There are usually two kinds, those having known heights, called bench marks, and those whose latitude and longitude have been determined, called horizontal control points. Usually they are bronze tablets, which may be set in concrete, rock outcrops, large permanent structures or affixed to pipes driven into the earth.

Typical projects that would use our surveyed markers are: the production of topographic maps, resource development, land administration, and scientific studies, such as determining the size and shape of the earth, crustal deformation and the movement of tectonic plates. Collectively they form a national reference system, which links together various forms of surveys throughout the country and offshore.

In each of the provinces a Responsible Agency is responsible for the construction, maintenance and access policies for the sites within their geographic area. It is recommended that the Responsible Agency be contacted before visiting any calibration site to ensure the site is still active and suitable for calibration purposes.

The backward solution uses the best estimates of the carrier-phase ambiguities obtained from the forward solution. The backward 'smoothing' improves the estimates of the parameters in the initial portion of the forward solution and following data breaks, when ambiguities are converging and the solution is not yet optimal. Normally ambiguities converge to their correct values within 30 minutes of the session start time or within a few minutes following a data break. Convergence time may vary depending on satellite geometry, quality of the observations, the duration of a data outage and the validity of the a-priori estimates of the parameters.

A 30 second sampling is usually sufficient for PPP in static mode.  While a higher sampling rate can improve cycle slip detection, it does not contribute to accelerating the solution convergence. While PPP can process data sampled at rates up to 10 Hz, the main issue when processing high rate data is the size of the file which increases data upload time to CSRS-PPP.

In order for PPP to provide its optimal accuracy with dual frequency pseudo-range and carrier-phase data, it needs to properly resolve the phase ambiguities.  This normally requires a minimum of 30 to 90 minutes of data -- the convergence period -- and is a function of the number and geometry of the satellites and observations quality.  Convergence takes longer in kinematic mode.  Solution convergence to centimetre precision should, therefore, normally be achieved with 4 to 6 hours of data (see Figure).  Lower precision levels can be achieved more rapidly and may be adequate for certain applications

Users may validate convergence time using their own equipment and operating environment by collecting and processing some sample data in the area they intend to work.  Using a point with a know position to perform the convergence test would also help confirm what accuracy you can expect from PPP at different latitudes.  Given the 55 degree inclination off GPS orbital planes, satellite distribution above a user’s horizon varies significantly from the equator to the poles (show sky views at equator, mid-latitude and pole) 

Depending on the accuracy and confidence you are trying to achieve with PPP, you may also consider collecting more data than the minimum required.  For example re-occupying points more than once for 4 to 6 hours or single, longer occupations (for example 24 hours in case some of the data is bad) would help increase your confidence in the results

You could try running your file through UNAVCO's TEQC program to determine the problem.
 
"TEQC (pronounced "tek") is a simple yet powerful and unified approach to solving many pre-processing problems with GPS, GLONASS, Galileo, and SBAS data:

- translation: binary data reading/translation of native binary formats (optional RINEX file creation for OBS, NAV, and/or MET files or optional creation of BINEX)

- editing and cut/splice: metadata extraction; editing, and/or correction of RINEX header metadata; or BINEX metadata records; as well as cutting/splicing of RINEX files or BINEX

- quality check: quality checking of GPS and/or GLONASS data (native binary, BINEX, or RINEX OBS; with or without ephemerides)

These three main functions (from which teqc gets its name: translation, editing, and quality check) can be performed altogether, in pairs, or separately."
http://facility.unavco.org/software/teqc/teqc.html

Executables for teqc
http://facility.unavco.org/software/teqc/teqc.html#executables

The table provides performance indicators after convergence which would normally have been achieved with a 4 hour observation session.

PPP Performance after convergence – Single and dual frequency – Static and kinematic

 Note (1): Quoted PPP code-only performance is for surveying grade receiver. Performance may vary for other types of receiver.

No ACP or base stations are used by PPP. PPP uses the so-called point positioning approach (users don’t need to be near an ACP, IGS or CORS station). Unlike differential GPS methods, PPP does not require tracking data from any other GPS receiver. Instead PPP uses a-priori values of the GPS satellite coordinates and estimates of the state of their clocks that have centimetre precision. These precise products are usually available from NRCan in hourly files available 90 minutes after data acquisition. The precise orbits and clocks remove a large part of the GPS errors. In addition, PPP processing must also properly account for several other effects on the position of the GPS receiver.

Below are links to papers describing how PPP works.

GPS Precise Point Positioning Using IGS Orbit Products (2000)

GPS Precise Point Positioning with a Difference (1995)

Precise GPS Point Positioning: the Future Alternative to Differential GPS Surveying (2002)

CSRS-PPP User's Guide

Ellipsoid model = GRS80
Semi major (A) = 6378137.0
Semi minor (B) = 6356752.3141
Flattening (F) = 1/298.257222101
1st Eccentricity Squared (ESQ) = 1-B/A*B/A or (A squared - B squared)/A squared = 0.006694380022901

NAD83 - North American Datum 1983.  The reference system adopted in 1986 by the U.S., Canada and Central America.  It is based on the GRS80 (Geodetic Reference System 1980) ellipsoid.  Although originally thought to be geocentric, it is now known to be offset from the true geocenter by about 2 m due to the limited accuracy of absolute positioning at the time.  There is only one NAD83 datum and it is considered fixed to the North-American tectonic plate.

With advances in positioning technology, new realizations/adjustments of NAD83 were produced.  The three main realizations are:

First realization (from 1986) based on the Doppler reference frame (same as WGS84 at the time).

NAD83(Original) adjustments were based on adding a set of VLBI constraints to the first NAD83 (1986).  This set of constraint coordinates was used as the definition of NAD83(Original) for horizontal control adjustments until 1995.  Note that these adjustments were 2-D horizontal (latitude and longitude) solutions only.  Some provinces still use these solutions for their horizontal networks but most have or are moving to the NAD83(CSRS) realization of the NAD83 reference system.

The absolute accuracy of NAD83(Original) varies from several cm to over a couple of meters, with an average of about 0.3 meters.  However, these errors are locally fairly coherent so that the relative accuracy can be much better.  If a survey marker's published coordinates are labelled simply NAD83, they are considered NAD83(Original).

NAD83(CSRS) - North American Datum 1983 (Canadian Spatial Reference System).  This is an updated high-accuracy, three-dimensional realization of the NAD83 reference system.  In agreement with the US, this realization is defined in terms of a common procedure for transforming from recent versions of the International Terrestrial Reference Frame (ITRF).  This realization is based on a 7-parameter transformation from ITRF96. Incremental transformations between ITRF96 and other ITRF realizations are used to update the transformation to any version of ITRF. NAD83(CSRS) is consequently a more accurate and stable realization of NAD83 that can be accurately referenced to other global and regional reference frames.  The absolute accuracy of NAD83(CSRS) is of the order of a couple of cm or better, the same accuracy as the most recent versions of the ITRF.

NAD83(Original) and NAD83(CSRS) are based on the same fundamental NAD83 reference system. They are merely different realizations or adjustments with different levels of accuracy.  Taking into account these errors, NAD83(Original) and NAD83(CSRS) are completely compatible with one another.

Provincial survey agencies have produced regional grid shift files to transform positions from NAD83(Original) to NAD83(CSRS).

In the late 1980's with advances in space-based technologies (Satellite Laser Ranging, GPS, VLBI) a much more accurate spatial reference frame was determined, the ITRF (International Terrestrial Reference Frame).  ITRF is a cartesian (X,Y,Z) system whose coordinates can be expressed both as (x,y,z) and lat/long and height above the GRS80 ellipsoid.

ITRF is dynamic and continues to be refined as more space-based data is acquired.  New realizations of ITRF have been published on a regular basis (ITRF88, ITRF89, ITRF90, ITRF91, ITRF92, ITRF93, ITRF94, ITRF96, ITRF97, ITRF2000 and ITRF2005). ITRF2005 is the current version.

Although ITRF2000 is quite stable, all subsequent realizations of the ITRF will see ITRF station coordinates slowly shift due to tectonic plate motion as well as crustal motion (uplift / subsidence).  The NAD83 reference system was fixed to the North American tectonic plate and the NAD83 coordinates of Canadian stations are not affected by tectonic plate motion.

WGS84 (World Geodetic System 1984) is a spatial reference system that was created to be used solely by the Global Positioning System (GPS).  From 1986 to 1994, WGS84 was considered the same as NAD83.   Both were based on the same Doppler reference frame and were therefore compatible with each other, at least to the accuracy of the Doppler reference frame (about a meter).  It was not uncommon for the datum to be called "NAD83/WGS84".

Compared to the ITRF, it turns out that "NAD83/WGS84" was mis-scaled, its origin was offset and its axes were misaligned resulting in a coordinate shift in Canada of 1 to 1.5 m.  Since GPS played a major role in refining the ITRF, it made practical sense that the GPS reference frame be truly geocentric and aligned with ITRF.  And so on Jan 2, 1994, WGS84 was realigned to ITRF91 and was no longer compatible with NAD83.  This new realization of WGS84 was called WGS84(G730) and could now be called "ITRF/WGS84".  The total coordinate difference between WGS84(G730) and NAD83 in Canada amounted to approximately 1-1.5 m with a magnitude and direction varying according to user location.

WGS84 was intended to remain compatible with the dynamic ITRF.  So on Sep 29, 1996, WGS84 was once again realigned to ITRF.  Based on the more accurate ITRF94, this new realization was called WGS84(G873). Finally, in Jan 2002, WGS84 was once again updated. Called WGS84(G1150), this version was realigned to the even more accurate and stable ITRF2000 and is still in use today.  WGS84(G873) and WGS84(G1150) provided only small incremental position shifts and for the majority of GPS users the difference between WGS84 and NAD83 today is essentially the same as it was in 1994, approximately 1.5 m.

With low-cost recreational GPS receiver this 1.5 m datum difference is inconsequential.  But when using higher-end GPS receivers and/or correction techniques that provide metre accuracy and better it must be considered.

WAAS (Wide-Area Augmentation System)…WAAS is a US real-time wide-area differential correction system built for aviation and now operated by the Satellite Navigation group of the  FAA.  Most GPS receivers can receive WAAS corrections and they can be as accurate in Southern Canada as in the U.S.

The near real-time Vertical Protection Level (VPL) map for the US and Canada provides aviators an estimate of the accuracy (vertical) provided with WAAS corrections.  It's a near-realtime indicator of the accuracy of WAAS across North-America.  See "New WAAS GeoStatus" to find out about the addition and availability of new WAAS satellites.  WAAS-corrected positions are in WGS84.

The short answer is "No, you are most likely still in WGS84".

All GPS receivers produce positions in WGS84 (in the "current" WGS84 realization).  However GPS users worldwide usually wish to obtain positions in their country's adopted reference frame (local datum) instead of WGS84.  Most GPS receivers let you switch from WGS84 to a selection of local datums using an on-board datum-shift table.  This table most often consists in a 3-parameter (dX, dY, dZ) transformation, equivalent to 3 translations (no axis rotation).  Most GPS manufacturers use datum shift values obtained from technical report "NIMA 8350.2" published by NGA (National Geospatial Agency).  See June 23, 2004 update, the latest publication.

Before January 2, 1994 WGS84 and NAD83 were considered the same and the NGA published WGS84-NAD83 datum shift parameters of (dX = 0, dY = 0, dZ = 0), indicating a "zero-shift".  On January 2, 1994 WGS84 was realigned with ITRF, instantly affecting shifts to all of the World's local datums, including NAD83, by about 1.5 m.

3-parameter datum shift values (dX, dY, dZ) can apply XYZ translations but no rotation therefore a single set of values cannot cover all of Canada.   NGA kept the WGS84-to-NAD83 datum shift as a "zero-shift" but now publishes the values as (dX = 0 +/- 2m, dY = 0 +/- 2m, dZ = 0 +/- 2m) to account for a +/- 2m datum error.

GPS manufacturers still all use by default the old "zero" WGS84-NAD83 datum shift, leaving you in fact in WGS84.

There are three basic approaches:

7-Parameter Shift

The more complex 7-parameter shift is actually the easiest to use as only one set of values can be used anywhere in Canada.  The following values were published in the 1998 paper:

Realization and Unification of NAD83 in Canada and the U.S. via the ITRF [PDF, 95.1 kb, viewer]

Values today would differ very little as ITRF is stable (at the cm level). These values could be entered in any GPS or GIS software to produce a good datum shift from WGS84/ITRF to NAD83(CSRS).

NRCan's online application TRNOBS applies the correct 7-parameter transformation and the values used are included in the TRNOBS output.

3-Parameter shift

The less complex 3-parameter shift is used in many GPS receivers and software.  A user can enter a “user-defined” datum shift (dX, dY, dZ) to transform from WGS84 to any local datum.   Recreational GPS receivers usually accept (dX, dY, dZ) values only to the nearest metre.

A 1.5 m coordinate shift can only effectively be applied in a GPS receiver or software that accept values to the decimetre or better.  Many geodetic receivers accept values to the millimetre or better.  However one set of 3-paramater values is valid in only one location.

Best results are obtained when using "specialized" GPS equipment (marine GPS units with integrated Coast Guard DGPS, GPS receivers with integrated CDGPS, Network RTK equipment).

"Non-specialized" GPS equipment can get real-time NAD83(CSRS) corrections by using a separate radio (Coast Guard beacon radio, CDGPS radio, cellphone with NTRIP client).  However users must take special precaution to understand how their equipment and software handle datums.  Most GIS software (such as ArcPad), use as input a GPS receiver's real-time NMEA (National Marine Electronics Association) data stream (standard format for position data).  The NMEA “GPGGA” string contains a data field (7th) that indicates if a position was corrected (code=2) or not (code=1) but it has no field to indicate the datum.

When positions are un-corrected or WAAS-corrected, they are in WGS84. When a NAD83(CSRS) real-time correction is applied then the corrected position is in NAD83.  Most GIS software expect NMEA positions to always be in WGS84 and could not detect the shift to NAD83(CSRS). Typically the NMEA stream is fed (for example) into ArcPad for logging as a shapefile.  When NAD83(CSRS) real-time corrections are used, identifying with certainty the datum of the positions takes care and can sometimes be problematic, especially if the NAD83(CSRS) corrections are intermittent.

Using "specialized" fully-integrated systems is recommended.

GPS receivers can only compute latitude, longitude and ellipsoidal height.  Because most GPS users want orthometric (or mean-sea-level) heights, GPS receivers often incorporate a lookup table of geoid height values (also called geoid separation or geoid undulation).  Unfortunately due to storage constraints in GPS receivers these values may not be very accurate.

For example many receivers use the DMA 10x10 geoid model which stores one geoid height value per 10° x 10° area.  In fact geoid heights can vary a great deal within a 10° x 10° area.  It is not uncommon for a GPS receiver's geoid heights to be erroneous by several metres.

Realtime corrections will improve your ellipsoid heights however the receiver's geoid heights will still be erroneous. Furthermore using NAD83(CSRS) real-time corrections introduces a vertical datum error that must be considered.

GPS receivers are designed to operate in WGS84 and the on-board geoid heights table is always in WGS84 (i.e. applies the separation between the geoid and the WGS84 ellipsoid).

When using uncorrected GPS or WAAS-corrected GPS, the receiver computes WGS84 ellipsoid heights (h), applies WGS84 geoid heights (N) and produces orthometric heights (H = h - N).  Keep in mind that inaccuracies in the geoid heights will be found in the orthometric height

When using NAD83 corrections a GPS receiver computes NAD83 ellipsoid heights but still applies WGS84 geoid heights which introduces a Vertical Datum Error.

The height difference between the WGS84 and NAD83 ellipsoids varies across Canada from +1m to -1m.  Depending on the accuracy the user is looking for this vertical datum error may or may not be a problem.

Users of CDGPS or Coast Guard Beacon NAD83 realtime corrections could expect (NAD83) ellipsoid heights to be accurate to 2 m at best so the +/- 1 m WGS84-NAD83 height difference may be considered negligeable.  However the on-board geoid heights table may be a large source of error.

To compute the best orthometric heights: