RTK is useful, but it will lie to you.

RTK is useful when you just don't push the buttons.
RTK is useful when you pay attention to your DOPs, satellite configuration, and signal to noise ratios.
RTK is useful when you learn how and why it works.
RTK is useful when you take the time to READ THE DAMN MANUALS for the controller, the receiver, adn the radio.
RTK is useful when you're taught and use proper procedure.
RTK is useful when you make some elementary checks.

Tell me and I'll forget.
Show me and I'll remember.
Involve me and I'll understand.

I know a local fellow that builds furniture.
He doesn't own any tools except a hatchet and a hammer.
The furniture looks like crap.

RTK is useful as long as you understand that on occasions a point will be wrong for inexplicable reasons, even though everything suggests the reading should be perfect. It's a fact of life.

If it's essential that the point is right, return to it later, after switching off and re-initialising. In any case always set a couple of check points at the start and pick them up again at the end. Then you can drive home worrying that the whole of the survey is suspect!

Nate,

Our trust level is the same and our procedures are almost identical. The only difference is if we can not come back to a point, we reinitialize and shoot the point again. I have seen some reinitialize and stakeout the point. I like to mean the values when they are close.

I agree with everything that Unmanned stated. However, experience has shown that if you do everything he suggested, you will occasionally get a wrong answer. I agree that most of the time the discrepancy can be explained by the violation of one of the elements of good proceedure. The problem is the surveyor and not the equipment.

:totalstation: :stakeout:

This guy is someone I would like to have a whiskey with. He said in one short paragraph what has been discussed in the longest most redundant threads I can remember.

Does it break when you sit on it, like some of the pretty stuff:-S

Another good check when you can't return is to set two swing tie points in a good configuration, say 6-10' from the point needed. I have found that you can get the same bad fix twice in the same spot, but moving just that little bit shows the error (and the same bad fix is a lot less likely in a different spot). Also, if all else fails I can compute the point from the swing ties (and manually check the distance between swing ties, then compare with inverse in the DC). Some points only need to be within a tenth.

I thought about building a folding 3 hole jig. One hole would be over/centered on the subject point the other two holes would be at a set distance and maybe an angle or exact straight line . Shoot all three and and inverse . Then stake all three .

Seems like something I could get printed .

Heck of an idea

I'm picturing a nice pair of legs that when folded out form a precise angle. Have a hole to center over the true shot and dimples for the other two known positions. Practice, practice, practice to confirm reliability. Only use it when you really need to. The main design thought would be how long to make the legs so that it would work in the vast majority of your situations. That is, how far out must you move to get reliable answers at the two dimples?

Heck of an idea

Seems like one foot legs would work. Of course you can do the same with 2 nails and the main point .

Heck of an idea

I was thinking more like eight-foot legs to move the shot away from the source of problem enough to prevent having the same problem for all three shots. Something that folds down like a pocket ruler yet locked in a uniform geometry every time.

Never trust a fix on important point

I've had good fixes in heavy canopy and bad fixes in open sky. I would never bet my life on a single RTK fix.

Heck of an idea

You flatlanders! Geesh

I don't believe I have ever shot a point where you could lay out a contraption like that of any size.

Heck of an idea

:good:

I've usually already got nails and a tape, anyway.

Never trust a fix on important point

:good:

Same is true for a total station. In the very first LSA class I ever took I remember hearing the statement "Anything you know from one observation, you don't really know at all".

Can anyone explain why the initialization would be different because you moved the receiver slightly?

-the short answer is because the objects causing multipath aren't miles away, and a change of this magnitude at the ground can cause a significant change in geometry of the multipath.

It has been my experience that 6-10' is usually enough (under canopy) to negate the effects of multipath that caused the bad integer fix at the previous location. I have nothing to offer on paper, and I'm not saying it always works. It has pretty much always worked for me, though.

With newer receivers, some times you don't even need to move - just wait three minutes. The geometry of the satellites will have changed enough to reveal/negate the multipath. Sometimes it will confirm the multipath with yet another bad fix, but it almost never equals the original (sometimes it does, though).

The crux of this whole "bad fix" discussion is to pay attention and know what you're equipment is capable of both good and bad - and that one observation never tells you anything.

I am in the camp that you pretty much have to export the vectors and include them in an LSA if you expect any meaningful data.

Bad data can be as useful as good data, if you allow it to teach you how to derive good data.

Never trust a fix on important point

:good: