Inconsistent GPS update rate on XGPS160

[spaRTan]

After having successfully used HLT with a XGPS150A for a long time I upgraded to the XGPS160.
Now I'm getting weird results and massive "lag" in the velocity graph.
The faster the acceleration the bigger the lag.

Could it be that HLT cannot process the information fast enough?
I haven't tried other tools yet and could do so. But maybe you have a hint for me.

Thx,
Robert

[Harry]

Hi,

To start with the processing question: no. LapTimer can easily process much higher update rates. Even better, as each GPS position comes with a time stamp, runtime effect can be completely eliminated by LapTimer. As an example, I use a NMEA stream recorded in 2006 for LapTimer regression tests. LapTimer does not care the stream is 8 years late and simple records a lap for 2006.

But maybe we start with the observation - what do you mean by lag and how is it related to the plots attached? GPS speed reported is always 1 to 2 seconds late when watching it in real time. As written above, this effect is compensated when storing and processing the data.

- Harry

[spaRTan]

Please have a look at the second acceleration graph of the Porsche Turbo.
Speed isn't rising consistently as the car accelerates but rises, stays, rises again and so on.
This is (of course) not how my cars accelerate at all.
Thus there are only few options left:

1. The XGPS160 offers less quality GPS data in exchange for a higher update rate

2. Something makes HLT process velocity in some kind of "chunks"

As said before I'll try the logging option of the XGPS160 without HLT connected so I can tell you exactly whether it's the XGPS or HLT. I still have my 150A and no problems whatsoever...

[Harry]

Except for interpolated positions (start / finish / split), LapTimer stores data *unchanged* when connected to an external sensor. So GPS speed data stored and shown in LapTimer is always raw data received from the sensor (except for the named exceptions). So #2 isn't it.

A different thought (#3): the graph looks pretty smooth, but speed comes in as 10 distinct values per second. I don't know which program you use to display the graph, but maybe this are some spline interpolations used by the viewer? Do you have a plot where we can see the discrete speed values?

- Harry

[spaRTan]

Thought #3 was my first one as well.
As you can see the original screenshot from HLT shows the same behaviour. At first i thought it's a problem of bad GPS quality or dropouts but it's always there (I had 13 satellites and a perfect signal position-wise).

The graph section of born2race uses all the ticks' values in the graph. No further interpolation (if there are very few ticks the line isn't that smooth any more) is done.

Could it be a faulty device? I've tried Peteolhead on my iPad Air and Rookie on my Nexus 5 - the result is the same...

[spaRTan]

Today I updated the firmware of the XGPS160 to 1.2.
Sadly it didn't improve the situation.

I took comparable logs using HLT Rookie and RaceChrono on my Nexus 5 (see screens below).
Conclusions:

1. The velocity sent by the XGPS is crap in both apps.
2. In RC it's a bit more smooth but way too bad to use it seriously.

Maybe I'll have to go back to the 150 for acceptable results or I get a new 160 and hope this one is OK...

[spaRTan]

Any ideas now that you've seen it on the HLT-screenshot?
The peaks are rolling in every 1 second all time long.

[Harry]

I have contacted Dual on this effect on Sunday. Please give some time so they can clarify on it. It seems I can't do too much on this.

- Harry

[spaRTan]

Thank you Harry.
I wanted to contact them myself tonight - now I won't have to.

I'm very eager to receive their opinion on this topic.
I don't understand how everybody else can use the device without any hassle.

[jwmelvin]

Any update on this (now that I have an XGPS160 on the way)?