Autosport Labs

i experience the same issue as toga with the distance reading.

not sure if this helps, but one thing i noticed on v2, my RPC's LED for GPS doesn't blink at the same rate it did when GPS lock was made on v1.

LED1
Indicates GPS Activity. Slow flash (1Hz) indicates GPS is actively acquiring. Fast flash (10 times per second) indicates GPS is locked on to satellites.

on v2, my RPC's LED for GPS doesn't blink at the same rate it did when GPS lock was made on v1.

That is expected. We should have noted that in the release notes.

Based on the files uploaded by Toga94m I can verify that the GPS data is in fact valid. To prove it I imported Toga's data into a spreadsheet and projected laps 7 and 8 onto a map. You can see his driving right along old post road and it flows nicely. So we can now safely eliminate incorrect or weird data as our culprit. I continue to dig...

Tonight I updated to 2.7.7, same results. I didn't expect any change since that part of the code doesn't seem to have been touched in this update, but thought I'd keep current anyway.

Tonight's drive started logging on sector 0 (not -1) and never moved up; sector time was at 0.0017 minutes and never changed. Interval started at 24300mSec so I guess I let it run for a while before hitting the LOG button. Lapcount started incrementing almost immediately, even though I hadn't crossed start/finish yet. When lapcount went from 0 to 1, LapTime = PredTime went from 0 to 0.1683 (10 sec).

There are some new random numbers in DISTANCE too, and more variety than usual. I've been hoping they'd be something like 65536 or another obvious math thing, but no such luck. Doesn't even look like it's related to earth's radius or one of the other big constants in the code.

These are all claiming to be in miles of course..
45127.539 45136.82 45136.18 3735928559 49023268 316226.469 -48694988 -49010056 -45134.773 -90269.477 -49099300 -931.552 49020928 931.552 85476224 -49011468 1083683.75 1084615.25 1083683.75 316175.688
I guess since these are decimal representations of floating-point binary random numbers, it's hard to tell what it's really saying.

I plotted the Lat/Long as an x-y chart, and it accurately shows my path around a few blocks in the neighborhood, so still getting good GPS data. Logfiles shows 6-8 satellites with the sensor on the passenger seat, so a pretty good fix.

So, I'm looking at this right now.

I thought I was able to reproduce it once statically on the bench, and now I cannot.

In order to see this behavior, do you have to cross the start finish line once before it starts happening?

Nope. The latest log I ran was already upcounting laps before I got out of the driveway, starting the log at Interval 24300 (24 second after powerup). By the time I first crossed start/finish, it was up to lap 22

The previous two logs I've posted show the sector at -1 until my first start/finish crossing, then sector goes to 0 and the upcounting begins.

We figured out what the problem was. We believe it's centered around the fast square root routine. It seems the toolchain for MK1 is calculating this incorrectly, but it's working correctly for MK2. We were debugging the root cause last night.

Good news ultimately, as we can quickly narrow down on the source of the problem.

Hello all,

we have a preliminary fix for the MK1 firmware. could you test this out before we release it officially? Please let us know asap. thanks!

http://autosportlabs.net/asl_dist/rcp/f ... _2.7.8.zip

It's flashed.. will try to do a few rips around the racetrack on the way to work this morning. We got a few more inches of snow, so cornering g's will be lower than usual.

It works! MK1 2.7.8, latest beta android app via bluetooth. Distance counts up once I've crossed start/finish, and lapcount increments as expected. My 3rd lap was slower because I was following a snowplow slaloming around garbage cans at the curb. And I got blue-flagged at an intersection and had to wait. Logfile available later today if you need to see it.

can confirm on the momostallion test track as well.

will try with coordinates instead of prebuilt track again later today.

Thanks for the update! Looking good so far.

Tested this afternoon with auto-detect turned off, just coordinates stored for the start/finish line 43.0957,-73.747. It worked great, and I didn't get arrested for suspicious behavior. GPSVisualizer.com provided this map from the RCP logfile data, with GPS coordinates / track colorized by lap count.

As expected, if you take too long in the start/finish line (like, 3mph) the lap counter will tick up one more than it should. But how often are you flying across start/finish at 3mph.. not a real-world problem.