Mapping the RCU number to the LBA/HBA number

Each polarisation from each antenna is processed by a separate channel. The electronics on this channel is called a Receiver Unit (RCU). KAIRA has 96 RCUs (numbered 0..95).

When looking at the signals coming into the system, they are labelled by their RCU number in the first instant. One needs to map this to the antennas to work out which antenna/polarisation is being shown.

The polarisation mapping is straight forward... even numbers for X-polarisation and odd numbers of Y-polarisation. However the match between the RCU numbers and the antenna numbers is a bit more confusing. Yes, there is a logic, but it is not obvious. Sometimes it is simply easier to look it up!

So, here is the mapping of RCU to antenna number for KAIRA (the scheme is different for other LOFAR stations).

For where those antennas are on the actual site, refer to the array layout map.

System status

Following the storm and the repair work, we carried out remote system checks for the bandpasses for all antennas. Here are the results, starting with the High-Band Antenna (HBA) array:

As seen, RCUs 80 and 81 (= tile #H75) remain dead. These are the two channels from this destroyed tile.

Here is the Low-Band Antenna (LBA) array results:

Here you can see RCUs 68 and 69 (= aerial #34) remain dead. This aerial was destroyed and we have not even found the remains of it! Following repairs, there remain two other damaged signal channels. These are RCU 11 (= aerial #L05, Y-polarisation) and RCU 42 (= aerial #L21, X-polarisation). These two aerials have suffered broken wires and are thus impared in this polarisation. Temporary repairs have been effected, but more work remains to be done.

However, in total, we only have lost 6 out of 192 signal channels. Throughout the storm and the repair work, KAIRA has continued to observe our riometry and interferometry programmes. That's pretty good.

Converting between subband and frequency

UPDATED : 28-Oct-2013

As KAIRA uses LOFAR hardware, it makes use of the LOFAR signal processing and nomenclature. When dealing with the signal from each polarisation of each antenna, the radio spectrum is split into 512 channels called "subbands".
Typically, the sampler clock runs at 200 MHz, this gives a Nyquist zone of 100 MHz bandwidth. This is what is split into the 512 channels.

So, to convert from subband number to frequency, you need to multiply by the sampler clock, and divide by 2 and the number of subbands. This gives the center of the subband. To get the top end of the subband, you need to use (subband + 0.5) and, for the lower frequency of the subband, you need to use (subband - 0.5). It is this middle frequency that is usually required. Thus:

     Freq_mid = subband * 200.0 MHz  /  2  /  512

The following is a python programme that prints the lower, centre and upper frequencies of each subband.

#!/usr/bin/python
# Make a table of subband number and equivalent 

# low / centre / high frequency for the RCU mode 3.

num_subbands = 512   # Number of subbands (fixed at 512)
clock_freq = 200.0   # Sample clock freq. MHz (typically 200)

print "Units = MHz"
print "Clock frequency =",clock_freq
print "Subband  LowFreq    MidFreq   HighFreq"
for subband in xrange(0,num_subbands,1):
    freq_low = (subband-0.5) * clock_freq / 2.0 / num_subbands
    freq_mid = subband * clock_freq / 2.0 / num_subbands
    freq_high = (subband+0.5) * clock_freq / 2.0 / num_subbands
    print "  %3d   %8.4f   %8.4f   %8.4f" % \
          (subband, freq_low, freq_mid, freq_high)
print "Done"
# End of file

So, for example, for RCU-mode 3, the centre of subband 256 is exactly 50 MHz

For subbands 0 and 511, this doesn't quite hold due to the way the sampling is done. However, these two subbands are completely outside the  filter ranges, so they would never be used for operational observing anyway.

Using LOFAR LBA and HBA arrays simultaneously

Most LOFAR stations have two antenna fields. These are the Low-Band Antenna (LBA) array and the High-Band Antenna (HBA) array. The LBA and HBA antennas are capable of receiving a range of 10-90 and 110-270 MHz respectively. However, the signal processing is arranged slightly differently. Each "channel" in the signal processing system has three inputs into which the antennas are connected. These receiver units (RCUs), then have several signal paths that can be used, each of which switch in different filters. Thus, the different antennas and frequency bands can be selected and sampled.

Because each individual receiver unit has both an LBA and an HBA polarisation connected to it, it is not possible to observe with both of these simultaneously. Furthermore, because of the filters used in the RCUs, it is not possible to observe across the entire frequency range of the HBA simultaneously either. You must select one of these "RCU modes" for each given channel.

•  RCU mode 3 = LBA input, 10-90 MHz filters
•  RCU mode 5 = HBA input, 110-190 MHz filters
•  RCU mode 7 = HBA input, 210-270 MHz filters

There are other modes as well, such as RCU mode 6, which uses not just different filters, but also a different clock rate (160, rather than 200 MHz) in order to sample the frequencies around 200 MHz (which would normally be aliased in the other modes). The point is that each RCU is limited to a single mode. This has led some to believe that LOFAR stations cannot observe with the LBA and HBA simultaneously.

Actually... they can.

Here at KAIRA, we have been using combinations of modes. In what we refer to as "RCU mode 357", we have been observing with RCU modes 3, 5 and 7 simultaneously. The modes are interleaved, so that there is still a distribution of antennas for each; this allows beam-forming to take place.

Of course, this is not for the faint-hearted, and one must be careful in mapping the channels and powering-up the HBA tiles in such a way so as not to cause power supply failures. However, when done correctly, multiple beams across all bands can be formed giving frequency coverage over nearly the entire VHF band. Here is one of our first results:

A small sample of data from KAIRA, ranging from ~15 to ~275 MHz. Each frequency channel has been divided by the median to bring out the radio features. The bright arcs are ionospheric scintillation during observations of Cas A. (D. McKay-Bukowski, J. Vierinen, and R. Fallows.)

Already we have seen some interesting things and we will be reporting these over the next few days.

LBA bandpass spectra

We're now in the midst of some intensive commissioning of KAIRA. Some of us have been working on the High-Band Antenna array and others on the Low-Band Antenna array. Today's report is about this LBA work. One of the things we need to do as part of the commissioning is confirm that all the low-noise amplifiers are generating the correct signal outputs. For each aerial there are two outputs: one for each polarisation. The bandpass spectrum for each of them should look the same (a bandpass spectrum is a plot showing the intensity received as a function of frequency, for the range of frequencies that the system can measure). In the case of a good working antenna, we should see something like this:

The lowest part (0-10 MHz) is filtered out. Then there are various short-wave radio frequency interference. The main part of the band rises up and then tapers off at the top end (going into the FM radio band).

However, we've noted that some channels are a bit weak:

And some channels are quite low across the main part of the band.

Before we can operate the entire LBA array, we will need to resolve these problems and ensure that all components are working correctly. This sort of trouble-shooting and tuning is a standard part of commissioning these sorts of systems.

RCU boards

The receiver unit (RCU) boards are the first part in the signal processing chain as the signals come into the RF-container. There are 96 of these in total in the KAIRA system (which is the equivalent of a Dutch Remote Station). There is one RCU per polarisation per antenna. Each RCU has three input channels. These are the high-band antennas (HBA), low-band-high (LBH) and low-band-low (LBL). On KAIRA, and other international LOFAR stations, only the HBA and LBH channels are used. These have the signals from the HBA and LBA arrays respectively. The third channel is for even lower frequency antennas (such as 10-30 MHz). At this stage, these antennas have not been realised, although there is some prototype work on this being carried out at the French LOFAR station (FR606, Nançay). On Dutch stations, the second part of the LBA arrays utilise this third channel.

The RCU subracks, with one of the RCU boards pulled out. (Photo: D. McKay-Bukowski)

What this means, is that it is not possible to use the HBA and LBA simultaneously on a given channel. While it is possible to mix and match between the different channel, each individual one is limited to a single input.

On these RCU boards are mostly switches and filters, but this then goes in to the sampler. This is a 12-bit analogue to digital converter, sampling at 160 or 200 megasamples per second (depending on the configuration). Once the signal has been digitised, down-converted and filtered the processing chain is common.