Another new GPS system thread- Geomax, Trimble, Igage,etc

[Norman_Oklahoma]

I think that you will see a significant performance upgrade over your R8-2’s no matter what new thing you buy.  And face it, those things are long in the tooth and due to crap out on you at the least convenient time.  

To reiterate the advise I always give – figure out what you want your field to office to finished product work flow to look like- shop for the data collection software that fits that best, and take the hardware that it runs.   

[lukenz]

With the usual caveat being is the dealer going to have good support read on. I-gage comes to mind especially if your wanting to use Carlson.

 

I’ve just started up solo over in New Zealand and bought a pair of zenith40s and zoom90 after two demo’s and six months of research (agonising) and confirming dealer can support. I’m running x-pad ultimate on Panasonic tablet. I’ve used Trimble access in past and most familiar with Leica gear.  Have used Javad ls’s too recently. Find the geomax hardware to be not to manufactured exactly Swiss standard (finishing is not as pretty) but better than other Chinese gear I demo’d and testing shows the same results as I’d expect from Leica ts15/fully optioned gs14 I was regularly using till now.

 

In terms of software functionality x-pad is a little more limited than Leica/Trimble but similar to surv-ce and I find way easier to navigate/use than Carlson. X-pad software still having features added all the time like Javad.

 

Here in NZ zoom90/zenith40 set is low 60k whereas Leica ts15/gs16 set more like mid 100k. I feel I get 90-95% performance of Leica for much less dollars.

 

Suggest investing in time to demo and working out how good the support will be should make decision for you?

[Andrew Clark]

We recently got the “Carlson” (it’s the Hemisphere re-brand for the US market) branded GPS units. They work great (even in some tree cover and near buildings too) and reasonably priced. I think a new GPS head was around 12k per head. Unsurprisingly enough, they pair quite well with the Carlson DC.

[richard-imrie]

@lukenz

Who is the dealer/support for Zoom90 in NZ?

[jhframe]

Posted by: @lukenz

In terms of software functionality x-pad is a little more limited than Leica/Trimble but similar to surv-ce and I find way easier to navigate/use than Carlson.

??

SurvCE *is* Carlson.

[lukenz]

@jim-frame

Sorry, made the assumption that was obvious. Strictly should have stuck to survCE throughout post.

[lukenz]

@richard-imrie

Buildmax, sister company to global survey, the one they try to keep away from surveyor s so you’ll stick buying Leica. Same owners, use same service centres just different sales and support channels.

[hpalmer]

@lukenz

I think some business model’s are outdated.  I hate it that my manufacturer’s sales rep has to compete with a same/similar product made by their parent company.  I hate it that much of our Leica equipment will not talk to non-Leica data collectors. We paid lots (and contributed technology) for Leica’s product development with little reciprocation other than the initial sale.  I hate it that those on-board programs we paid for cannot be used on other Leica instruments.

I like that much of the Geomax line is built by and has same quality control as many of the Leica instruments.  I like it that our Geomax equipment will talk with many of the non-Leica data collectors.  I like it that our Geomax came fully loaded.

[Luke7574]

@lukenz I was wondering if you stiil feel the same way about x-pad being “limited” I have been using for the last few months and I find it has many advantages over trimble access in that it has a cad program that is very robust and allows for OSNAP functions for creating points from imported dwg/dxf files as well it lets you snap on points of your base map and translate /rotate to collected points.In my opinion Trimble is very limited in map functions which in my opinion is essential to a comprehensive Survey program.I am going to go out on a limb here and say that besides field-2-finish functions XPAD has them all beat.

[lukenz]

@luke7574

Still of the same opinion that for a boundary/geodetic surveyor X-pad Ultimate software is more limited than access/captivate (if Leica wasn’t close to twice the price I’d have got that as I like all the technical functionality they make available to the user on the controller). Also manual doesn’t explain all of behind the scenes calculations as well as Leica/Trimble do. That said none of the missing Cogo or other functionality in x-pad is a deal breaker and due to the price differential still an excellent option (wouldn’t have shelled out personally for it otherwise). Hardware performance just as good as the Leica gear.

 

I’m not doing much topo/stakeout on simple grid/site coordinates nor does my topo workflow require creating DWG in field (I use feature codes in field and desktop software applies map file) but I can see it has additional functions compared with other software in that area. As you mention x-pad has a heap of CAD functionality that is solid just not much use to me. You can also use background maps from web map services too which would be very useful for some users.

 

Where I think it is superior is the user interface, so easy to pick up how to use it and find what you are looking for and feels like a smartphone which everyone is familiar with; be easy to train people to use it. I like having it on a tablet too as 7inch not heavier than standard controller but much better for viewing plans on it. I just bring all the boundary plans on a memory stick and plug it in if I need to add to my pre-calculations.

[Luke7574]

@lukenz yeah,Im not doing much work with site coordinates either.Im working with state plane coordinates if that makes any difference to you and how x-pad is functioning as a “geodetic surveying” data collector. Im just curious what exactly do you feel like is missing in x-pad ultimate to where is doesn’t stack up to trimble access or captivate as a “survey/geodetic” collector,what cogo functions are you missing?

[lukenz]

@luke7574

Here in NZ we use transverse Mercator projections; a national one which I think is similar to your state ones and local circuit which sounds similar to LDPs I hear about on here. Our cadastral system is vector based plans with grid bearings and ellipsoidal distances. We use grid coordinates on local TM circuit. Ground distances are only for construction work here.

 

With xpad ultimate there are no fully functioning inverse or traverse functions which are the ones I want most doing boundary work.  It’s not totally non existent just not well executed.  Xpad makes a point of saying they built it from a new angle but for these functions I wish they’d stuck to tradition.

 

In the COGO/CAD menus the distance function is as close to inverse and does return grid bearings but for distance Trimble/Leica/Javad all provide drop down options to select grid/ground/ellipsoidal for distance output whereas xpad will return grid and maybe ground or ellipsoidal depending on options you use in the co-ordinate system menu; ie if I set base up on new point, survey a control point with rover and do a single point gnss site calibration (scale factor at 1, just n,E,z translation) the distance function stops displaying ellipsoidal distances which is a PIA. Also in COGO the default is to measure to current position which means software won’t let you start distance function without being connected to gear so if you want to turn off gear while you do calcs you can’t use that function.  Need this function to create vectors between old marks GNSS’d to check against old plans for bearing swings and to confirm marks still reliable.

 

Also in COGO menu there is no traverse function which the other three above provide and even sdr33 had! In the CAD menu there is the layout function but again you can not user select the distance type to input.  Also a PIA as all our survey plans have ellipsoidal distances which I want to input and have software calc the grid coordinates not have to work out the required scale factor, and apply to distance with calculator and then input into software. Need this function to extend office pre-calculations in field to look for extra marks etc.

 

Also the GNSS site calibration doesn’t work as well as Javads m-local or Trimble’s site calibration (mostly because they don’t publish the maths/process behind it so hard to know what it is doing, don’t like using functions without knowing exactly what they are doing), Leica two-step transformation is more painful however.

 

The offset from 2 points for GNSS, reference line and offset left/right/back from prism functions are awesome and well coded so I know they can do it.  Still of opinion it’s good value and easy to use just not as comprehensive as Trimble/Leica offerings but their packages are close to twice the price so for shops not wedded to a brand it’s great value and very simple to learn/operate. Plus hardware performance just as good as Leica.

[plumb-bill]

Posted by: @norman-oklahoma

I think that you will see a significant performance upgrade over your R8-2’s no matter what new thing you buy. 

Yep

[Tom Bushelman]

I am a Javad guy, drank all the kool aid and love it, formerly a Trimble man.  A friend of mine had I-Gage and he borrowed my Javad.  As I was showing him how it worked, he seemed to like the functionality of it much better than the I-gage, so much so that he purchased his own base/rover set.  A selling point for the Javad equipment is that it requires no data collector, freeing up your options for robotic equipment future purchases(get the S-5).  Shawn Billings is from Texas I believe and he was/is instrumental in the development of the Javad software and he can’t be that far.  Texas is a little bitty place.  It is worth a test drive.  In my very generalized opinion, Trimble, Leica, and Javad are the top performers not necessarily in that order and you can save $20,000 or so with one of them.

[lukenz]

@tom-bushelman

Having used a pair of Javad LS’s in NZ after watching video’s of fixes under tree canopy we were disappointed with hardware/measurement engine performance (software ok once you get your head around it) compared to Leica gs14 had been using (and the geomax Zenith40 I’m using now).

 

The Javad would get more fixes under trees using default “boundary” setting but not as precise as Leica in open using rtk with <1km from base and similar observation periods (30 sec – 1 min). For boundary/control work we prefer the greater precision/repeatiblity at the <8-15mm level horizontal of Leica whereas Javad was more like <10-20mm. Our limit between control points for boundary work is 20mm + 10mm per 100m and Javad was often just on it.

 

Not sure I agree with data collector comment either, I prefer to be able to start job using GNSS and then connect to robot to get the last few points and have all data in one job.

 

My 2c

[matt8200]

Posted by: @lukenz

The Javad would get more fixes under trees using default “boundary” setting but not as precise as Leica in open using rtk with <1km from base and similar observation periods (30 sec – 1 min). For boundary/control work we prefer the greater precision/repeatiblity at the <8-15mm level horizontal of Leica whereas Javad was more like <10-20mm. Our limit between control points for boundary work is 20mm + 10mm per 100m and Javad was often just on it.

I did a precision test in April with a Javad T1-M base and Javad Triumph-LS setup in the open.  The “Precise Topo” profile was used and “Auto Repeat” was used to collect 789 points.  The average duration of the measurements for points was 11.42 seconds.  Both the horizontal and vertical precisions were very good:

100% of the points were within 0.023 ft (7 mm) horizontally and 0.062 ft (19mm) vertically of the averaged position of all the points. 

 

I then also post-processed the base rover vectors with DPOS that is built into the software.  The results were rather spectacular with 100% of the points were within 0.013 ft (4 mm) horizontally and 0.026 ft (8mm) vertically of the averaged position of all the points. 

Note that DPOS uses absolute calibrations in processing and it is critical that both receivers be aligned to north to achieve these precisions. 

[lukenz]

@matt8200

So that’s just under three hours observing at 12 seconds a shot? Long enough to give the necessary satellite geometry change for sure. My numbers are from repeat shots (double ties 0.5-4hr between shots) with same base setup per job but over lots of different jobs/days (i.e. not necessarily ideal satellite geo or iono activity every time).  Don’t have a great handle on height values as were doing boundary work mostly.

 

Also assume you had both receivers on tripod/bipod for observation duration? My numbers are repeatiblity revisits with base in same location on tripod but pole plumbed each time which would add 2-4mm to numbers above. 

 

Even with the differences noted above and being in a different part of the world we seemed to not be able to repeat your results. Both the boss and I read all the guides and the online forum but no local dealer support. We also used the boundary measurement profile which we understood to be likely to deliver a more repeatible result than precise topo.

 

Your comment about orientation for receivers, does that apply for rtk too? And would it have made a difference we had an LS as a base?

[shawn-billings]

In 2015, I ran quite a few precision tests which I think ultimately led Javad GNSS to revise the accuracy claims in our literature. I tested at ranges from 0-15km (0-9 miles). This was my conclusion from those tests and comments on a baseline distance of 25km (15 miles):

If my error estimates for the new engines are correct, He(RMS) = 5mm + 0.7ppm x BL and Ve(RMS) = 10+0.7ppm x BL, and the baseline length is 24895m, then the Horizontal Error would be 0.022m (1 standard deviation) and the Vertical would be 0.027m (1 standard deviation). The Horizontal Error at 2 standard deviations is 0.035m and Vertical Error at 2 standard deviations is 0.053m.

[these estimates] are based on 240 epochs…

From memory the difference between a 240 second observation and a 30 second observation was about 15% improvement in accuracy. 

 

Precision will open up a bit if you:

• use tilt correction – more measurements from more sensors, each with its own precision
• poorly calibrated hardware (poles, level bubbles, points, etc.)
• phase center offsets (mis-alignment of base and rover)

Back in 2015, we advertised a horizontal precision (RMS) of 1cm+1ppm and vertical (RMS) of 2cm+2ppm (if I recall correctly), just like everyone else did. Today, we advertise a precision of:

Horizontal (RMS): 4mm + 1ppm

Vertical (RMS): 7mm + 1.5ppm

Which seems to agree pretty well with my testing from 2015. 

Longer observations did appear to improve accuracy somewhat, but not as much as I would have expected. I studied observations from 10 seconds to 20 minutes in length and at the time felt like their was an inflection point at about 240 seconds (four minutes). Four minutes seemed to mark a point of diminishing returns, but as stated above, the difference between a 30 second observation and a 4 minute observation was about 15%.

 

 

 

[lukenz]

@shawn-billings

Glad to see you point out RMS vs 95% CI. I always tell staff to double the RMS values to get closer to true repeatiblity numbers (for instrument error, plumbing errors on top also). We were using a braced pole that was regularly checked for plumb and bubble adjustment. Also try to take obs with base and rover facing approx north.

 

I know all manufacturers publish RMS values but for a professional surveyor they are nonsense, at minimum we should be considering 95% CI values if not 99% CI when quoting repeatiblity of our work.

 

The 95% CI precision figures you have above (0.022m HZ) fit with what we were seeing in the field.

[shawn-billings]

@lukenz

Note that the horizontal accuracy of 0.022m is at a distance of 25km in the above quote. I agree with you regarding 95% vs 68%. Our product literature is written to “industry standard” to make comparisons for future customers simpler. But in our software, we often show accuracy estimates at the 95% confidence interval, because, as you point out, it is much more realistic usually.