I'm considering a PM2, asked for a demo and ended up with the following disappointment:

Of 20 known points surveyed in stop-go, 6 differed from the known values by between 20 and 60cm horizontally. Short vectors 300m max.

OK so I'm no GPS boffin but the salesperson is supposed to know which buttons to push. Plenty of sats (8+), low pdop. Sooo.... are these the sort of accuracies one has to live with, or to rephrase: is the 1cm claimed accuracy if fact no better than sub-metre ?

What I found particularly disturbing is that all the vectors seemed to have processed ok, small (sub cm) std errors etc. Or again to rphrase: there are a number of (insert expletive)-ups but nothing tells me so.

Please help a stoopid newbie......

I don't agree. Stoopid would have just said, "I measured this with GPS it has to be good."

The list of things that could be wrong is long but let's begin with the elementary.

Did you put the correct control point coordinates in for the base station. Some people actually assume that raw GPS is exact. It is not.

What makes these points "known" and what are they "known" by? Are we talking about the differenct between accuracy and precision?

Do you have the Ashtech Solutions manual. If you do not please get it from the ftp server at

ftp://ftp.thalesnavigation.com

Yes, I know it is a big download but the most skilled users have generally Read That Fine Manual.

Once you get it find the section called Post Processing Data Analysis and keep reading to the end of the chapter.

Let's just say for a moment that the salesperson made a mistake. Could you use what is in the training folder on the ftp server to determine what mistake was made and perhaps, even now, fix it.

Last, there are many ways to get a survey done wrong. That has been true since long before there was GPS. Getting the claimed accuracy is easy when proper procedures are used in appropriate places.

I'm not a Pro Mark 2 user, but I do have an arsenal of Locus units, Z 12s' an M 12 and a Z Surveyor, and none of them have let me down. I'd say the Pro Mark 2 is as good if not better than the specs published.
How did you establish your "known" points? With a total station? Other GPS gear? Could there be a rotational problem involved? Surely you aren't looking at the reported vector lengths between the points and then inversing your "known" coordinates and seeing a difference---because you will and seems everybody falls for this "discrepency" in the beginning.
I have had more than 1 instance where we set static GPS control for a project, only to have another "crew" traverse between the control and "mistie" by 8 (yes, eight) feet in a mile in a State Plane grid solution. I have then personally traversed between the same points with my total station, using a scale factor and have found the points I set to be in error of no more than 2cm.
Give us some more details of the "test track" and I bet a bunch of us will jump in on this!
Modified By Dave Huff on 4/30/2003 at 4:34 PM

Check if you are looking at ground or grid coordinates, rotational issues, rod heights, any number of things could be wrong.

I've been experimenting with the stop and go as well. I was getting crap results and was about ready to forget it. After about 4 or 5 email exchanges with Phil, I have figured out that it was operator error (I blame it on fat fingers and bad eyes...). I suggest reading the manual more than once, experimenting and when you get flustered, call or email Phil. I'm still not 100% sure what I did today, but I went from 0.4' differences to 0.04' differences.

PM2's are great machines. If you're not getting EXCELLENT results, then I'd wager it's a case of "equipment superior to the operator".

I've done stop & go topo work (locating edge of pavement) with 3" intervals and 9" observations, 98%± of the vectors solved ±0.1'

Also topo'd borrow pits in kinematic mode by putting mag-mount on the top of the pick-up (again, 3" intervals) and drivung all the breaklines while enjoying a cold coke & listening to Don williams music. Again 98%± of the vectors solved within a couple tenths.

Currently doing a corridor/boundary survey in wide-open skies of Wyo, 3" intervals, 60" observations on property corners. Worst vector residual I've had is 0.06'.

In case you're curious, I use 3" intervals to insure that I can collect 10-12 hours of data with the Locus that I use for a base station.

Jerry

Although people from Ashtech told me there was no tutorial. I found that in the support area of the Ashtech page, you can navigate to the "Reference Manuals/ Ashtech Solutions" and you will find a PDF file named "Old Tutorial". It helped me a lot.

I wonder when they will make a NEW TUTORIAL.