I hope you guys won't mind me harping on about an issue which the majority of you won't face, but I have a query...

Over here in the UK we have a national reference system setup by the OS, which is similar to the US CORS system, except it has (I think) a slightly denser network of active stations. I have used this many times with dual frequency receivers and it's extremely useful, now I am attempting to utilise it with Promark2 and Promark X-CM L1 receivers.

The OS Active stations are set to log at 15 second intervals (presumably to cut down on file size when downloading from the net). The Promark X-CM can be set to log at 3 sec intervals thus giving overlaps at the 15 second timers, the Promark2's with their fixed 10 second interval only overlap at the Zero and 30 second timers.

Has anyone the knowledge and experience (both of which I lack in this area!) to hazard a guess at the consequences of this for relatively short occupation times and baselines of up to 20km? My tests so far show that the Promark X-CM is much more reliable at providing fixed solutions (over the same occupation span time) due to the extra data it has to compare.

Whilst it would be perfectly possible to set up a Promark2 on a fixed point, relate this to the active base station, and then use one or more other Promark2's to collect the 'roving data', it seems a waste of resources. I am tempted just to buy more Promark X-CM's.

If anyone has any comments on this I would be grateful.

Jon Charlton.

Jon,

If I remember correctly the PX-cm is really a sub-meter unit. The PM2 is a sub cm unit, in which you appear to be trying to use it for a different purpose.

If you time on point was extended to 1 minute then the PM2 would have 3 epochs to compare. The data in the PX-cm between the 15 sec epochs is not being used. Therefore in one minute you have 5 epochs to compare.

At 20km I doubt would have to doubt the accuracy of either system for one minute obs. but this is not from experience, just a guess.

Jimbo

Jimbo,

Sorry..I didn't make myself clear, the obs time is not just one minute, you are right there is no way such a short baseline occupation would fix. When I said short occupations I guess I was meaning anywhere from 8 to 30 minutes in length.

You are also right in that the data on the Promark X-CM between the 15 sec epochs is not being used...the trouble is you cannot set this receiver for longer than 3 seconds, so this is the maximum I can observe.

The X-CM is not however a sub-meter only receiver (you may be thinking of the Promark X)..the X-CM is a true centimeter level receiver. It logs carrier phase data which can be post-processed just like the Promark 2's. It has the advantage though of having real-time features not built into the Promark 2's.

Jon.

You can try the NGS program interpo, which interpolates RINEX files to a desired interval. I've not done any experimenting with this, and I notice that it's not easy to find on the NGS site anymore, so it may be that NGS feels it's not reliable.

Jim,

Thanks for the suggestion, I will give it a try. I have always been dubious about this sort of interpolation though..I am sure I have read or heard somewhere that it is not a good idea...although where or why I am not sure!

Jon.

Jon,

While I have gotten a fixed solution on a 1/2 mile (sorry 1000 meter) line before with a 5 minute observation, I certainly wouldn't expect L1 to do this reliably all the time. We have seen no such difficulties with the PM2 with regards to fixed solutions, but we also tend to be extremely pessimistic about our occupation times. What's an extra 15 minutes when it takes 30 to get to the site and set up? Also, it is my understanding that L1 requires a substantial change in s/v geometry to get a sub cm solution. I believe this to be true based on my own experimentation on a process known as psuedo kinematic. (A quick search on this board will probably tell you more than you wanted to know about pk). But anyway, the change in sv geo is more important than the quantity of data. I really don't know what the limit is on less data vs. good solutions is, but I feel pretty confident that unless you are measuring something under 3000 meters that 8 minutes isn't enough for any L1 system, and 30 minutes is just barely enough for a 10 km solution. Maybe I am misunderstanding your point. I do appreciate the heads up on the week roll over problem, and hope you don't take my post as an offense to you.

Shawn

Shawn,

Thanks for the reply, and I certainly don't take any offense..all info is useful! What is at fault is my ability to put a clear and concise post together! lol

I totally agree with what you say...when I mentioned 8 minutes occupation times, I was referring to trials we have done for baselines under 3km in length. We have found that generally these do provide fix solutions with good accuracy, but we would always increase this to at least 15 minutes for 'real' use, even at this very short length.

The thrust of my question was really this: If you have two sets of receivers, one set measuring at 15 sec intervals, the other at 30 sec intervals, what would be the implications of this....ie would the receivers at 15 secs require significantly greater obs times, and is there a baseline length at which this difference disappears.

Maybe I should also have explained my interest in this short baseline/short occupation scenario. As a company we undertake two types of static work, one is control establishment where we often use dual frequency receivers and where occupation times are not so critical. The other is where we have to gather a large number of points in a relatively small area, where time at each point comes into play. We have used RTK extensively for this but are always looking at alternative strategies.

I appreciate you guys taking time to reply to my ramblings...I dare say I should really go away and read all the books and manuals on the subject...but this is easier :-)

Jon.

don't worry about the concise part to much.......providing plenty of info helps the rest of us. Not in just understanding your problem, but in how it relates to our own procedures, thoughts and ideas.

And yeah, it's a lot easier here with a bunch of people, many quite capable, helping out. Nothing wrong with that in my book !

Jimbo

PS: Of course you might not want to post epistle's like some (me !!!) have done in the past.

And you are right, I did have the wrong receiver model in mind...me and my alphabet......
Modified By James Webb on 11/8/2001 at 6:46 AM

I wasn't aware the NGS had done away with Interpo, but I can bet they had a good reason...
The SV's, Iono, ect change 'in the blink of an eye' and interpolation is really in the realm of GIS grade receivers and should not be used for static or quality work..

Bottom line..If you have receivers logging at 30 second intervals and some at 15 second intervals, then you will have to get enough data for the 30 second intervals to process your baselines..They (your software)should 'toss' the intermediate records out and not use them in any ambiguity solution...

I've done extensive testing on short baseline with short occupations (by cutting the data down to minutes)..99% of the under 3km baselines process fine with 2 minutes of data (meaning they correctly solve the integer), but I would never survey this way...Our receivers log 20 minutes for a fast static...I go longer, but never shorter..

RTK is a complete other beast and doesn't have the constraints that static and fast static have, since the OTF solution is maintained during the survey and unless you lose lock, then the interger stays fixed..(usually)..

Dual Frequency receivers can usually 'fix' in less time (my opinion) and based on the number of SV's, PDOP, then you can easily get by with 7 minutes of data for a short baseline solution..

TM

And one other thing I might add..I have yet to see any manufactures spec's that did not err on the side of caution...Phil and I both have 'cheated' and with experience, then you know when you can and when you can't..

My 2c worth:

Jimbo, the old Promark X-CM provide carrier phase data usable in MSTAR for fixed solutions AND usable in SOLUTIONS..I do it today with a Promark X, a decimeter Ashtech Reliance (code and carrier), and Promark2's. Mixing and matching requires attention to antenna heights, as the Promark X antenna is not a tested unit by NGS to my knowledge, so I have to use my own antenna testing parameters (test and retest:-)

As for interpo, I have to admit I use it regularly on carrier phase data for fixed integer solutions. I feel strongly that interpolating 30 second or even 20 second data is a stretch and provides WORSE results than using the original 30 second data by itself...BUT interpolating 10 second data to five or even two seconds to fill in the gaps DOES result in stronger solutions, IMHO.

BE CAREFUL!!! Better statistics, ie . std dev or confidence intevals (in the case of MSTAR) does not mean better vectors...Use the adjustment software to identify bad vectors and toss them after analysis....