Bad RTK or Not?

I've heard that as long as your RTK is generating numbers, you're pretty much golden. I trust that you won't charge your client for actually checking a number generated via RTK, right?

I've done enough RTK over the last 15 years to know you can get bad numbers. Of course what is bad and how far is bad is a relative term (to some of us). But anyway, I considered this a problem zone and so I checked it. Also the UDOT's were likely only going to be measured with a total station.

Anyway, for what I'm doing a tenth or so is plenty good for me, the county and my clients. They don't get to pay for an extra five days work to get every new corner measured in under a hundredth. My expectation is that the future lot owners will use the monuments as set and found and not fuss to have them tuned up (moved or kicked about) to make them measurement perfect.

But hey, if you are ever in Utah surveying, you're welcome to go criticize the work, point out the deficiency of my RTK, and move these 62 new lot corners a few hundredths to measurement perfect. If you can get paid for the exercise then you truly are a survey god. Nobody else cares!

> I've done enough RTK over the last 15 years to know you can get bad numbers.

Okay, so is there something WRONG with your RTK system that you're worried about getting "bad numbers" (as opposed to "good" numbers, presumably)?

I mean, you did get NUMBERS, even if it took you a while. Isn't that the main thing?

I've had bad fixes before but these usually bust 5 feet or so. In pine trees you can get say a half foot difference with observations taken one right after another (while fixed). But a total station will whack out also in certain conditions. So there is nothing perfect. I'd say the real danger is technology in the hands of those that don't have enough experience and realize the limitations of the tools they are using.

> I've had bad fixes before but these usually bust 5 feet or so. In pine trees you can get say a half foot difference with observations taken one right after another (while fixed).

So, what you're saying is that you really don't have any confidence in your RTK-generated positions until you've checked them by some other means?

Note to Leon: starting a thread titled "Bad RTK or Not?" is like handing Kent a hammer and saying, "Hit me, please."

:good:

I don't have total confidence in any measurement system until I've done some redundant checking. There are certain conditions where I've experienced problems with measurements of all sorts where I don't trust the "numbers" until I've either done enough redundant measurements to gain some confidence or used two different systems one which I trust more than another for those conditions.

Under good conditions RTK will give excellent results. Over long distances (a mile or more) I would have much more faith in a RTK measurement than a multiple leg total station traverse.

Every RTK shot has up to a cm in slop. So if you are measuring 2 cm might not get such good results. But I've tried a lot of things over the years. I've compared 3 minute RTK observations to fast static. I've compared 5 second RTK to everything including long term static. I've compared RTK to total station measurements. My conclusion, yup, you got that cm floating around in there, other than that RTK works very good. Get the cm in opposite directions at the ends of an inverse between the shots and you got a couple of cm. Sometimes it's about spot on. For the work I do my clients can live with it and the savings in time and convenience.

One of the last things I'd do is go out and traverse say 3 miles thought the mountains with a good total station, miss close a few tenths (after least squares repair) from some RTK surveyor's direct measurement and declare that RTK is bad, the guy was an idiot and should be turned into the board for poor surveying.

Jim, that strikes me as completely unfair. Leon was just discussing both (a) his misgivings about the correctness of his RTK results and (b) how he went about "checking" them.

We cannot fault him for that, can we? I mean, my impression of Leon is that he's a somewhat earnest guy who uses RTK because his clients want the speedy anwers it provides and don't really care whether they are correct or not, but he himself is savvy enough to realize that he needs to check it any way he can.

I realise there are economics involved of course but I haven't used a GPS only unit since early 2007.

GNSS units are not that expensive anymore so if you are using RTK for anything, why would you not avail yourself of the extra satellites and productivity?

And the static snob returns to the board to lambast anyone who dares use a procedure that differs from his.

Glad to see you back.

> I realise there are economics involved of course but I haven't used a GPS only unit since early 2007.
>
> GNSS units are not that expensive anymore so if you are using RTK for anything, why would you not avail yourself of the extra satellites and productivity?

:good:

I doubt Kent would use a GNSS RTK if you gave it to him free! For the rest of us RTK seems to work pretty good.

My equipment has about reached (or past) it's better days. Considering that at this point if I bill an hour 90% goes straight to the bottom line, (everything paid for for some time) and I also don't work all the time at surveying, life is OK. Most large firms are doing good if 10% of gross ends up on the bottom line. But they have employees to manage and all the finest toys.

If I see the world picking up and a brighter future maybe I'll go ahead and invest a half years billing in new equipment and work an extra half year to pay for it. Maybe I'll do something else. I'd guess I'm about twice your age, so another decade is about all I could plan on being able to work in surveying assuming my health holds good.

LRDay,

I work considerably with the BLM and their chief geodesist locally. He's been a wealth of information. He has always said, if you can reinitialize on a point it is usually good. Have you tried that technique? I often will change the H.I. as a check on the reintialization.

Pablo

I don't have total confidence in any measurement system until I've done some redundant checking

:good: :good: :good:

On more important points (like section corners) I always reinitialize. For this little sub I did it for every set corner. Get a fix, make an observation, then dump the antenna, wait to get a new fix and shoot the marker again. If the two measurements aren't "about" the same (a few hundredths) then you may have a bad fix in which case you would do it a few more times and hopefully at different times of day. Even if fixed if you can't get good comparable results you are just in a condition where GPS is not that good. Maybe at a different time of day and SV arrangement it will work. You need all the SV signals you are using coming in without going through vegetation. If you have more than enough SV's sometimes you can figure out which SV's you may disable to improve the result's (or even enable fix). At my first visit to this point I couldn't get a fix at all. Some points just can't be measured with GPS so you have to use a different method for those. I been at this a long time.

I went back later in the day and after waiting for SV's in the western sky, got a fix, dumped it and got another fix which turns out were good. I just didn't like the condition of the site and wanted some further conformation (plus the UDOT markers were under the wall (didn't even try GPS on those).

This is what I consider a more critical boundary line because it's between federal (Forest Service) and my client (private). I not marking the section corners just using the Forest Service markers from 1969. Can't be to cautious on these now can we?

Another way to do this, is to assess the multipath in your environment, and move the gps to a LESS multipath environment, that is within 15' or so, and re-observe, then use compass and tape to CHECK between them. This gives you the redundancy, that lets you find a bad init.

When you do find a bad init, then, the games begin... pick a third location, and compass and tape between all 3 points, and now you find the bad one. OR TWO. I have found that certain times of day, when the sats are just right, well, they just yield more bad inits than others.

N

I appreciate this report of your results. It is a beautiful story. When things check that well it almost brings a tear to my eye.

The ugly little secret is that RTK is actually processing the data, like static does. Of course the big difference is you can look at the data in real time, if you wish to invest the time you can let the RTK "cook" and get a position.

I've never, never seen one of those be wrong, it's the quick ones that get you in trouble.

If you think about it RTK offers such an advantage over static.

You don't need to leave the site (or download to your laptop and process) to find out the real numbers, and you can check everything, use Static at the same time as the RTK and you have another check. Most people just don't use it that way or think that way, they think it's all separate processes that can't be intermingled.

One thing I don't understand, and maybe someone has done it, why not have the DC import local CORS data over wireless and then internally process a static base session or even a rover point right there in the field, seems so easy. The data is all available.