Successful trip to LPO in October 2007, ND and Erik attending.
Report on 2nd Commissioning of HaC
Arrival 29 October 2007, to see what progress had been achieved in reducing the detector noise compared to the first commissioning run, and possibly assist. The "bleeding" effects shown in the halpha.html page were traced to a fault on the controller causing the clock voltage to be a little low. But this was not the cause of the very high noise levels, which went away when the newly manufactured RSAA CCD mount was included.
When Nigel arrived he made contact with Andy Ridings who was working on the PN.S in the Instrument Test Area. Andy continued to support us through our entire stay on La Palma, including a public holiday, which was great.
Andy had all three dewars cooled, mounted and connected.
Despite the improvements mentioned above the noise levels were still above normal levels.
Andy's initial concern was for unearthed components, such as the plate provided by RSAA, which was (black) anodized so that there was no electrical connection to the rear of the instrument. We examined the assembly drawings of the HaC which I had brought along , with a view to disassembling and earthing the front end, but we chickened out when it seemed that this would involve removal of the iris shutter and or the focus mechanisms. Instead, Andy added an earth strap (as well as scratching away some of the anodization).
Subsequently, Andy showed us read noise levels (median of 100 pixels) of 3.0 e, but with a coherent, wavy modulation of 30 e.
In further discussions with ING personnel it was stated that (a) there is little point in trying to improve these levels until the instrument is actually at the telescope, as optimisation is a black art in which the position of a cable can be decisive, and (b) detector noise seems to be the curse of the ING and it has never been fully understood.
In Andy's words: "understanding and correcting the earthing made the noise more stable. This involved the shorting of 2 unexpected electrical breaks. One by removing anodising, the other probably caused by anodising, by the addition of an earth strap as we were unhappy about the risks involved with further disassembly. There is still some low level noise that comes and goes but there is little point in pursuing this off the telescope as it is likely to change when we put the instrument on the telescope. However we do now know that it is possible to get the noise down on all three detectors, there were periods when it was perfect. Before with the old mount and with the earthing not understood it wasn't clear that this was achievable." So we closed the file on this issue for the time being. We should make sure that Andy, and/or Simon, is available to help optimise the noise at the start of the December run.
30 October Erik de Wit arrived, to install and test the V2 controller. This was very convenient, with the PN.S being located in the test room. In this phase the controller was used manually only, i.e. pressing buttons on the controller.
Erik's first priority was to determine the "polarity" of the switches in the PN.S which tell the electronics whether the shutter (etc) is closed or not. Documentation of this had been lost. Also, we had determined that the motors used in the PN.S and HaC operated at 27V whereas Controller-1 (used to date) generated 6V. Again, there is no documentation of this discrepancy.
As a test we set the voltage at which Controller-2 drives the motors to the nominal 27V and noted a startling increase in the speed at which everything moved. Indeed we worried about damage occuring. The voltage was then lowered to 23V at which point the "clang" of the mask is less threatening but still loud enough to wake up Lodovico, if necessary.
Erik and I then turned to the implementaion of the control functions.
Erik had already implemented the basic functions (open/close mask and shutter) and now turned to the familar macros "focus", "calib" and "expose".
This required a lot of on-the-fly programming and in truth we ended up with little time to test every possible situation - The best I can say is that I was unable to crash the software with fast pushing of random combinations of buttons. I confirmed that the iris shutter of the HaC worked, as did the PN.S motors, via commands from the laptop with a serial connection.
We packed up, and reorganised the PN.S cupboard, and left on November 2.
Nigel Douglas 16/11/2007