Up until this point, we’ve focused a lot on the engineering and electronics of the telescope. That part of the development process is thankfully thriving. However, we’ve not taken a lot of time to look at the final component of the pipeline; the computer.
The plan from the start has been to computerise the telescope. This was the motivation behind our choice of the Spectracyber II spectrograph. Not only was this a complete solution integrated into a single neat box, but it also came with the software required to communicate with the box from the computer. We also discussed the possibility of a solution involving the relatively new technology of Software-Defined Radio (SDR), wherein a computer simulates the radio receiving hardware to demodulate the signal. Although this showed promise, we opted for an in-hardware solution. Due to the spectrograph being an older, well tested design, it connects to the computer via a serial COM ports, as opposed to a more modern standard like USB. This drastically limited our choice of computer, as most don’t ship with this connection any more. There was also a large debate as to whether we should use a laptop computer for this purpose. We eventually came to the conclusion that the form factor was not worth the extra price, as it would remain anchored to the telescope most of the time anyway. Thankfully, the University of Exeter’s IT department was able to find us a system that met our requirements.
The computer in question was an HP EliteDesk 705 G1, running Windows 7 Professional 64-bit. The software installed fine but when it came to connecting to the receiver we started getting errors. After some troubleshooting, and emails back and forth to the IT department, it seemed like the issue was with driver for the COM port. So I ventured into the department to figure out the issues plaguing the system once and for all.
The first cause was misconfiguration of the hardware. There is a low noise amplifier (LNA) in the signal path between the feed horn and the receiver. This is powered using a small 12V spur (red wire), with accompanying ground (black wire), run from the receiver end of the mains power cable. In our initial setup these were wired the wrong way around causing the system to be improperly grounded and for the light on the front of the receiver to pulsate. Due to this the receiver wasn’t properly able to power up and communicate with the computer. Changing the wires around resolves this problem.
The software was also causing issues. When attempting to initiate communication with the receiver we got a variety of different errors, all pointing to the fact that the program was unable to create the file required to access the COM port. This issue was resolved by going into the properties of the SpectraCyber executable and changing the compatibility mode to Windows 98 / Windows Me. This runs the program with a privileged user level and allows access to the COM port. It’s worth remembering that this program was originally developed back in 1998 on Windows 98. Since then operating systems have become a lot more secure and don’t allow such easy access to the serial connections. Putting the program into compatibility mode is allowing it that access and letting it communicate with the receiver.
With these issues resolved and the system in a bare bones configuration, with the receiver wired directly to the feed horn via the LNA, we got the first flickers of life from XRT-C. Even in this configuration, and with the feed horn pointed toward a nearby window by using a book as a crude stand, we notice possible indications of the 21cm hydrogen line near 0 kHZ Doppler shift.
When we get the full telescope built later in the year we’ll have to set the system up properly and carefully calibrate both the receiver and gearing system that moves the dish. The next job for the science team is to read up in literature and figure out the best way to go about doing this.
by Sam Morrell – Science, Outreach and Press & Publicity Teams