What to make of this? Erratic Compass

[ZeMartelo]

Today while I was out, there was a cache nearby that I had found before and decided to make a test with the Colorado.

Compass is set to Auto, Waas enabled (although no sat were corrected on the sat screen), accuracy was 6meters.

 

I selected the cache and switched to the compass screen.

 

One of the map fields is set up to display the GPS coordinates (I wish there was an option to also show the destination coords).

 

The red pointer pointed me in the right direction (cache was approx 150 meters away).

 

At approx 15 meters from the cache the pointer started to swing and not point to the cache anymore.

 

I stopped when I was approx 7 meters away from the cache on my left.

The pointer was basically pointing 45deg to my right while cache is 90degrees to my left.

 

I calibrated the compass at that spot, and even turned it off, and got the same result.

Moved back and forth a few meters and the pointer would always point in the same direction.

This means that at this point I could not rely on the compass anymore so I decided to look at the GPS coords instead.

I had written the cache coordinates on paper and decided to walk towards the coords using the coords on the data field instead of relying on the pointer.

The result surprised me. Using the coords on the data field it put me approx 1 meter away from the cache and to find it only needed to move my head to spot the cache, the coords were spot on and they matched the cache coords, yet the red pointer was pointing the other way all the time??

I just now realized that I should've walked in the direction of the pointer to see how far it would take for it to point again towards the cache...

 

How does the Colorado calculate the pointer position? It seems obvious that it is not using the displayed coords on the screen to calculate its position but some other data.

It was telling me that the final coords were South while the GPS coords were sending me North?

 

How much of a factor would cold be to skew the pointer? It's winter and its pretty cold.

BTW this pretty much applied to the 60Csx that I had as well, and thats why I sold it.

[+LostMontanan]

I can't answer your question, but my 300t does this as well. I came from the eXplorist 500 and the compass would walk you right to the cache. The 300t doesn't do this. The only way I know of to remedy this is to watch the numbers decrease as you get closer to the cache. (I have it set to Distance To Next or similar) It is frustrating, because the compass is the easiest way to find caches, but with this unit it is useless. Repost here if you find a solution and I'll do the same.

[+myotis]

Do not forget the cache's coordiantes will be off and so will your GPS. GPSs have errors so you cannot assume the cache coordiantes are correct. GPS errors can cancle each others out or make it much worse. So the trick to finding caches is use your caching sense when you get close. Once you get close, paying too much time looking at your GPS instead of the area makes it take longer to find the cache.

 

On my 400T sometimes the arrow gets confused so I just watch the distance when it is pointing the wrong way.

[ZeMartelo]

Do not forget the cache's coordiantes will be off and so will your GPS. GPSs have errors so you cannot assume the cache coordiantes are correct.

I know that, but in my test the cache coords were actually pretty close to the actual cache location.

 

The point is that while the compass pointer was pointing in the other direction all the time, walking towards the listed cache coords using the GPS coords took me right to it.

The compass pointer should be using the gps coords and calculate the bearing to get to the final coords.

Its not doing that, I dont know what it is doing.

[+twolpert]

This is characteristic of high-sensitivity receivers, although the CO is worse than some. The basic problem is multi-path reception. Multi-path occurs when satellite signals bounce off nearby surfaces such as foliage, steep hills, buildings, etc. The receiver sees multiple instances of the signal from a single satellite. Each instance has travelled a slightly different distance from satellite to receiver. The unit makes its best guess as to which signal is the "real" one. From one sample to the next, it may switch from one signal to another. Usually, more than one satellite is involved. The net result is sudden apparent shifts in postion, 30, 40, 50 feet or more. If you were to examine the track leg corresponding to the shift, you would see very high apparent speeds -- sometimes as much as a few hundred miles an hour.

 

Unfortunately, on the CO, these quick shifts also interact with the "bearing to next" as shown by the pointer. Remember, the pointer represents the direction you should go to reach the next WP. So the pointer takes into account your current position, the position of the target WP, AND the direction of travel. With the compass in auto mode, if you are standing still or moving below about 10 MPH, the unit uses the compass to determine direction of travel (or the direction you are facing if standing still). Above 10 MPH, the unit uses direction of travel as determined from satellite data.

 

So, as you near the cache in a multi-path situation, you get a perfect storm of adverse conditions. Your apparent position is moving, perhaps even putting you on the other side of the cache from time to time. More importantly, the unit is shifting back and forth between compass heading and bogus satellite direction of motion information. So the compass swings.

 

In most cases, the "wandering" motions are very fast and converge on your real location. If you look at your track, you will see a characteristic "snowflake" or (after a while) "ball of yarn" track around your real location. The excursions take place quickly, so you are not likely to see them if you are watching your current coordinates. Also, the unit makes some attempt to "average" your current location -- so the coordinates are unlikely to vary as much as the track.

 

This is why some cachers recommend disabling the electronic compass. And why most experienced cachers recommend switching from GPSr to eyeballs when you get close to the cache.

[ZeMartelo]

Twolpert thank you very much for your explanation.

It is so simple that even I understand it and it makes total sense.

 

You are right about snowball effect, I've left my GPS on all afternoon and it is visible.

The coords as you say are more stable, they barely change, why doesnt the unit use it in its calculations for the pointer?

 

I am going test the unit with the electronic compass off to see if it improves accuracy.

I've done a few tests in my driveway (I know very scientific tests) and I liked how it performed with the electronic compass turned off.

[+Prime Suspect]

The navigation arrow was simply reporting the normal GPS errors. The further away from the cache you are the less the error seems, because the angle of change is much less. The closer you get, the greater the angle change. You would see the same effect if you used a magnetic compass, and were able to take readings as fast and as often as your GPS can.

 

However, one drawback that the newer units have, that the 60CSx doesn't, is the compass error message. With the 60CSx, if you're not holding it level (required for an accurate reading), a HOLD LEVEL error message is displayed. The newer units will not warn you that you're holding it incorrectly and that the reading is not reliable.

[+twolpert]

I am going test the unit with the electronic compass off to see if it improves accuracy.

I've done a few tests in my driveway (I know very scientific tests) and I liked how it performed with the electronic compass turned off.

It won't improve accuracy (in the sense of giving you smaller estimated or real position error). But it will cause the pointer to settle down somewhat. The drawback is that when you are standing still, the unit will continue to use the last known direction of travel from the satellite data. And, if the unit is "wandering", you will continue to see some shifts in the pointer.

[ZeMartelo]

However, one drawback that the newer units have, that the 60CSx doesn't, is the compass error message. With the 60CSx, if you're not holding it level (required for an accurate reading), a HOLD LEVEL error message is displayed. The newer units will not warn you that you're holding it incorrectly and that the reading is not reliable.

The Colorado does show the "HOLD LEVEL" in the compass page. Its a pain to get rid of it too.

Edited February 16, 2009 by ZeMartelo

[ZeMartelo]

I am going test the unit with the electronic compass off to see if it improves accuracy.

I've done a few tests in my driveway (I know very scientific tests) and I liked how it performed with the electronic compass turned off.

It won't improve accuracy (in the sense of giving you smaller estimated or real position error). But it will cause the pointer to settle down somewhat. The drawback is that when you are standing still, the unit will continue to use the last known direction of travel from the satellite data. And, if the unit is "wandering", you will continue to see some shifts in the pointer.

From the small tests that I have done the pointer seems to be more accurate and keeps pointing in the right direction till it reaches the destination coords. That doesnt happen with the electronic compass on.

The drawback of course is that it requires movement and I am not sure just how sensitive the unit is to detect movement.

[+Cache Viking]

My 400t did the same thing. I turned off the electronic compass and have not had the trouble since. Only problem with the pointer these days if I get out of the car with the cache like 20-30 feet away. With that the pointer again does not want to point. If I take a walk for about 50-100 ft away then head back it hooks up and works fine.

[+qlenfg]

If it hasn't been suggested already, make sure you have the latest GPS chipset firmware and GPS firmware. The compass issues have been reduced quite a bit from the early firmware.

 

I usually keep my compass turned off for the reasons you saw. However, the Colorado sometimes gets 'behind' when you start getting close to the cache, so you walk past the actual location to a position 20-30 feet away, which the GPS says is ground zero. Within 30 seconds or so the GPS will catch up and show you are off by that distance.

 

What I do when I'm still several hundred feet away from the cache is guestimate where its going to be, then keep going until the pointer shows I'm right on it. If I think the position is off, I'll continue another 20-30 feet, then walk back slowly, watching the direction and distance. Sometimes going out and back in one of the cardinal directions will help.

 

The satellite constellation and the amount of tree cover or terrain cover seems to affect this quite a bit.

 

As far as being stopped and the pointer showing the last direction it thought you were headed, just keep tabs on where north is and use the north indicator on the screen to orient the GPS so the indicator is pointing the right way. Saves a lot of heartaches!

Edited February 16, 2009 by qlenfg