Tech Support FAQ for SiteSense and SiteStumbler

Why aren’t some/any AP’s showing up in the Wireless Client window?

Make sure the intended wireless card/driver pair is selected from the Device menu in the Wireless Client window. Next, make sure the intended survey Profile is active for the wireless card. Profiles are setup via the utility for the wireless card. In the profile, select the desired RF band (2.4 GHz and/or 5 GHz), the power output, and specify an SSID for scanning. For a passive survey, it is usually best to specify a false SSID (one that doesn’t match any active SSID’s in the survey area). If this does not give the desired results, try a blank SSID. For an active survey, specify the particular SSID as well as the necessary authentication and authorization credentials. The Wireless Client obtains Wi-Fi data from AP’s by sending out probe requests across all channels (some client utilities allow one to select the scan channels) and listening for the probe responses. If there is no data from an AP known to be active in a particular area, then that AP is not responding to the probe request. Modifying the target SSID in the profile may help. Some wireless networks can be setup to ignore probe requests from unauthorized client devices. In addition, some systems can be configured to specifically ignore NetStumbler packets. For the survey, these settings will need to be temporarily disabled in order to acquire the necessary data.

Why are all the reported Noise values -100 dBm?

An unsupported card utilizing the NDIS 5.1 driver will report Noise as -100 dBm. No Noise data is retrieved, but the -100 dBm value is passed to the client window. In addition, the signal strength values obtained via NDIS are probably not accurate. The most accurate signal strength and noise data is obtained using a supported wireless adapter.

In SiteSense, why is there no Detail data displayed during mouse-over of a survey path even though the path shows coverage?

There is a known bug that results from the use of the Detached Mode feature and attempting to view the coverage data in the Analysis screen during a survey. Any path information acquired after using Detached Mode can be offset in the Analysis screen so that mouse-over of that section of path does not display results in the Details window. The Details can be displayed by moving the cursor in the area around the path (not on the path itself). If the file is saved and viewed at a later time, the offset anomaly is gone. The display offset is a screen artifact, the underlying actual position data is true. Pending the bug fix, to prevent visualization of the display offset while surveying, acquire Detached Mode data at the end of a survey, i.e. after all Attached Mode data is obtained, or simply save the file and re-open it.

I start and end the survey at the same physical location, but the starting and ending points don’t match up in SiteSense, why not? What can I do to make sure they match?

This is a consequence of measurement drift and the Snap To algorithm used in SiteSense. SiteSense measures the rotation of each of the rear wheels to determine distance traveled as well as direction. Any difference in rotation between the wheels is perceived as a change in heading. Consequently, subtle changes in floor slope can be perceived as a heading change even if there was no change (i.e. still walking straight down the hallway). The Snap To algorithm draws a straight line on the screen even if the intended and actual headings deviate (up to a threshold value). All the real position data is logged, but the Snap To algorithm filters it for display purposes. Choosing Snap To "None" for a previously acquired data set will display the actual path detected by SiteSense. Sometimes this will be enough to get the start and endpoints to match up. Discrepancies are also caused by drift accumulated through the survey process. Drift occurs any time there is a difference between the actual heading and intended heading. The larger the difference and the longer it occurs result in greater drift. As the survey progresses, the drift errors increase. The extent of drift can be minimized by resetting the x,y coordinates during the survey. This is accomplished using the Snap to Waypoint functionality. Snap to Waypoint resets the SiteSense x,y coordinates to the coordinates of a previously establish Waypoint. To minimize drift, break the survey area into multiple paths, and always start a path from a waypoint that has minimal propagation error. The Starting Waypoint is the best, since it has no positional error. With this technique, it is good practice to initially set up a survey boundary path with a large number of "error-free" waypoints.

Windows (or the application) crashed during the course of data collection, is the survey data automatically saved?

SiteSense (versions 1.2 and earlier) and SiteStumbler (version 1.0) do not have an Automatic Save feature. It is up to the user to save the data as the survey progresses. It is a good idea to save multiple versions of the data set during a survey by using the Save As option.

Can I use the software to do a quick check of AP’s operating in an area without hooking up the cart?

The Wireless Client part of the software (both for SiteSense and SiteStumbler) is a standalone application and can be used to get a snapshot of AP activity. The Wireless Client is a proprietary version of the popular NetStumbler application.

How do I carry out an active survey?

For an active survey, the wireless adapter must be associated with the wireless network of interest. Therefore, a profile must be setup and activated that enables this association (i.e. the necessary authentication and security parameters must be specified in the profile). Once associated, a roaming Ping Test can be carried out during the survey. The Ping Test can be configured (IP address, packet size, etc.) under Options in the Wireless Client window. It is best to specify an IP address of a wired device (switch or router) that resides behind the AP’s in the network. The Ping Test can be started and stopped with the Start/Stop toggle button in the Wireless Client window. During the test, the statistics for the associated AP are displayed live. Each time the Client roams to a new AP, the statistics reset. The ping data is saved to file, but is unavailable for coverage map display in SiteSense versions 1.2 and earlier. The ping results can be saved for the report by adding the data to Notes during the survey. To restrict data capture to only the associated AP, click the Lock-on toggle button in the button bar of the Wireless Client window.

How do I carry out a design survey?

For a design survey, there is no existing installed wireless network and temporary AP’s are used to map out the RF characteristics of the facility. For each AP position, the AP is given a unique name (click in the Name column within the Wireless Client window at the row corresponding to the MAC address of the AP). Once the coverage has been mapped out for a particular AP position, the survey is paused, the AP moved to a new position, the AP is renamed in the Wireless Client window, and then the survey is resumed. This procedure is repeated until the entire facility is mapped out.

The software auto-placed the AP, but the location is wrong, how can it be re-positioned?

To re-position an AP, simply click on the AP icon, hold the click and drag the icon to its desired location.

How does the software AutoPlace the AP’s?

The auto-place is based on a triangulation algorithm. The algorithm uses the three strongest signal strength readings (above a specified threshold) from the AP to calculate its estimated position. By default, the threshold is -50 dBm. If the software doesn’t have 3 readings above the threshold, the AP will not be auto-placed. For multi-floor buildings, the -50 dBm setting works well to prevent placing AP’s on the wrong floor. For warehouses or outdoor settings, the threshold may need to be lowered to get the software to auto-place AP’s. This setting can be accessed via the helium.properties file that resides in the same folder as the application. A text editor application such as Notepad can used to edit the file. The AutoPlace setting is found at the line : strategy.minimumSignalThreshold=-50
The accuracy of the algorithm increases as the threshold value increases (e.g. -40 is better than -60).

Why don’t the Expanded or Gradient Maps show coverage from AP’s that are located on floors other than the current view floor?

For SiteSense versions 1.2 and earlier as well as SiteStumbler version 1.0, the projected coverage maps (Expanded, Gradient) are calculated in two dimensions only. Therefore, even though an AP may have an x,y position on another floor, its projected coverage on the current floor will not be displayed. Three dimensional coverage can only be viewed on the maps that use the actual path (e.g. Max Signal Strength).

I’m using a supported card (Cisco CB21AG) but I’m getting unexpected behavior (e.g. signal strength readings at a position near an AP are too weak) in the Wireless Client, what’s wrong?

Under some circumstances it has been found that the Cisco card/Cisco driver combo gives unrealistic results in the Wireless Client. The problem seems to be isolated to passive surveys (i.e. blank or false SSID). The phenomenon has not been readily reproducible, but it has been encountered more than once. The Cisco card uses its own proprietary driver instead of the default Atheros driver. Each driver has its own unique conversion scale for converting RF energy to RSSI values. Until the issue can be resolved within the Wireless Client, it is best to be cautious when using the Cisco card, i.e. make sure the readings make sense before you acquire survey data. Experimentation with profile settings (e.g. SSID used for scanning) may help the situation. In addition, a custom Atheros driver is available which can help with this problem. Contact Helium Networks Support (412-231-7770 x6) for the custom Atheros driver.

After gathering and saving survey data with SiteStumbler, when I attempt to open the saved file, the following error is displayed: "Unable to parse the coverage map file xxxx.hil. No access point with name ‘yyy’ and bssid 'zz:zz:zz:zz:zz:zz' exists in this CoverageMap". The data looked OK when it was analyzed at the time of the survey, what went wrong?

For some reason, some access point data is not being written to the file. This problem has not been readily reproduced, but it has been encountered more than once. The survey data is saved, however the particular AP information is missing from the AP list within the file. The file can be resurrected by adding the missing AP data. The AP list is present at the beginning of the file. Using a text editor such as NotePad, open the file and then duplicate an existing AP entry. For the copied listing, change the values for the SSID and BSSID to correspond to the missing AP. If the channel is known, change this value also. Save the file under a different name.