RTK Occupation TImes

Norman,
I've invested considerable time and effort into this question. Our procedures aren't perfect but they work.
We occupy all control and boundary ties 30 by 30 seconds, rotating the pole 180 between. If we have prior dimensions to check (and they work) we call it good. Most times we follow up with a second 30 by 30 under a different constellation.
We cut off RTK and switch to rapid static if quality goes down. In my experience longer RTK observations in a bad environment or with high PDOP only gives me more bad data to stare at. The continuous apparent improvement in position is due to the abuse of the 'square root of the number of observations' programmed into most RTK units. Contrary to popular belief an RTK epoch cannot be treated as an independent observation (flame away)...
As the units 'improve' I am finding it more important to hold to these procedures. We kicked and screamed and whined to get things that would lock under canopy. Well, we got them. I see more bad inits in a month than I used to in a year.
My suggestion is to test the units you have under varying conditions. See if you can get it to display a quality indicator that matches what your testing shows to be true. Above all I would set the PDOP and QC limits as a stop RTK, not extend occupation.
My .02, Tom

> We occupy all control and boundary ties 30 by 30 seconds, rotating the pole 180 between.

I'm curious about the pole rotation. This is sometimes done to cancel out a maladjusted bubble when measuring to a prism, but given the magnitude of nominal RTK errors, the bubble would have to be way out for this to make a difference. Is there another reason for spinning the pole?

> I'm curious about the pole rotation. ... Is there another reason for spinning the pole?
I do that, too. As proof that the pole was plumb. If I was in the field myself I probably wouldn't bother with it, but when working in a team environment you have to build these sort of things into your procedures, IMHO. And as long as the field crew knows that I'm checking, the pole is always plumb. That may not always be so if they thought I wasn't checking.

> And as long as the field crew knows that I'm checking, the pole is always plumb. That may not always be so if they thought I wasn't checking.

I'm not going to knock anyone's procedures -- I have plenty of quirky ones of my own -- but if you don't trust the crew to keep the pole plumb, what makes you trust them to rotate it for the second shot?

For important points I take five shots in succession with the same point number, and the unit averages them. I never rotate the pole. Rather, I always face North, reasoning that if the bubble is indeed off it will be off the same amount on all the important shots and they will thus be relatively correct.

For less important points, like parking stripes, I arrive, hesitate a second to allow the unit to catch up, and take one shot.

When I later use reflectorless to shoot between my control points, they never exceed 0.025 feet difference, and they average 0.01 feet difference. All, obviously, in somewhat open areas, not near trees or buildings.

Are we are talking ties to boundary control? The length of time spent at a station is relative to the length of vector being measured. The longer the vector, the longer the observation and all the more reason for a repeat observation under different satellite geometry. I always collect raw data while doing RTK. I set it up to collect data at 5 second intervals. Vectors of a few hundred feet; observations of under a minute are fine. Barring topo shots, I like at least a minute in case I have to post-process the raw data for whatever reason. Stretch that out to a mile or so and I will do 10-15 minute obs and post-process. Beyond that, I'm in full static mode collecting at 15 second intervals with 30 minute sessions. Whether I'm doing RTK, Fast-Static or Static observations, I perform repeat observations. If it's worth recovering and measuring once, its worth measuring again for confidence.

Another note: If I'm not post-processing raw data, I will always run the RTK vectors through Star*Net and look at the residuals.

Back in the day when we would do 1 hour sessions and be at a station for 3 or 4 hours (3-4 sessions), I would rotate the fixed height tripod 180° between sessions. The bubble gets banged around and goes out of whack. A good thing to do if nothing else but to see if the bubble is still good. If I was using a tripod/tribrach, I would break the setup and get a different HI for the second and third sessions.

> Above all I would set the PDOP and QC limits as a stop RTK, not extend occupation.
:good:

There is a guide in the UK from the TSA and Ordnance Survey which I have attached.

It might be of interest.

https://www.dropbox.com/s/f7r1e22gl167nol/Network_RTK_Full_Report.pdf?dl=0

Regards,

Tim

I'm not 100% sure but could you just adjust your mask angle to get a better pdop longer

> For important points I take five shots in succession with the same point number, and the unit averages them. I never rotate the pole. Rather, I always face North, reasoning that if the bubble is indeed off it will be off the same amount on all the important shots and they will thus be relatively correct.
>
> For less important points, like parking stripes, I arrive, hesitate a second to allow the unit to catch up, and take one shot.
>
> When I later use reflectorless to shoot between my control points, they never exceed 0.025 feet difference, and they average 0.01 feet difference. All, obviously, in somewhat open areas, not near trees or buildings.

I'm with Bruce on this one. We have been using RTN here for nearly two years. here are some of the techniques we have grown into:

for the more important measurements, we take four measurements, meaned, on a station. we often will take four more measurements under a different constellation

we monitor GDOP and increase occupy time to suit when necessary

we monitor coordinate quality (ies) and keep in mind that most CQ's are displayed one sigma. we are concerned with two sigma results, so i multiply the CQ by two or two and a half to get a more reasonable estimate (as two sigma)

we always record the kinematic track for post processing from CORS. it is a small price to pay for redundant measurements. i never have had a comfort level with only recording vectors. keep in mind also that RTN still does not operate with the benefit of precise ephemeris. if you don't record it, you cannot get it back for a second chance.

on the most important measurements, we will bring one or two other static receivers, in addition to our RTN receiver, to get local vectors as well.

Poor PDOP Can Be UP To 15 Minutes Duration

And poor PDOP does not just include low satellite counts. I have seen up to 8 satellites in sky positions that give a poor geometric solution. Most PDOP calculators do not factor in sky positions.

Reoccupation at a different time is the preferred solution.

Paul in PA

State of Connecticut guidelines

GNSS guidelines

I found this very helpful.

Which RTK?

There are a couple of variables we need to eliminate here. First, why type of RTK? Traditional, where you are also setting up a base and broadcasting your correction? If so, how was this point established? How are you achieving ground coordinates?

Are you utilizing a network that pushes out single-base corrections via an IP/port? If so, how long of baseline do you typically run?

OR, is this a network solution (VRS since I'm a Trimble guy) that is modeling and eliminating a lot of the PPM error? Again, how long of a baseline do you typically use?

The answer varies quite a bit between the methods.

~Raybies

Which RTK?

> There are a couple of variables we need to eliminate here. First, why type of RTK?
Oregon has a VRS system. So does Washington. The virtual baselines are typically a few kilometers in length. We also sometimes operate at the fringes of, or outside of, cell phone service areas and are obliged to switch to out own base. In those cases the baselines would typically be short, perhaps a mile, maybe two.

The rotation was something I added in after a few experiments. I am not certain of everything it compensates for. I suspect it is mostly the bubble and user bias. As a results oriented person I simply kept the routine after seeing significant position improvement,
We continue to migrate further from RTK as time passes. We post process everything anyway. I get a lot more sleep with dual base rapid static...

Welcome back to the Pacific NW Mark. I'd recommend talking to Mark R. down in Marion County. A wealth of knowledge, he's been invaluable to our efforts in understanding the benefits and pitfalls, the dos and don'ts of RTK. And, his work is all over on the heavily treed west side.

Sound advice!

I would suggest that you also speak with Mark Armstrong (NGS advisor). I applaud your efforts.

Welcome back!

> ...if you don't trust the crew to keep the pole plumb, what makes you trust them to rotate it for the second shot?
True that. I'm thinking more of pole bubble adjustment, and user bias, rather than the mechanics of shot taking.

> There is a guide in the UK from the TSA and Ordnance Survey which I have attached.
Thanks, Tim. That is useful.