Using two gps receivers to replace a compass

ianj99

ianj99

Well-Known Member
Joined
11 Nov 2009
Messages
2,097
Location
UK
Visit site
If you have two gps receivers mounted at either ends of the vessel, you should be able to generate the hdg sentence by comparing the positions and thereby use this in place of a fluxgate compass. This could be of particular benefit on a steel vessel.

Yes?

Would a pair of cheap 5hz usb receivers be accurate enough?

Where's Angus - its one for him I think.

Ian
 
Already done - you can buy a "GPS compass" which consists of a beam about three or four feet wide with a GPS aerial on each end. Rather specialised and quite expensive though.

Pete
 
Yep. Satellite compasses are the most accurate and use this principal. They are much better heading sensors than the best gyrostabilised compass and are used in the best autopilot systems.
Very expensive. GPS chips are cheap so if you can make a DIY version that would be great.
 
Last edited:
GPS compasses don't work simply by measuring a position at each antenna and calculating the vector between them. As well as the GPS code broadcast by each satellite, they track the carrier wave. By comparing the phase differences between the antennae for each satellite in view, they can calculate the azimuth between them.
 
geoid96 said:
GPS compasses don't work simply by measuring a position at each antenna and calculating the vector between them. As well as the GPS code broadcast by each satellite, they track the carrier wave. By comparing the phase differences between the antennae for each satellite in view, they can calculate the azimuth between them.
Click to expand...

Quite. So presumably not possible with two standalone units.
 
geoid96 said:
GPS compasses don't work simply by measuring a position at each antenna and calculating the vector between them. As well as the GPS code broadcast by each satellite, they track the carrier wave. By comparing the phase differences between the antennae for each satellite in view, they can calculate the azimuth between them.
Click to expand...

It is very possible to do this.
But normal GPS receivers just output position data.

I have played around a little with a couple of identical receivers. If they were 12m apart, and both had a good view of the sky it could work moderately well.
If something causes one GPS to look at ne cluster of satellites while the other is looking at others, like a big wet sail in the way perhaps, or a steel cabin, then the errors would be bigger.
Many modern GPS IC's now have inputs for inertial sensors, as these are used in car systems to help fill in the gaps in 'urban canyon' conditions.
So maybe integrating it with a gyro of some sort would be good?
A lot depends on what sort of 'averaging' you are willing to use, which is the flip side of how fast a response time you need.
You could do a lot of averaging if you were using it to 'swing' a mag compass.
A lot less if you were trying to steer to it.

There is a lot of tosh about tracking racing using GPS. It works very well on the Fastnet or AC boats, but its a bit rubbish on dinghies. The scale is too small and they change direction too quickly.
 
lw395 said:
It is very possible to do this.
But normal GPS receivers just output position data.

I have played around a little with a couple of identical receivers. If they were 12m apart, and both had a good view of the sky it could work moderately well.
If something causes one GPS to look at ne cluster of satellites while the other is looking at others, like a big wet sail in the way perhaps, or a steel cabin, then the errors would be bigger.
Many modern GPS IC's now have inputs for inertial sensors, as these are used in car systems to help fill in the gaps in 'urban canyon' conditions.
So maybe integrating it with a gyro of some sort would be good?
A lot depends on what sort of 'averaging' you are willing to use, which is the flip side of how fast a response time you need.
You could do a lot of averaging if you were using it to 'swing' a mag compass.
A lot less if you were trying to steer to it.
Click to expand...

The £30 Columbus 800 5HZ usb gps module I bought recently works well so I can mount a pair 10m apart so it would be intersting to compare the resulting heading vector with my Raymarine fluxgate compass. (its barely 2m above 10t of steel motorsailer).

I found references to a Kalman Filters for the averaging but this project is way outside my capabilites. Maybe Angus would do a YAPP: 2 nmea inputs and one output sending hchdg to see how it compares to the fluxgate...
Ian
 
ianj99 said:
The £30 Columbus 800 5HZ usb gps module I bought recently works well so I can mount a pair 10m apart so it would be intersting to compare the resulting heading vector with my Raymarine fluxgate compass. (its barely 2m above 10t of steel motorsailer).

I found references to a Kalman Filters for the averaging but this project is way outside my capabilites. Maybe Angus would do a YAPP: 2 nmea inputs and one output sending hchdg to see how it compares to the fluxgate...
Ian
Click to expand...
Must admit I was just cutting and pasting NMEA from hyperterminal into excel.
Do you know what chip is in the module out of interest?
There are some interesting magnetic compass sensor IC's about now, I got a few on ebay a while back, yet to do any work with them though.
You could mount a sensor much higher quite easily.
 
The best way to find out might be....

Get a pair of GPS Phones (Andriod?) write the software so that one broadcasts its GPS position to the other. Then the software on the other to receive the position and use it as a way-point to calculate the bearing bearing form its own position to the other GPS.

If the Phones where a known distance apart it might give some error estimation. If both did the same to each other they could be used to cross check the bearings between the positions?

Not saying I know how to do any of this far to clever for me...
 
ianj99 said:
The £30 Columbus 800 5HZ usb gps module I bought recently works well so I can mount a pair 10m apart so it would be intersting to compare the resulting heading vector with my Raymarine fluxgate compass. (its barely 2m above 10t of steel motorsailer).

I found references to a Kalman Filters for the averaging but this project is way outside my capabilites. Maybe Angus would do a YAPP: 2 nmea inputs and one output sending hchdg to see how it compares to the fluxgate...
Ian
Click to expand...
Kalman filtering would slow the results down to be unusable in any real world situation on any yacht; it's dreadful enough using them underwater for delay in keeping up. Isn't this a nice idea that's possible of course by having 2 separate units, but really impractical? After all, compasses work, right?
 
Last edited:
ianj99 said:
If you have two gps receivers mounted at either ends of the vessel, you should be able to generate the hdg sentence by comparing the positions and thereby use this in place of a fluxgate compass.
Click to expand...

I would expect that the accuracy of the received positions would not be sufficiently high to make it worthwhile, but that is just a suspicion. I haven't tried it. What I do know from seeing evaluations of commercially available fluxgate and gyro-compasses is that the accuracy you get is what you pay for. At the moment the best available for the leisure market is the one made by KVH. Some of the cheap ones are dreadful.

http://www.kvh.com/Leisure/Marine-Systems/Compasses/All-Compass-Systems/GyroTrac.aspx

The output from this will of course need correcting for deviation on steel vessel.
 
lw395 said:
Must admit I was just cutting and pasting NMEA from hyperterminal into excel.
Do you know what chip is in the module out of interest?
There are some interesting magnetic compass sensor IC's about now, I got a few on ebay a while back, yet to do any work with them though.
You could mount a sensor much higher quite easily.
Click to expand...

The spec of the module is here:

http://www.cbgps.com/v800/spec_en.htm

After engaging a few more brain cells, (another 4 of 10 available) I think 10m separation is nowhere near enough when the accuracy of the module is 2-8m.

A solid state compass module at the top of the mizzen mast might be a better way.
 
Why? when compasses rarely fail catastrophically and GPS's can and do in the marine environment.

Even a handheld walking compass will be better than a dead gps, so I can't see the point of replacing a compass with electronics. Supplement maybe.
 
Neil_Y said:
Why? when compasses rarely fail catastrophically and GPS's can and do in the marine environment.

Even a handheld walking compass will be better than a dead gps, so I can't see the point of replacing a compass with electronics. Supplement maybe.
Click to expand...

I was simply curious as to whether 2 low cost gps dongles could be used in place of my Raymarine fluxgate compass and I need the hchdg sentence for my old analogue AP. (Angus has already done me a YAPP that converts hchdg on Seatalk to 0 to 5volts which I need for it)

I do, of course, have a good old fashioned binacle compass as well: its not about a money no object alternative to this and my plotters.

Thanks for the replies.
Ian

PS in view of the $4199 cost of the SC50, I'm sure there is a market for a similar lower cost gps based compass for use on steel vessels
 
Last edited:
You must log in or register to reply here.

Share this page

Top