WAPP user guide

WAPP user guide


This document describes how to use the WAPP (Wide Band Arecibo Pulsar Processor) for pulsar observations with the Arecibo telescope. The current system is the first unit of four independent 100-MHz correlators that can be used for fast-dump pulsar and radar work as well as regular spectral line observations. Currently (September 2000) the pulsar mode is the most developed of these applications and is in routine use for search and polarimetry work. The one remaining extension to the existing pulsar mode is an on-line folding capability and this is currently under development.

Before the observing run

It is a good idea to familiarize yourself with running the WAPP before your observing session. The current unit is located in a single crate next to the interim correlator. Ask the operator to show you where this is if necessary. The WAPP is equipped with a Mammoth tape drive that allows archiving of up to 20 GB of data per tape. Blank Mammoth tapes can be obtained from the operator. The WAPP is controlled by a PC running Linux which you can log into from the AO computer network. The PC is called ``wapp'' and the username required to login is also ``wapp''. The simplest route in is to invoke the Expect script


which will ask you for the wapp password (your scientific liason will know what this is) and log you into a tcsh session and set the display to your local machine automatically.

The fast-sampled pulsar data get written to a disk array (currently 3 disks each of 30 GB capacity). The disks are seen from the WAPP PC as /data1 /data2 /data3 so you can check the current status of thes from the WAPP PC as follows:

> df -k /data*
Filesystem 1k-blocks Used Available Use% Mounted on
/dev/sda1 34134305 29996452 2344198 93% /data1
/dev/sdb1 34134305 23929383 8411267 74% /data2
/dev/sdc1 34134305 11442637 20898013 35% /data3

in this example, we have around 30 GB of free space available for use. The amount of space needed for a single observation is roughly:

nbits * nlags * nifs * tobs / tsamp / 8388608 MB

where nbits is either 16 or 32 depending on the whether the correlation coefficients are written as 16 or 32 bit numbers, nlags is the number of lags, nifs is the number of IFs recorded (1, 2, or 4 in polarimetry mode) and tobs and tsamp are respectively the observation and sample times (in the same units). Contact your scientific liason if there is not enough space for your observations!


The WAPP is controlled via an easy-to-use Graphical User Interface (GUI). During your observing run, the GUI can be started by simply clicking on the ``WAPP'' button on the pulsar observing widget. It can, however, also be tried out off-line before the run by typing

> wappgui

from the WAPP PC. The GUI is fairly simple in appearance as seen below

The GUI allows you to configure the WAPP to a particular mode, adjust the levels to get the proper input power and start and stop the machine. What is actually happening is that the GUI is talking to a lower-level datataking program through a socket connection, the log of all the commands sent and received by the GUI via this route is shown at the bottom of the display seen in the date-stamped lines beginning with ``WAPP says.....''. All these lines get logged into an ASCII file in the directory /export/wapp/logs (as seen by the WAPP PC) which seen by Solaris machines on the AO network as the directory /share/wapp/logs. For any particular day, the ASCII file is names wapplog.YYYYMMDD. Each time the user starts a scan the correlator mode and filename is written to another logfile in this directory, scanlog.YYYMMDD. This latter file is particularly useful after the observing session when the user wishes to locate and backup the data taken.

Let's now look at the various features of the GUI in detail, starting with the top row of buttons:

As its name suggests, the ``Restart PC'' button restarts the datataking program which will be necessary in the rare event that it crashes or the user does something silly (see below). The ``Sanity Check'' button asks the datataking program to check the currently-selected correlator mode and report back in the message window whether this will result in sensible data being taken. For example, in the default mode the GUI will perform a sanity check upon startup. This resulted in the following messages to appear:

An invalid mode (e.g. too many lags) might produce the following response:

which requires the user to alter the correlator mode. This is usually done by adjusting the menubuttons or via the ``Customize...'' button, both of these are described below.

The ``Monitor'' button will start the WAPP running in the currently-selected mode but will not write data to disk. This allows the user to set levels etc. The WAPP will run in this mode until the user clicks on the ``Monitor'' button again. The ``TakeData'' button will start the WAPP running in the current mode and write to disk. Obviously, both these modes cannot be requested at the same time and the GUI will not let you do this. So, if you are hitting the ``TakeData'' button and wondering why nothing is happening, the most likely cause is that the ``Monitor'' mode is still running.

Let's now look at the control of the WAPP configuration which is done via the row of menubuttons shown below which show, from left to right, the total bandwidth, number of lags, dump time and requested integration time:

The configuration can be readily changed by moving the mouse over any of the fields and clicking to reveal a pull-down menu. For example, here are the various correlator modes:

The preset numbers of lags and the sample time are integer powers of two: 32, 64, 128... lags and 16, 32, 64 us etc. Although this makes post-processing with FFTs easier, there is no reason that the correlator cannot be run for, say, 36 lags dumping every 54.3 us. To set up such a configuration, and also to choose arbitrary integration times, click on the ``Customize settings...'' button to reveal the following widget

Note that this also allows the selection between 16 and 32-bit modes as well as the setting of polyco-style parameters for passing to the header bit selection for 16-bit modes (see later). Clicking on ``Abort'' will not change any of the settings, if you wish them to take effect, click on the ``Done'' button.

Once you have selected the correlator mode you want, you should check that the PC can cope with it by clicking on the ``Sanity Check'' button described above. If you don't see the ``sanity check passed'' message, adjust the parameters and try again. Common causes of failed sanity checks are long sampling times (which will cause the 16-bit registers to overflow), too many lags, and/or too short dump times.

Selecting the correct IF

When you're actually observing and running the WAPP GUI on the observer workstation via the AO-Control user interface you'll need to set up the levels to their optimum values before taking data. The input IF frequency expected by the WAPP depends on the bandwidth being used. In 100-MHz bandwidth mode the IF is 250 MHz:

this is the default IF that AO-Control will select for the L-band receivers. For the 50-MHz bandwidth mode, the filter requires an IF centered at 275 MHz:

When run from AO-Control, the WAPP GUI will interact with the vx-Works command line to change the LO as required. Should you need to do this manually, the relevant vw% command from AO-Control is:

vw% getwapp 50

and to change back to the 100-MHz mode

vw% getwapp 100

This information is included only for completeness - you shouldn't have to send these commands when running the GUI under normal circumstances. Note also that this section has discussed only Gregorian receivers. Should you require a 250 or a 275 MHz signal from the carriage house, this has to be done with good old-fashioned BNC cables. Consult the detailed cabling instructions in the carriage house section of the pulsar user manual.

Setting the IF power levels

The AO-Control GUI will attempt to set the optimum power levels in the upstairs and downstairs IF chains (see the AO-Control documentation for details). For fine tuning of the signal, a set of variable attenuators can be controlled from within the WAPP GUI via the two sliders and the ``Set Attentuation'' ``Query Attenuation'' and ``AutoSet'' buttons. To check and set the levels, click on the ``Monitor'' button once you have selected the desired correlator mode. During monitor mode, the GUI receives information from the datataking program about the power levels. These are displayed above the two sliders as shown below:

For each channel, the power and dB step required to reach the optimum power level are shown. In this example, the left channel shows 3.6 for the power and the dB step is 5.6. The units of power are such that 1.0 is the optimum value. To get the optimum settings while in monitor mode, simply click on the ``AutoSet'' button. This will temporarily stop ``Monitor'', adjust the attenuators by the required dB step and restart monitor mode again. In the above example, here's what happens after the AutoSet button was used.

Now the power levels are close to unity and the dB step is much less than a 1 dB. In this case you are set to take data. Hit ``Monitor'' to stop the monitor mode and hit ``TakeData''.

Normally, the ``AutoSet'' button is all you will need. For some applications, manual adjustment of the attentuators is desirable. While in ``Monitor'' mode, the attenuator settings can be moved manually by moving the slider widgets for the left and/or right channels. The settings are applied by hitting the ``Set Attenuator'' button. The ``Query Attenuator'' button recalls the current settings if you move the sliders and decide to leave them at their original values.

Taking Data

Once you're happy with the levels, stop the monitor mode and set your integration time as described above. You're all set to go! Click on the "TakeData" button on the main panel and the WAPP should begin taking data. If all is well WAPPGUI should begin counting down the number of seconds to go until the scan is over (3124 s in the following example)

Note how the WAPP has assigned a file for your observation and placed it in one of the 3 disks currently being used by the PC. The file naming scheme is sourcename.scannumber. For long scans, where the files exceed 2 GB, multiple files are opened at the beginning of the observation and the WAPP writes to them in order, starting with the lowest scan number. This is the case in the above example for the files C0341+321B.328-332 where the log shows that writing to file number 329 has just begun following completion of scan 328.

When the scan finally does end, the message window will look something like this:

Data Monitoring using SNAP

During datataking, or also in ``Monitor'' mode, you can take a look at the data quality and the passband using the SNAP online monitor program. SNAP takes and processes occasional blocks of data from a shared memory segment of the WAPP datataking program when the PC has time for the processing. Don't worry - running SNAP will NOT degrade your data in anyway! To start SNAP, log into the WAPP PC as described above and type:

> snap

This will fire up three windows. The main display window looks something like this:

where you can see the header of the incoming data. To select the next available block of data, hit the ``Next Scan'' button. To run SNAP continuously, hit the ``continuous'' checkbox. This will keep taking and processing data blocks until you click off the checkbox.

The two other windows deal with diagnostic plots. The smaller of the two (labeled ``Plot Window'') can be used to show a number of interesting displays. By default you see a total power time series of the current data block:

Click on the ``Select'' menu to look at other plots. Here is the reference band of the same block:

The total power and passband plots are useful diagnostics to run in ``Monitor'' mode to see whether any significant levels of radio frequency interference prevail. Finally, a useful plot when looking at strong pulsars is the so-called ``waterfall plot'' (radio frequency versus time) displayed in the third snap window for the current data block.

In this example, a strong dispersed pulse can be seen in each of the two IFs (the two diagonal lines in the green plot) and also the total power (the points shown in the bottom plot).

After the observing run

Once you are finished, it's a good idea to back up your data onto Mammoth tape as soon as possible. Ask the operator for a blank tape and insert it in the tape drive which is located in the front of the WAPP crate. Log into the WAPP PC as described above and use the tar program to copy individual files to tape. To find the locations of the files go to the log directory:

> cd /export/wapp/logs/

and look at the file corresponding to the date you're interested in. Here's what was taken on Sept 19, 2000, for example:

> cat scanlog.20000919
/data1/wapp/B1919+21.430 3-level, polarization 100 MHz 256 lags 256 us
/data1/wapp/B1919+21.431 3-level, polarization 100 MHz 256 lags 256 us
/data1/wapp/B1919+21.432 3-level, polarization 100 MHz 256 lags 256 us
/data1/wapp/B1919+21.433 3-level, polarization 100 MHz 256 lags 256 us

Then go to the appropriate directory and issue successive tar commands to copy the data onto tape. For example:

> cd /data1/wapp
> tar cvf $TAPE B1919+21.430

each Mammoth tape will hold up to 20 GB.

Routines for reading WAPP data

The directory /export/wapp/utility contains a number of programs that can be used to read and process the data. For looking at headers, the program asciidump reads the header information and pipes the header info to the standard output. For example, try:

asciidump datafile | more

to browse the header of a particular file. To look at the header and the raw data, the program wholedump is also very useful. This prints out the lags for each dump in octal form. For an ASCII dump of the zero lag (total power) data, the program zlagdump provides a quick means of getting a time series from a data file.

This document was generated by Duncan Lorimer on September 20, 2000