Compare the Garmin 25LP and the Rockwell Jupiter
Nearly all tests that I performed on the Garmin 25LP and the Rockwell Jupiter OEM receivers were carried out simultaneously, making a comparison of these (nearly identical priced) receivers easy and obvious. Some tests were carried out simultaneously with a third receiver : my NovAtel reference receiver (a more expensive and more potent OEM receiver). Below the condensed results.
1. Multipath- and noise errors.
1.1. C/A Code Multipath
|25LP||2.5 to 4 m, 1 sigma|
|Jupiter||Could not be determined|
1.2. C/A Code Noise
|25LP||0.8 m, 1 sigma|
|Jupiter||Could not be determined|
1.3. C/A Carrier Noise
|25LP||4 mm to 8 mm, 1 sigma|
|Jupiter||2 mm, 1 sigma|
It is clear, that The Jupiter scores much better with its low carrier noise error.
2. Position errors
|Receiver||condition||dLat, 95%||dLon, 95%||dAlt, 95%|
|25LP||low multipath, non-smoothed||2.50||2.84||4.10|
|25LP||low multipath, smoothed||1.94||1.92||2.91|
|Jupiter||low multipath, smoothed||2.65||2.47||5.15|
|25LP||medium multipath, non-smoothed||7.0||4.0||14.0|
|25LP||medium multipath, smoothed||3.5||2.0||6.0|
|Jupiter||medium multipath, smoothed||2.0||0.8||4.9|
|25LP||high multipath, non-smoothed||23||17||48|
|25LP||high multipath, smoothed||19||14||39|
|Jupiter||high multipath, smoothed||16||16||26|
It is somewhat strange that the 25LP scores better under low multipath conditions, whereas the Jupiter scores better under medium- and high multipath conditions. The smoothing process of the Jupiter really does a good job in less benign multipath environments !
3. Tracking Performance
3.1. Noise statistics
In this graph the pseudorange noise statistics are compared. When less than
4 satellites were available, no position calculation could be carried out,
and hence no noise could be calculated. When 4 satellites were available,
the position could be calculated, but only with more than 4 satellites the
noise could be determined as well.
The 25LP is the receiver which tracked most often more than 4 sat's, but on the other hand it had the most observations with noise greater than 5 m. No clear winner here.
3.2. PDOP statistics
Where less than 4 sat's were available, no position could be calculated,
and hence no PDOP could be determined.
The performance of the three receivers is pretty much the same with just a marginally better behaviour of the 25LP.
3.3. Satellite Statistics
This graph gives the satellite PRN's together with the number of valid
observations per satellite. The total number of observations was 856, PRN
26 was tracked nearly always by all receivers.Little differences can be found
in the other PRN's, but the 25LP has a slight advantage over the other
3.4. Channel statistics
In this graph, the number of valid satellites per observation are given.
At least 4 satellites are required to calculate a position and PDOP, at least
5 satellites are required to calculate pseudorange noise. Excessive pseudorange
noise is an indication for an incorrect position calculation and allows
discarding of this calculation.
The Jupiter has 17 observations with less than 4 satellites, the 25LP only 8. The Jupiter has 49 observations with 4 satellites, the 25LP scores slightly better with 39. A more detailed analysis of the horizontal plot of the receivers in the urban canyon and under foliage (see the 25LP- and the Jupiter test pages) shows, that most out liers are observations with only 4 satellites, where no internal check on the correctness of the position could be made.
There is no clear overall winner. The 25LP scores slightly better in the
Urban Canyon/ Foliage Test, where the Jupiter scores better under less benign
multipath conditions. In my opinion the behaviour of the 25LP is better
predictable because it delivers the raw data as they were measured
(non-smoothed). Although the smoothing process of the Jupiter is very effective
nearly all the time, once it goes wrong, it goes wrong bad and unpredictable.
Your own application will be decisive in your choice, prices are nearly