Ok, been a while so I figure what the heck, lets go for a kinematic session with a pair of the Locii. The players: 2 Locii, base on a tripod/tribrach and we gonna use the bar, rover goes to the 2 meter pole and we're controlling it all with Survey Control II on the iPAQ 3150. Oh, and the triple mag mount for the roof of the Subie for the "continuous kinematic" part.
Wide open area. In over 2 hours of "working" I never lost lock. Set both up for 2 second epochs, time on each stop and go point set to 12 seconds. I know, 12 aint that sporty but it's better than 10. Located edge of pavement in S&G mode, then on to the mag mount and lets drive the centerline and record a spot every 2 seconds, set a few new points and observe with bipod for 5 minutes each (these worked out Swell), hit an NGS benchmark there on the site---everything goes really smooth.
Processing. AS 2.7 with the Locus download patch. One Locii a bit cranky on the download but finally gets the job done. D file out of the iPAQ goes in without a hitch. Loosen up the "desired accuracy" because I know those ones in continuous mode aren't going to be so sporty.
Process all.

Oh, boy is this ugly! 2 sigma aint gonna happen. 1 sigma everything looks OK except 3 or 4 points in the S&G mode----floats. Big time floats. Working with right click on the floaty vectors and raising the mask does nothing but process the entire file.

And whats up with my observations? I set the S&G for 12 seconds. The counter on the iPAQ displayed the countdown each time as 12 seconds. However, the observations list the session as anywhere from 6 to 12 seconds and one of the floaters is 14 seconds?? Most were in the range of 8 or 10 seconds.

Thoughts?

Dangerous Dave

Dave

It's just the nature of the beast to occassionally have a couple of total crap shots when doing S&G. I would say percentage wise on a good day, with decent constellation and workable pdop, around 3% may be "suspect". Even these can be eleminated with enough shots to allow for a few bummers.

Also, raising the mask affects every observation in the S&G file. It ain't nothing but one long session, so you can't tweak a single point in the session.

Can't offer anything concerning the way the display is acting regarding the observation session counter. Sounds a bit flaky though.

jd

ps....you really ought to consider standing still for more than 12 seconds at a time :-)

Are you sure that BOTH, your data collector(HP?) was to collect 12 seconds of data?

The locus doesn't keep track of the data timing . . . it's basically along for the ride.

As far as the Locus is concerned(even the rover) this is just one big session. It doesn't know(or care) that it's moving around.

Maybe the synchronization(sp) was off by 8 seconds, leaving only 4 seconds for observations.

And if you don't HAVE a kinematic controller, there's one for sale on Ebay.
Modified By Jerry Anderson on 10/20/2004 at 10:52 PM

Actually Jerry, this particular controller and card was purchased from Hayes. I've got another in a 3835 iPAQ. The 48GX as well if I have to go back to that route.
Yes John, both receivers were set to log 2 second epochs. The only thing I can think that may have "corrupted?" the D file would have been the iPAQ turning itself off after so many minutes. This it did during the 4 sessions that I observed 5 minutes each on. When turned back on the program continues and the countdown looks reasonable.
Ebay. Nothing there tonight but some old obsolete broken M's. Have at 'em.

Dave

Same thing happened to us with the Sokkia Stratus, and we always thought the best thing to do was RAISE the mask.

The Stratus logs with default mask of 5 deg, and so all the fixes were based on THAT mask. We raised the mask from 10 to 20 to 30. Nothing!Floats everywhere.

As soon as we put it back to 5 deg we got all fixes!!

Ran a bit more this afternoon, I tried various session lengths for the stop and go kinematic. Still recording every 2 seconds at the base and the rover, I tried sessions of 15, 17 and 20 seconds for the S&G points. Downloading it looks like yesterday---seems I'm losing either an epoch on the front end, the back end or both ends of each session. 20 results in either 20 (rarely), 18 or 16. 17 you would think would show as 16, but more often 14 or 12. Anyways, the results from just those few more seconds of data were much better than yesterday. Got to be the clock in the iPAQ. Regardless, the Locus and the iPAQ with Survey Control II being the obsolete animals they are, I'd be interested in the PM2 users with regard to the session lengths as displayed in the occupations tab of Solutions. If you ran a S&G with say 15 second occupations, does the data agree with the interval?



Dangerous Dave
Modified By Dave Huff on 10/21/2004 at 9:03 PM

Dave, humor me for a sec.....Take your base and rover files, convert to rinex, run them through the interpolation routine in Reliance.. Re-convert to Ashtech and put them back into AS. (I don't know if the d-file would work when paired with a rinex, kinda doubt it).

Reprocess and see if you picked up your missing epochs..If so, the time.delay..was...caused....by.....the......latency.......between........stop.........and..............go..........

No interpolation routine of Reliance to be had on my end, however, I did take a look at the rover files in FileTool, particularly the horizontal speed plot amongst other things. That looks pretty cool,that looks like the initialization period on the bar, here we go with the stop and go's, oops, stopped and lit a smoke there, oh yeah, here we go big time speed headed back for base. Size 13 shoes and I'm covering a lot of ground hehe. Maybe I should hang tight just a bit longer after moving before I log the next point? Ya think?

Dave

ps--Interesting thought on that one Bob, Thanks! Wonder why JD didn't have the same idear. (Oh, he's gonna shoot me now for sure!)
Modified By Dave Huff on 10/22/2004 at 1:02 AM

DD,

it's almpost gotta be the IPAQ or SCII software. I've run a bunch of the.....wait let me rephrase that....I've run several K surveys with Locii and the HP48 and NEVER had the problem you mention with occupation times vs actual epochs nated as being recorded.

While it does take a bit longer, I generally have used 20-30 second occupation times...as slow as it was to edit in the HP you were on the point that long anyhows !!

Never tried the Continuous K but have considered it, even had a little rig built to trailer behind a 4 wheeler or a vehicle...jst have got aroundtouit yet. Always wondered how them 1 epoch elevations would work out.

Jimbo

Dave, Reliance processor used to be available on the FTP server as a 16 diskette download....The interpolation routine in it is probably a version of NGS Interpo that runs as a dos program...I used to use that before Reliance was upgraded (ver 1.4 or later, currently at ver4.beta) Anyway, interpo would give you a test to see if the d-file settings would pick up those first and last epochs that you missed at 2 seconds.......b

Bob,

Thanks, I may look in to that.

Jimbo,

The continuous kinematic is a hoot. I put the Locus rover on a triple mag mount atop the Subaru Outback and drove the roads, had it mark a shot every 2 seconds. The "testing ground" is a park in a nearby suburb. It is actually the old Corp of Engineers Mississippi river basin model. Once a scale model of the MS river where they conducted all kinds of neat tests, it is now, the model basin that is, all grown up and abandoned. The property is generally wide open, it is one of the places we fly the R/C models.
I plan on "proofing" the S&G and continuous one weekend with the robot and the RTK. Should be fun! Here is some info on the site:
http://www.kilroywashere.org/004-Pages/JAN-Area/04-D-Jackson-POW.html

Dangerous Dave
Modified By Dave Huff on 10/22/2004 at 7:08 PM