Track data recording

[timurrrr]

Originally Posted by X25

GPS speed also goes a bit wrong when you're taking a corner (since you're no longer on a straight line)

GPS data of course can be slightly off when not going at constant speed in a straight line, but...

Looking at your video, it's obvious that you simply have the GPS data not aligned with the video correctly. Look at the gear shift at 0:18 in the video. The GPS speed keeps going up for a third of a second or so, which is of course incorrect. The OBD speed channel though is well aligned with the video and the speed briefly stops going up as you shift.
The GPS-vs-OBD offset will become even more apparent if you plot those two speeds at the same time.
Good news is that RaceChrono provides an ability to easily offset any data source, you can find it in the devices/channels list in the "Analysis" mode. I suspect you need to delay your OBDLink MX+ data by ~350 ms.
Once you align the OBD data to the GPS data, try re-making that video and make sure to re-align the OBD+GPS data to the video.

I'm pretty sure the scale of this timing problem is orders of magnitude larger than the scale of the theoretical "GPS speed is incorrect when cornering" you mentioned.

Originally Posted by X25

so you should use this data for speedometer anyway.

There are many real-life examples where this statement is incorrect.
Try doing a burnout and you'll see 40 mph on your speedometer while the car is clearly standing still.
Drop the brake to start accelerating, and gently go back to moderate throttle — you'll see the speedometer data doing DOWN while the car is still accelerating.
If you're driving fast enough for tires to start micro-slipping, the speedometer will read low as you decelerate, and high while accelerating.

If you want to compare how fast you're going between two laps, using the speedometer data can result in wrong conclusions.

GPS speed has none of those effects. It has other bad effects, e.g. when driving under a bridge. But as long as the bridge is not in the middle of the corner, and you don't have other GPS issues (e.g. bad mounting location), comparing GPS speeds between two laps should be apples-to-apples.

That being said, you should really log both, not just say that one is bad and ignore it. When one of them looks suspicious, look at the other one for sanity checking.

[X25]

Based on your responses, I'm guessing you think I'm bashing GPS over OBD speed data? That's not my intent; they're both very valuable.

I log them all, and use whatever works best if I need to analyze something. I already 'synchronize' the GPS data with the OBD data, but even when you make it perfect, the GPS transponder will sometimes 'delay' the update. In my experience, compensating for that delay often created negative artifacts elsewhere.

You should definitely use OBD speed data for the speedometer gauge, since even when perfect, GPS speed will be very much off in any long turn. I care about my speed mostly at the entry to, or at the end of straightways (where GPS is correct), and care about my speed data for the whole corner. Using only GPS data would be very misleading in long sweeper corners, and you could have wrong conclusions : )

[jflogerzi]

Look at race Chrono pro. Use external gps and get the recommended ob2 reader and you can tap into CAN bus and get data much quicker. This combo is much cheaper that other things on the market and works well for timing, data logging and analysis.

Sent from my Pixel 5a using Tapatalk

[X25]

Originally Posted by jflogerzi

Look at race Chrono pro. Use external gps and get the recommended ob2 reader and you can tap into CAN bus and get data much quicker. This combo is much cheaper that other things on the market and works well for timing, data logging and analysis.

Sent from my Pixel 5a using Tapatalk

Agreed!!