VRS and Javad GPS L1 only OEM RTK board

the http://www.javad.com/downloads/javadgnss/sheets/TRH-G2_Datasheet.pdf&apos ;">datasheet

Yuriy Lutsyshyn, post: 454906, member: 2507 wrote: Hi,
Because RTK networks can almost always create VRS literally a few meter away from a rover I was wondering why not to use L1 RTK units with good multipath reduction algorithms?

I came across https://www.javad.com/dynamic/Shop/Product/2856&apos ;">Javad GPS L1 only RTK OEM board and was wondering how good are they ?
I am interested in this combination of options: RTK Rate 1 Hz + Advanced Multipath Reduction, which makes total $1,276.88

How good the board performs with Advanced Multipath Reduction option compared to performance without this option?
How long does it take to initialize ? and how reliable the initialization is ?

Thanks a lot.

Hi Yuriy, [USER=10007]@Michael Glutting[/USER] may be able to help you out. Here is the number to the San Jose headquarters. +1(408)770-1770

Hello https://surveyorconnect.com/community/members/yuriy-lutsyshyn.2507/&apos ;">Yuriy,

This could be an interesting experiment. Our European team will be in contact with you Monday. I look forward to reading if your hypothesis is supported or not by the data.

Do you have a "high precision geodetic" GNSS antenna in mind yet? I suggest considering the GrAnt-G3.

JAVAD GNSS users (and I imagine other manufacturer's users) can test this hypothesis as well by disabling L2 & L5 tracking.

Regards,
Michael Glutting
JAVAD GNSS, Inc. | Sales
M 408.375.9135 | F 408.663.6626

Thanks a lot,

It would be really interesting if someone who actually owns the board could share his experience with the board's performance on short baselines ( just dreaming 😉
http://www.tallysman.com/index.php/gnss/products/&apos ;">Tallyssman is making good OEM antennas, some of them are NGS calibrated, and they are inexpensive. There are other cheap and still good alternatives.

Yuriy,

I just set my LS to be L1 only. I was connected to the Ohio RTN which is a Trimble VRS system with the fictional nearby base. In the open, initialization time was between 10 and 30 seconds. Accuracy was right around 1cm horizontal and vertical. I only ran the test for about half an hour.

Under the tree canopy using "L1 only" was not fun. I did successfully verify and repeat a location, but I also had many bad fixes.

My opinion is that in the open, you would have a wonderfully productive system.

Interesting thread. How does the built-in LS antenna compare to an external? Wondering if that would make any difference for an L1 only receiver as i understand the LS and its algorithms are geared to all GPS and glonass.

Cool topic.

I have accidentally run the LS in canopy with L1 only corrections in base rover (not VRS ) a couple of times. It was slower but I was surprised that it worked at all. In fact performance was so good that it took me a while to discover. Of course base-rover is going to be better than VRS when working under canopy, so I am sure John's experience is accurate.

You sure read a lot into an abbreviated comment regarding canopy and comparative differences between RTN and local base corrections. Sensitive much?

The OP didn't ask about ocean tides, he asked about multipath. I've not heard anyone suggest performance with an RTN is better than a local base when working under canopy. Perhaps your experience, testing, or anecdotes from users is different.

I did not, nor would I, suggest that RTN's "totally suck". There are significant logistical, and depending on the subscription costs, economical advantages of RTN over local base. Straw-man much? I do not believe that Time-To-Fix, Accuracy, and performance under canopy are strengths of RTN solutions, but I'm fully prepared to be corrected on these considerations.

Gavin, of late, you've been resurrecting debate tactics that we both once derided. I had suspicions that you were better than that, but perhaps you'll prove me wrong yet.

I sold L1 only systems for quite some time... up until a few years ago. The ProMark 3 (L1 GPS Only) and then the ProMark 100 & 120 (L1 GPS+GLO) were examples.

Five to ten years ago these receivers had a reasonable niche: they were 1/6 the cost of a R8 and the RTK performance in a VRS network was not horrible.

The time to fix was quite long, usually 90 seconds, but sometimes as long as 10 minutes and other times in-explicitly they never fixed. Once they fixed, they typically were just as accurate as L1/L2.

Even though a VRS network solution appear to be within a meter of your location, the solution is still a mathematical function of possibly distant CORS sites. I think that we empirically found that the L1 network solution was approximately as good as the shortest 'real baseline' (the distance to the nearest CORS site) divided by 2. So if you were between distant CORS sites with the shortest true baseline being 20 KM, it would perform about as well as having your own L1 base about 10 KM away. (Which is to say, not very well.) If there was a CORS site 6 KM distant, the performance was about the same as a 3 KM true baseline which would not be bad.

From a price perspective, a PM3 Network Rover was about $5,000 street price with software (RTK option + FAST Survey). The ProMark 100's which could optionally track GLO L1 (and then later the PM120 which could do L1L2G1G2) were about $6,500 in the L1G1 configuration. Dual frequency units at the time were MUCH more expensive.

Today, I think that you will find a much smaller price differential to excellent full constellation network rovers (for example I sell OEM6 based network rovers for around $6k or less). I know that you are looking at OEM Boards and I think that you will find that L1L2G1G2 solutions are only marginally more expensive than the L1G1 only boards. I personally believe that the much better TTF for dual frequency boards will be worth the small increase in price.

I also believe that that the time has come for all new receivers to track and L2C, L5, GLO L3, and Galileo. Again, the price differential between full constellation receivers and marginal tracking is not worth the savings, except perhaps for machine control applications where lock is never lost. Since I personally specify these engines, I have a very good idea what the cost of every component including the engine brings. I really think Dual Frequency performance is now absolutely worth the very small price differential.

BTW, you should look at the Emlid receivers. They are built on this L1G1 only concept. They are VERY inexpensive. They are sold under an interesting sales model. https://www.swiftnav.com/&apos ;">Swiftnav and http://www.sapcorda.com/&apos ;">Sapcorda (both L1L2 I think) have a potential to drop L1 only devices totally out of the picture... In addition, there are a bunch of Chinese engine manufacturers who have decent engines at very reasonable price points.

Just my $0.02 here, your application may be special.

Mark Silver, post: 456366, member: 1087 wrote: In addition, there are a bunch of Chinese engine manufacturers who have decent engines at very reasonable price points.

I would certainly defer to your knowledge on these sorts of things, however, you do raise a good point. Signal tracking and engine algorithms are not the same thing. I know nothing of the Swiftnav receiver, for example, but I'd be curious to know how their engine is "tuned". Just because it is an L1/L2 receiver, may not translate to performance. Again, I have not idea. I do know that spec sheets do not always give a full or accurate picture of performance.

I worked with a lot of L1 only receivers back when. Today, I wouldn't even consider using L1 only anything, as my experiences ended with the ProMark3 and were similar to what you describe. Ok, but not great. As a result, it's been a long time since I even tried to use L1 only receivers. A lot has apparently changed over the past 15 years regarding L1 only performance though. I can't speak to the other brands. I can barely speak to Javad's handling of L1 only performance because of unintended experiences. I was truly shocked. I would agree that if the price difference is not too much, I'd still lean toward L2/L2c also.

First, what misconception were you addressing exactly?

Second, you are reading way too much into the Arkansas test, and this isn't the thread to discuss it.

Hi John, Thank you very much for sharing your experience !

I think there is no significant difference between GNNS raw data generated by Real or Virtual reference stations, VRS raw data might be even better because they should be multi-path free (just guessing), unlike Real reference stations raw data.(I mean, in both cases, raw data that are transmitted to rovers in RTCM 3.x, CMR etc)

I hope I will get Reach RS next week to play with, will share my experiences if it happens.
Thanks a lot,

Yuriy Lutsyshyn, post: 456492, member: 2507 wrote: I think there is no significant difference between GNNS raw data generated by Real or Virtual reference stations, VRS raw data might be even better because they should be multi-path free (just guessing), unlike Real reference stations raw data.

What? I am probably misunderstanding your statement. So you are saying that if we have a rover in the middle of three CORS stations each with baselines of 20 KM to the rover location, the VRS solution is going to be better than a 200 meter single baseline to a base in an open location?

I have not found this to EVER be the case. All you need to do is look at the VRS driven rover elevations bang around for a few minutes...

I don't want to disparage our local VRS network with this statement, but... 😮

I hope so, we had a multi frequency/constellations unit and measured a point from VRS station and than from 18km real station and coordinates agreed within 1-2 cm. We may be lucky to have a dense RTN network here which allows good modelling ion/trop delays I do not know, but VRS looks to be reliable to me in my particular case.