I've gone over Lee's files, and I can make a couple of
educated guesses as to what's happening.
To get a good looking map, Lee used -0.014g adjustment to
end speed, and 0.068 adjustment to lateral, which increases
left turn. Both of these adjustments seem to be the right
adjustment for the course, as the result of both end speed
and course shape look about right.
With this all done, the absolute peak left is 0.85g, while
the absolute peak right is 1.13. This certainly does make
things look like the left g's are underdeveloped vs the
right g's.
So we can look for the easiest explanation for this
first--the car doesn't pull as many g's left as it does
right. While that's possible, it doesn't explain the
required 0.068 adjustment left required to make the map look
good. As most everyone here will agree, 0.068 adjustment is
a huge adjustment!
We know that the G-Cube will read 1g in each direction.
Each cube reads 1g slightly differently, and the purpose of
the calibration is to set your individual cube to read the
value correctly at whatever level it requires. In other
words, if a Cube were to for some reason be twice as
sensitive to right g's as left g's, the calibration would
also show the result, and hence the outcome from the
software would still show 1g each direction. In fact, cubes
generally vary in their left and right g's by as much as
about 5%; the calibration routine keeps all values right in
balance.
One way to test that is to "drive the desk". Go to
recording options, and set your roll and pitch to 0 (or turn
off roll and pitch compensation). Set minimum run to 10,
and autostop duration to 5. Now go to RECORD, set the cube
flat and press AUTOSTART. Simulate a start by setting the
cube down on it's back side, so that the front side is up.
Hold this for 3 seconds. This simulates a 1g launch for 3
seconds. (Wow, what a car!) Now switch to left side up for
3, then right side for 3, then back side for 3, then set the
cube flat again until it stops. When you pull the run up
and look at it, you should see an S shaped run, and the
charts should show 1g in each direction left and right.
You'll also see 1g levels in gas and brake. I would suggest
doing this test just to be sure that everything in your
calibration is right. If you get good one g levels
everywhere, then all is fine. If not, perhaps you should
recalibrate and try again.
Because this is essentially the same process as calibrating,
it should obviously produce readings which are the same as
in the calibration, and hence, they'll be 1g levels in all
directions. If you find that your first test produces small
left g readings, then yes, that was the problem. So
recalibrate and try again. (Be sure to set your recording
options back where you started when you are done, btw.)
But if the "zero calibration", that is, the level point at
which we first press autostart, if that were off, we'd
certainly see some offset in the g's. So this is even more
likely. The zero calibration takes place each time you
press autostart or record. It's the little half a second
"CALIBRATE" that you see in the status window. In that 1/2
second, the cube takes 10 readings, averages them, and calls
the resulting numbers 0,0 for lat and lon g's. If the car
is on a slope when this happens, the slope becomes 0,0, and
when we return to level, we have a built in bias in some
direction. This is a very possible scenario. If Lee's
AutoStart was set with the car parked on a lateral slope,
with the left side of the car high, the result would be
pretty much what we have here. How much slope? To produce
0.068g, the slope would have to be about 4 degrees. That's
a pretty good tilt, but certainly something that you could
inadvertently do.
The solution to this is simple of course, be sure the ground
you are on is level when you press AutoStart. If it's flat
enough that the car won't roll, no matter which way it's
aimed, it's fine.
If that's not it, I come back to the position and stability
of the cube. The velcro is usually not my favorite mounting
method. Use good ol' racer tape or duct tape if you can,
simply taping the G-Cube to a firm level surface. If the
velcro is allowing your cube to tilt, again by only 4
degrees while you are driving, you can see how easy it would
be to not get accurate readings. Additionally, the cube
MUST be facing straight in line with the car. If your
heading is off by a few degrees you'll also get some strange
"off center" readings.
And finally, it's not common, but a low battery can produce
some flaky readings too. So check the battery in your power
supply. And don't use PC power to power your G-Cube.
Battery power is much cleaner and steadier than PC power.
Perhaps one or more of these items will strike a chord with
you as you read it, Lee. With this all said, though, I want
to congratulate you on your adjustments. You did a fine job
of adjusting your map. The resulting information will
certainly help you to find spots to improve your driving,
and should will provide a wealth of info for you. Getting
your maps to come out of the G-Cube a little better right
away is usually just some simple basics. But no matter
what, you seem to be able to adjust the maps to look great.
--Byron
800-775-9511
Lee Witkowski wrote:
>
> Arthur,
>
> I just sent the zip file to Byron and cc'ed myself. I received it "just
> fine." Your analysis of the Team.Net server problem appears to be "spot
> on."
>
> Thanks for your help.
>
> Does anyone else want me to send this file to them for analysis? Steve?
> Mark?
>
> Lee
>
> Arthur Emerson wrote:
>
> > Lee Witkowski <lwitkowski@ameritech.net> wrote:
> > >
> > > Well, the single file method didn't work. So here's the zipped file
> > > method.
> >
> > I think that the problem is with Majordomo on the new Team.Net
> > server. I don't see a content-type or encoding header line
> > in Yahoo mail, which could mean that Majordomo is stripping
> > them off. (It could also mean that Yahoo mail is brain-dead,
> > but I'm leaning towards Majordomo.)
> >
> > The solution is simple. Just send the files directly to
> > Byron, not the list. If that fails, somebody should start
> > a GEEZ group at e-groups.com. They include file space to
> > exchange uploads. Another solution is to open up a free
> > Yahoo briefcase and upload the files there. I use it
> > all the time, and you can feel free to check out some
> > pictures of the "delete paint option" on my car at
> > http://briefcase.yahoo.com/vreihen .
> >
> > I'll spare you my rant about how e-mail is not supposed
> > to be a file transfer medium, and e-mail file transfers
> > are responsible for half the ills on the Internet today.....
> >
> > -Arthur
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Send instant messages & get email alerts with Yahoo! Messenger.
> > http://im.yahoo.com/
>
> --
> ============================
> Lee Witkowski
> President, Wit-Arts, Inc.
> 1359 State Street
> Lemont, IL 60439
> Voice: (630) 257-5025
> lwitkowski@ameritech.net
> ============================
|