Data Alignment Issues

turfburn
Posts: 11
Joined: Wed Apr 20, 2011 3:45 pm

Data Alignment Issues

Postby turfburn » Thu Feb 16, 2012 5:38 pm

I sent in a request to the US Support contact, but figure I may get additional points from some of the regulars on here. So minor apology for the two pronged push for information.

I've been trying to do a lot of positional analysis of my autocross runs to see variances in turn in location and braking etc to determine how much they are affecting the run performance. However, the alignment between runs on the data is all over the place for some reason. I am trying to align the graphs based on distance as that should theoretically show if braked sooner or later based on position relative to the corner. However, I have had very little success doing this. I'll try to get screen shots to add to the post to help explain it better, but basically even though the data is aligned that it believes I'm at ".36 km" at a given spot for example, the data is clearly not really at that spot, it shows positions of turns being 100 feet different from one run to another based on the lateral and longitudinal forces and that really isn't possible.

Any ideas/thoughts on what could cause this? It improved a little with the option to correct GPS position data enabled, but it is still ugly. The GPS receiver is on the top of the car completely clear of obstructions so it should be a solid signal, and it usually flashes that it has at least 5 satellites locked, so I'm unsure what could be the issue here - the raw data, or something in the analysis.

Thanks!
Steve

Support

Postby Support » Thu Feb 16, 2012 5:46 pm

It sounds like the GPS data just isn't very good - I would suggest that you send a run file to the US office for him to look over. The most common reason is that the GPS antenna is starting to fail.

Regards, Support

Rob Stevens
Posts: 36
Joined: Sun Oct 31, 2010 9:36 am

Postby Rob Stevens » Fri Feb 17, 2012 4:23 pm

Try enabling the wheel speed option, this should help the speed/distance/gps accuracy, however I've only noticed a slight improvement, also check to see if vibration might be corrupting your acceleration data. I assume autocross is like our sprinting, one car at a time not on a circuit, a sprint?

turfburn
Posts: 11
Joined: Wed Apr 20, 2011 3:45 pm

Postby turfburn » Sat Feb 18, 2012 4:27 am

Support wrote:It sounds like the GPS data just isn't very good - I would suggest that you send a run file to the US office for him to look over. The most common reason is that the GPS antenna is starting to fail.

Regards, Support


I see this issue in the data back to day 1 of the equipment (about 2 years ago). We also have been meticulous and not pinching or abusing the antenna. I'll try to get some files to the US office and see what he can get. I thought there was something about a positional accuracy value that is in the file that someone mentioned to graph to see what it thinks.

osborni
Posts: 497
Joined: Fri Dec 07, 2007 9:08 pm
Location: USA, Michigan

Postby osborni » Sat Feb 18, 2012 2:19 pm

I had an antenna fail with no visible damage to the wire....
BMW 2000 M Coupe

turfburn
Posts: 11
Joined: Wed Apr 20, 2011 3:45 pm

Postby turfburn » Sat Feb 18, 2012 4:31 pm

osborni wrote:I had an antenna fail with no visible damage to the wire....


is there a reasonable way to test if the antenna is okay?

turfburn
Posts: 11
Joined: Wed Apr 20, 2011 3:45 pm

Postby turfburn » Sun Feb 19, 2012 4:17 am

I pulled up some data, I need to dig through more, but I'm seeing positional accuracy values provided between 2.52 and 3.4 on average with a spike here and there to 4 or 5. I'm assuming the units are meters. Not sure if that is good/bad or average.

Thanks!

jlucas
Posts: 302
Joined: Tue Nov 29, 2005 12:45 am
Location: East Liberty, OH
Contact:

Postby jlucas » Sun Feb 26, 2012 1:47 pm

That sounds like normal averages to me.
Jeremy Lucas
Fast Tech Limited - a Race Technology Dealer

turfburn
Posts: 11
Joined: Wed Apr 20, 2011 3:45 pm

Postby turfburn » Sun Feb 26, 2012 5:20 pm

jlucas wrote:That sounds like normal averages to me.


That's what confuses me then if the positional data is reasonably accurate per that data, yet the distances items don't line up at all in the data. Off by much more than several meters.

That said, if the 3-6M is normal I'm not sure that I will be able to see what I want then any way as I need accuracy probably closer to 1-2 M to gain anything useful. I think I'll wire up a sensor for the brake pedal and overlay that in to the data and that would give me a better indicator for brake position. Also would be interesting to wire in steering angle. Can I get those off the OBDII maybe and just add that instead??

jlucas
Posts: 302
Joined: Tue Nov 29, 2005 12:45 am
Location: East Liberty, OH
Contact:

Postby jlucas » Sun Feb 26, 2012 5:47 pm

I'm not sure this right term but I think of that as an absolute GPS accuracy, the run to run accuracy is greater than that.

I don't have personal autox experience with a DL1 so I can't be much further help, but you can move the data in the run manager so that your start & end points line up between runs to account for any GPS drift (same goes with road race stuff).

You cannot get steering angle off of OBD2. You need either the manufacturer CAN code (and CAN adapter) or a dedicated sensor.
Jeremy Lucas
Fast Tech Limited - a Race Technology Dealer

turfburn
Posts: 11
Joined: Wed Apr 20, 2011 3:45 pm

Postby turfburn » Sun Feb 26, 2012 5:49 pm

jlucas wrote:I'm not sure this right term but I think of that as an absolute GPS accuracy, the run to run accuracy is greater than that.

I don't have personal autox experience with a DL1 so I can't be much further help, but you can move the data in the run manager so that your start & end points line up between runs to account for any GPS drift (same goes with road race stuff).

You cannot get steering angle off of OBD2. You need either the manufacturer CAN code (and CAN adapter) or a dedicated sensor.


Sounds good. I think I can build up a steering angle sensor off a potentiometer and some rigging. Can't easily/directly access the stock steering angle sensor on the car. So I'll try to rig that and a brake sensor and see where it gets me.


Return to “General software support”

Who is online

Users browsing this forum: No registered users and 4 guests