Unusual Drifting Behavior With 920T

[TomChat]

Got a 920T last week. At intersections, the 920 will sometimes "assume" that I have already made the turn and change perspectives, even though I haven't. For example, I will be facing north. The 920 will indicate, turn left. While waiting for the light to change, the 920 will suddenly go on to the next step of the instruction, as though I have turned, even though I haven't.

As far as I know, the unit is up to date. I don't know why it is doing this. It's not often, but when it happens, it is incredibly annoying and confusing.

 

[Boyd]

It happens to me sometimes too, and it happened even more often on my Garmin Nuvi. I think this is the price we pay for the SiRFstar III GPS chips and their ability to lock onto satellites so quickly.

Basically, I think what's happening is there's a sort of scatter pattern to the position data which comes from the GPS, and if you average the readings it's very accurate. But that can confuse the software about exactly where your car is if it's not moving.

 

[tomlouie]

Got a 920T last week. At intersections, the 920 will sometimes "assume" that I have already made the turn and change perspectives, even though I haven't. For example, I will be facing north. The 920 will indicate, turn left. While waiting for the light to change, the 920 will suddenly go on to the next step of the instruction, as though I have turned, even though I haven't.

As far as I know, the unit is up to date. I don't know why it is doing this. It's not often, but when it happens, it is incredibly annoying and confusing.

Do you have the GPS located deep within the car? If the unit doesn't have clear reception of the sat signals, it could throw off the positioning calculations.

 

[TomChat]

Do you have the GPS located deep within the car? If the unit doesn't have clear reception of the sat signals, it could through off the positioning calculations.

Not at all. It's pretty close to the rear view mirror. I wish there were a way to fix that, because it gets confusing.

 

[tomlouie]

Not at all. It's pretty close to the rear view mirror. I wish there were a way to fix that, because it gets confusing.

Hmmm. The times when this has happened, is it usually near when you just turned on the unit? Or does it happen even after the unit has been on for a while, like 30 minutes? The reason I ask is that once the unit has been powered up and had a chance to lock onto more satellites, the more accurate it will be.

My guess is that the GPS receiver is getting just enough signal to allow navigation, but not enough to have a solid definitive lock.

As boyd said, recent GPS chips are more sensitive, and will be more aggressive in allowing navigation in marginal signal reception conditions. In situations when it receives signals from fewer than 4 sats, the unit is essentially guessing where you are/where you're going.

 

[bseidner]

this problem too

I have seen this problem as well, but when standing still no matter how many sats it sees. I think the solution is to have the TT sample the sats more often. I think the TT polls the GPS chip at a low rate, maybe speeds up as movement starts to occur. I am looking for a config file in Linux to increate the sample rate, although this will affect power consumption. Should not be a problem if the unit is plugged in.

 

[terbolee]

I had the same problem just yesterday when I tried to get to Sydney's Luna Park. With tall buildings all around, and with cloud cover, my guess is that my TomTom simply could not get a good fix on the satellites.

I'm probably guessing but my old Navman PiN 570 did not have the problem because I got an external antenna for it.

I might have to get an external antenna for my TomTom.

Regards

 

[mrlong]

Mine's done that too and it's sometimes confusing. I wonder if the motion sensor might have something to do with it. If you make a sharp brake or excelloration, do you think it might fool it somehow? The motions sensor is used for going through tunnels to keep it somewhat accurate as to position.

 

[Boyd]

I also think it happens at times when you have a perfectly good satellite fix, and that it's just part of the SiRFStar III chips. I have a Garmin 60csx handheld with the same chips, and you can zoom WAY in to the 20 ft level on its screen. When you do that and turn on tracking you see a lot of drifting even when you have a good signal. The samples look like a shotgun blast and are scattered pretty randomly within a certain radius. But if you were to average them all, you would have a very accuate fix at the middle of the pattern.

I think that this can confuse the TomTom software when you're standing still, and it thinks you may have turned 90 degrees. Like I said, my Garmin Nuvi 650 did the same thing with the same regularity.

 

[TomChat]

Hmmm. The times when this has happened, is it usually near when you just turned on the unit? Or does it happen even after the unit has been on for a while, like 30 minutes? The reason I ask is that once the unit has been powered up and had a chance to lock onto more satellites, the more accurate it will be.

My guess is that the GPS receiver is getting just enough signal to allow navigation, but not enough to have a solid definitive lock.

As boyd said, recent GPS chips are more sensitive, and will be more aggressive in allowing navigation in marginal signal reception conditions. In situations when it receives signals from fewer than 4 sats, the unit is essentially guessing where you are/where you're going.

Wow, that was a good diagnosis. It's happened both ways, but more often, it will happen when I start up the unit and have been driving for less than 5 minutes. I guess it's a combination of a "cold start" satellite acquisition and something to do with the SiRF chip set itself.