. Averaging over several minutes should help minimize the variation in GPS velocities.

I may write a Kalman filter to estimate instrumentation biases and scale factor errors. Then repeat some of the tests to validate against data that weren't used to estimate the parameters originally.

Once I have the instruments calibrated with bare poles, I need to start working on position error. I may be able to use windup turns to get a handle on speedo position error with leeway angle. Wind instrument position error will be a bit harder to estimate. By again sailing at slack tide, and perhaps stopping periodically to do a GPS current check, I may be able to estimate leeway angle from GPS minus heading.

I'm not sure how to handle leeway angle when sailing under general conditions. I don't see how to distinguish leeway from current. A vertical vane would be a handy sensor, but I have no intention of putting another hole in the bottom to mount it!

Ultimately I intend to develop a VPP for the boat. It would be cool to incorporate the VPP into a Kalman filter to process the measured data in real time to both smooth the data and estimate the typical computed quantities such as true wind direction and current.

[RHough]

Originally Posted by tspeer

All true. I've not made any attempts at calibration, yet. I wanted to get everything talking so I could ensure the calibrations I entered were mutually consistent. I'm not sure what I can use as an accurate reference, except GPS. Hence my urgency at getting GPS into the mix.

One test I intend to do is to pick a day with calm winds (not too difficult in Puget Sound!) at slack tide. Then I'm going to motor in a big, square figure 8, all the time recording data. The straight legs on cardinal headings, plus looking up the local magnetic variation, should get the compass calibrated against the GPS true course. It may also help to identify mounting bias and scale factor for the wind instruments, at least for the low speed range (<10 kt). A similar maneuver with wind will further calibrate the wind instruments at different apparent wind angles with bare pole. The assumption for all of this will be that leeway is zero and current is zero, so the GPS velocity is down the centerline. Averaging over several minutes should help minimize the variation in GPS velocities.

I may write a Kalman filter to estimate instrumentation biases and scale factor errors. Then repeat some of the tests to validate against data that weren't used to estimate the parameters originally.

Once I have the instruments calibrated with bare poles, I need to start working on position error. I may be able to use windup turns to get a handle on speedo position error with leeway angle. Wind instrument position error will be a bit harder to estimate. By again sailing at slack tide, and perhaps stopping periodically to do a GPS current check, I may be able to estimate leeway angle from GPS minus heading.

I'm not sure how to handle leeway angle when sailing under general conditions. I don't see how to distinguish leeway from current. A vertical vane would be a handy sensor, but I have no intention of putting another hole in the bottom to mount it!

Ultimately I intend to develop a VPP for the boat. It would be cool to incorporate the VPP into a Kalman filter to process the measured data in real time to both smooth the data and estimate the typical computed quantities such as true wind direction and current.

I would hope that after a 360 deg circle the heading sensor should write its own deviation table. The fact that the HDG sentence has blanks where the deviation information should be, may be an indicator that the heading sensor does not have that ability.

If the heading sensor does not need to be swung, the masthead unit has to be checked to make sure that MWV (R) and VWR read 0 win the boat is headed into the wind. Some units need a full circle in calibration mode, then halving the difference between tacks adjusts the on center alignment. Once that alignment is done, the MWV (T) reading should be correct. I have to assume that the wind instrument can use either GPS speed (AACCCKKK!) or knotmeter input to display VPW (VMG), why that sentence is not in the data stream may indicate that the information is not output in NMEA.

The heading information is coming from the compass and the data stream shows a 17E variance. This is picked up from the RMC line from the GPS output. No need to fuss with charts and do math to get variation.

I don't see the need for accurate leeway information for what you intend to do. VPW and VHW give you L/D of the boat. The only need I know of to have leeway information is to calibrate wind instruments so the boat can be compared to VPP polars. Since you are creating polars as a base point, you don't have targets to hit, you will have a toll to measure changes, but not measure against predicted performance.

It sounds like you are going to write a lot of code to do what an off the shelf software package can do for you.

Raymarine's RayTech software it based on the kiwitec stuff that the AC guys and the whitbread guys used. Documentation sucks, but you can enter corrections for wind angle and speed to compensate for sail induced errors. It will do strip charts on almost any data stream and you can use best averages to create/correct polars. Once it has polar information it does laylines and all that good stuff, plus it has a starting line function that calls distance and time to the line and to the laylines at each end. Once you head offshore, it uses GRIBBs and the boat's polars to do routing. Strangely enough RNS does not calculate set/drift, I have been assured that this ability is on the wish list for the next version (6.1) that should be out this month for Vista (gag) and the new G-series.

Cheers,

Randy

[tspeer]

Originally Posted by RHough

I would hope that after a 360 deg circle the heading sensor should write its own deviation table.

You're right, it does have that capability. I had just skimmed over much of the calibration section in the documentation before, because most of it required manual entry of the calibration parameters. But now that I'm ready for the information, I see that it will self-calibrate by doing a turn.

I don't see the need for accurate leeway information for what you intend to do. ... Since you are creating polars as a base point, you don't have targets to hit, you will have a toll to measure changes, but not measure against predicted performance.

My objective in this is not to just build an empirical VPP, but to build up a VPP from first principles using CFD and parameter estimation, and validate it with the performance data. So it's not just a matter of being able to know what the boat will do in a given set of conditions, rather it's forming a baseline for engineering future changes to the boat. I need to know why it performs as it does, not just what it does.

It sounds like you are going to write a lot of code to do what an off the shelf software package can do for you.

Actually, I plan to write as little code as possible. But some is inevitable.

Raymarine's RayTech software it based on the kiwitec stuff that the AC guys and the whitbread guys used. ...

I'm currently using TacticTool. It does most of that, too, and it's a lot cheaper. They've been responsive to the questions I've asked so far, so I suspect that they may be willing to collaborate on adding additional capabilities. But I need to get the basics down, first.

[yipster]

thats nice new sailing software i see at TacticTool

[RHough]

Originally Posted by tspeer

You're right, it does have that capability. I had just skimmed over much of the calibration section in the documentation before, because most of it required manual entry of the calibration parameters. But now that I'm ready for the information, I see that it will self-calibrate by doing a turn.

Good ... reading the instructions is always a good thing

I've chased this sort of thing more times than I care to admit.

On the other hand I've also done a complete install only to find that an instrument did not output the data listed in the specs. Tech supports solution? Change the documentation ...

It would be nice to see the data stream after all the base calibration is done. I'd like to see if the MWV and VWR sentences sort themselves out afterwards.

You are having too much fun!

R

[tspeer]

I'm making some progress, but I'm not there, yet. This past weekend, I swung the compass (2 deg deviation) and calibrated the speedo (reading high, as I suspected). A 0.9 factor seemed to produce the best correlation between GPS velocity and boat speed while motoring in different directions near the head of Sinclair Inlet.

But the wind data recorded by the computer are still highly erratic. I'm beginning to suspect that the wind data sent by the NMEA interface and the wind data shown on the Dual Digital and Analog displays are not quite the same. I was thinking that the Dual Digital Display might have been the brains, with the Analog Display and NMEA Interface being basically repeaters. It makes sense for each unit to receive the raw apparent wind data from the masthead unit and then do its own calculations. So, for example, true wind speed on the digital display would be calculated independently of true wind speed displayed on the analog display, and the NMEA interface would do a third calculation. That would make it easy to mix-and-match displays in any desired configuration.

However, it also means that each unit has to have a good wireless signal from all the other units, not just, say, the nearest unit. I think I may be suffering from a lack of signal strength from the masthead unit at the NMEA interface. Not only is it shadowed by the mast, but it's under a stainless steel sink, too. That would explain the erratic wind data, when all the other data were solid. The Dual Digital Display will show the signal strength from each unit, so I should be able to use it as a sniffer to confirm my suspicion.

If that's the case, then I need to relocate the NMEA interface. I'm going to see if there's room to mount it near where the hull unit and compass are. I've already run DC power there, so I can put them both on the same circuit. That leaves the question of what to do with the GPS.

I could leave the MiniPlex41BT multiplexer where it is and hook the GPS signal directly to it. But that probably means not having the GPS data available to the instruments because of the return loop to the multiplexer. I'd have to be very careful about which sentences are exchanged in each direction.

I'll probably end up just extending the GPS cable and keeping the old architecture. I know that basically works, even if it may mean running some more cables.