For background reference looking for <10mm 2d repeatability using base/rover so short (<2km) baselines for control level work. Not NRTK which is more like 20mm+ 2d repeatability here. Using Leica GS18's mostly.
Curious about using a cellular data link rather than UHF radio, not to increase range but to solve:
1) Conflict on limited public frequencies available.
2) Wondering if message length with 40+ satellites tracked may be maxing out RTCM 3.2 MSM message length?
For those using cell link for RTK GNSS wondering what extra kit I'll need to try it. Base will need a SIM card and rover either SIM card or share data off the controller which already has a SIM card. Is there anything else I'm missing?
My experience is the cell is good base n rover but it can cause latency especially during peak times. Kids out of school talking and texting etc. of course depending on the coverage that can cause issues as well cel coverage. But I would assume that if a receiver is capable of doing Network RTK full constellation then base and rover should be no problem. Here we ran SIM card in controller. Same company also in another state had SIM cards at receiver because they ran old data collector that could no longer keep up with the new SIM technology. The we’re slowly upgrading those as well.
For the simplest approach to a GSM approach, the base will need a static IP address. That way you can broadcast directly from the base. With out a static IP address you would need to set up an NTRIP caster.
As for your reasoning behind the decision to try GSM, I would agree that radio interference can be a bit of an issue. This can be mitigated, at least partially, by setting your rover to look for a specific base. I spent a summer breaking down several townships in a remote (really remote) part of northern California as part of a large dam removal project. One day after hiking/climbing through tick and mountain lion infested terrain, I went to RTK in the point via our UHF radios and found that one survey firms in the area was stepping all over my transmission. I ended up going F/S. When I got back to the base, I had it start broadcasting its ID and never had the problem again.
As for the message, I would use the proprietary format (in your case Leica 4G) unless you are mixing receiver brands. In my experience with UHF, RTCM msm is good for 3 constellations at max before it starts cutting out data. While it does not seem to be a problem with GSM, why take a chance.
Setting up an internet base can be a right pain.
Ntrip casting works well for a fixed base station where you have power to run a computer, not so good for a more mobile operation.
I know as a Leica user it wont help, but Trimble now has IBSS built into Access - which makes it a little easier
Would suggest that you look at getting your own frequency if interference on the public channels frequencies is an issue for you. See rsm.govt.nz for details.
Jim,
When using a static IP address, it is pretty easy to set up a GSM base with Leica. It acts just like a radio vs an NTRIP server. It does have a limit on how many rovers can hit it, but I've had it set to 5 rovers with no problem.
I'm already not using one of the standard three frequencies the dealers load, https://gazette.govt.nz/notice/id/2015-go884, was very helpful to find another 25khz one in 463mhz I can use!
I'm still wondering if I'm getting some interference though as getting drop out where I didn't used to, not sure if it's the RTCM 3.2 MSM not compressed enough for number of satellites or some pirate on the frequency. I'm even running a 15° mask to cut down number of satellites in data stream.
Think your a Trimble user? How many sats are you seeing with R12 base/rover and is Trimble CMRx compressed enough to get the whole message in 1 sec still?
Leica 4g copes with 44+ satellites but wondering how many more till I have to turn glonass off at both ends to keep the message short enough. Would love to just block out all the geostationary ones but haven't worked that out yet.
With Leica 4g, 8fsk, 1m whip at base, 25khz in 460mhz and internal 1w can use 44+. With RTCM 3.2 MSM, using same as above, somewhere around 37+ is too many. So yes use 4g where all Leica.
With all four constellations what are the maximum satellite numbers you are seeing?
I must be missing something as how does a base ID help if someone is already on the same channel, don't you just drown each other out still like two people talking at you at once?
Does anyone else have experience with the maximum number of satellites using UHF for Trimble CMRx or Leica 4g before you max out message length?
I have not bounced up against the limit using four constellations and Leice 4G over UHF.
As for the base ID, think of any given frequency as a crowded room with multiple conversations happening at once. Specifying a base ID basically allows the rover to hone in on the specified base's message and ignore the drone of other conversations. Not that this works situations where the frequency is being used by more than a couple of transmitters and the space is just overloaded. But how often are you running into those situations. If it is often, then you need to go through FCC coordination and get a protected frequency. I can work in the Portland metropolitan area on an itinerate frequency without much interference.
Not sure about the limit, but I regularly see 34 to 37 satellites over uhf using CMRx
Luke,
We have a documented workflow to set up IP based RTK using a static IP sim card with the Leica GS18. If you would like that information just let me know. Also Leica should have an NTRIP based solution for this type setup likely later this year. For UHF radio where satellites approach 30+ you can try changing your transmission speed to 2 sec instead of 1 - for our users that have tried this it has worked. Also only use Leica4G message format if you must, RTCMV3 allows more satellite signals to be used VS Leica4G. Lastly there is a new UHF radio standard (Modulation type/Format/OTA Speed Standards) coming out soon that should provide better performance for more satellites and a more stable working environment for mixing radio manufacturers.
Best regards,
Terry
I've never had the base ID do squat for me. I think the only reason it's really there is to help you not connect to the wrong base during your initial startup, but even after you connect you'll still get tramped on anyway.
Crowded channels seems to be a major problem in the Boise area in my experience.
I use Javad equipment. I use cellular almost exclusively for base/rover RTK. I really like it. I use RTCM MSM3.2 for my protocol. I do occasionally need to use UHF due to poor cellular service at the project site. I use RTCM MSM 3.2 for both cellular and UHF, but for UHF I broadcast at a 2 second rate and use a 9600 bps equivalent modulation. Giving the transmission 2 seconds instead of 1 second gives the radio plenty of time to push out the full message and gives the radio a chance to "rest" a bit between transmissions by reducing the duty cycle. (I think I recall that the full RTCM MSM3.2 message usually took about 1.2-1.4 seconds to broadcast) Reducing the duty cycle reduces wear on the radio by reducing heat build-up, and it also increases battery life. The slower bit rate will also improve radio range compared to faster bit rates (e.g. 14,400).
In my own experience, I saw no degradation in performance using a 2 second rate (0.5hz), over a 1 second rate (1hz). Fixed solutions were acquired just as quickly, range was just as good, precision was unaffected, etc. Your mileage may vary. With the cellular link, there is no reason to use a 2 second rate, as the bandwidth in cellular internet is more than sufficient for 1 second (and much more actually) multi-constellation RTK corrections, unless you just want to reduce the data used.
Anyone ever used Starlink for GPS?
@tnygaard that link was very interesting. Leica 4g and Trimble CMRx are clearly more efficient that RTCM offerings so improvement there would be great.
We're still able to use 25khz, not narrow banded so with RTCM3.2 I think the message length tops out in the high 30's and I'm already using a 15° elevation mask to get down to those numbers. On the GS18 using Lecia 4g/8fsk, fec off its happy pushing though 44 at least.
@shawn-billings splitting the message over 2 seconds is not something I've though about, not sure if I can set it up either with my gear. But logic is good and usually RTK seems to hold a fix as long as getting a whole message <3 seconds.
@MightyMoe I know the local Lecia dealer here is trying it, suspect it suffers the same latency problems as cellular as Telco networks seem to be built to prioritise calls and then data turns up in bigger 'lumps'. This isn't good for GNSS as we want a tiny steam of data to come through consistently and as fast as possible.
