Will, I finally got my LS to lie to me....

[Nate The Surveyor]

Here's the story.
I have a Triumph 2 on base.
I have a 35 watt radio, running this day at 10 watts and 1 second period.
I have a Javad LS on rover.

Settings:
I recently set my Confidence Guard to 0.25' horiz, and 0.40' Vert.

Confidence Level 10
show on screen 6 groups
Flashlight blink on.
Verify with V6 reset, (on)
Consistency 10
Min RTK Engines 2
And, here are the screenshots, in the same order that I find them:
1.)

2.)

3.)

4.)
Here is a very interesting shot.

5.)

6.)
Well, this forum won't allow any more pics. per posting. I'll add some in a little bit.

 

[Nate The Surveyor]

Here is the last pic

OK, This is called a "string-line" shot. It just keeps GROWING the string across the screen. It never did stop. It was set to 200 seconds and 120 epochs, minimum.
this is the FIRST and only time, I have EVER gotten a shot, that was wrong, by over 0.16', that I know of.
I think it COULD be possible to have something in the software that detects "String-lining" And, causes a jump-fail. However, I also tend to think/assume, that it is RELATED to the opening up of the confidence guard.
This shot was wrong by 7.0 feet. It had open air to the east, and a pine thicket to the west. It was under the pine thicket, a little.
This is ALSO the largest I can recall seeing, the Left had square being.. (0.40') It is on a prism pole, with a bipod. So, it has little to no wiggle.
Height of the LS was at 5.1'.
I later shot this same point 3x more. The last 3 shots were within 0.05'.
I have been trying to figure out how to "Break" the Javad system, for, well ever since I got it. (I want to know what NOT to do).
So, I have this to say: Watch for String-line shots. Especially the ones with excessive size of the box, on the left.
I have to go back to this job, and set corners. If you would like a pic of the site, I can do that.

Respectfully,

Nate

 

[Nate The Surveyor]

One more relevant thing.
In the after noon, around here, (Arkansas) GPS data, between 4:00 and 6:00 PM has been poorer than average. This was around 6:00 ish. (Maybe 6:28PM from the NE corner of pic 4).

Anyway, boundary shots, in woods, often get 3 shots from me. (Maybe I'm paranoid... but it's my comfort zone..)
My previous GPS was a Topcon Legacy E. It lied alot more than this...

Nate

 

[Matt Johnson]

I recently set my Confidence Guard to 0.25' horiz, and 0.40' Vert.

This is the problem here. The Confidence Gaurd settings are too high.

 

[Nate The Surveyor]

too high.

Pic number 5 above, says "Guard=0.164 ft"
And, where you set the confidence Guard, I had set it to 0.25' Horiz and 0.40' vert.
What is the "Guard" above, in pic 5?
And, what would you recommend, for the confidence guard?
Thanks,

N

 

[Shawn Billings]

That's the 3D engine guard. Different animal. It determines what tolerance is required for engines to be considered in agreement.

 

[Nate The Surveyor]

Thanks Shawn.
I have been running
0.20 h
0.25 v
Until eclipse day.
Before that, I was running
.17h
.20v

Matt,

Are you saying I should be running
0.30 or 0.40 h
And 0.50 or 0.60 v?

Did I hear you right?

I've never run this loose before.

If memory serves correctly, when I first got mine, it was:
0.165 h
0.250 v

N

 

[Nate The Surveyor]

Thanks Shawn.
I have been running
0.20 h
0.25 v
Until eclipse day.
Before that, I was running
.17h
.20v

Matt,

Are you saying I should be running
0.30 or 0.40 h
And 0.50 or 0.60 v?

Did I hear you right?

I've never run this loose before.

If memory serves correctly, when I first got mine, it was:
0.165 h
0.250 v

N

 

[Nate The Surveyor]

Oops

 

[Adam]

The default confidence gaurds are what I recommend for all critical points (boundary and control). I have a set up for very rough locations that I use 1/.5 as the confidence gaurd but I wouldn't ever use it for anything but rough data.

 

[Matt Johnson]

I am saying that you should leave these at the default values of 4 and 7 cm (0.13 and 0.23 ft).

 

[Nate The Surveyor]

Okay Matt thank you. At default value it would have done a jump fail. Got it.

 

[Shawn Billings]

I've been using 0.2' H and 0.3' V with good success. I was noticing split groups with default settings. Opening the confidence guard a little has all but eliminated this.

Stringing is definitely something to watch for. It's a sign of multipath influence. Not always a sign of a bad fix, but definitely a sign of multipath.

 

[Darren Clemons]

I've been using 0.2' H and 0.3' V with good success. I was noticing split groups with default settings. Opening the confidence guard a little has all but eliminated this.

Stringing is definitely something to watch for. It's a sign of multipath influence. Not always a sign of a bad fix, but definitely a sign of multipath.

I did exactly the same thing Shawn.....was getting several groups filling up two or three at a time in coverage....upping it a tad really helped that.
Such an incredible machine to be able to "dial" things in for your own personal taste.

 

[Nate The Surveyor]

Well, that's the same thing I did.
Started out at default, 4 cm and 7 cm.
then, went to 0.17' and 0.25'

then, went to 0.20' and ....

Finally went to 0.25' and 0.40', And, guess what? TOO MUCH of a good thing...

Now, IF it were allowed to keep cooking, (longer time) on the shot above, I do believe it would have "Jump Failed". But, 200 seconds was NOT ENOUGH time, for it to discover it was "in the bad".


I am saying this, because, there will be others, coming after us, and we should leave the "Footsteps plain".

Thanks everyone, for what you have said. It's an evolving tool. And, we are learning with it.

Nate

 

[Matthew D. Sibole]

I just got this picture from a client.

I thought it fit in well with this discussion.

In this instance I would have accepted this shot long ago. Typically after getting a few epochs over 240 seconds apart from each other. I would have then repeated the observation 2 more times.

What was done here is not necessarily bad. What this does show is that over a significant time frame many epochs were collected that all fall within 0.316' of each other and are being averaged out. Good accuracy, poor precision.

Repeated observations will give good accuracy and good precision.

 

[Shawn Billings]

Nate,
I see that the screen capture of the bad point shows exactly 200 seconds. Validate would have added at least 2 or 3 seconds more. Looks like the validate bug was present in this observation.

What happens currently is the validate reset will happen, but if the new fix doesn't agree with the average, the epochs will not be added but the accept reject prompt will be given suggesting that the test passed.

In this instance, the validate phase occurred but the new fix didn't agree with your 200 second running average. The validate epochs weren't added to the point, but it didn't reattempt the fix or eventually fail on jumps as it should have.

 

[Shawn Billings]

I could kick myself for not spotting this when you posted it. The clues are all right there in the description of your settings and the screen capture.