The Low Frequency Instrument

The LFI (Figure 1 below) is designed to cover the 30 - 100 GHz band, with an array of 56 detectors split into 4 channels, centered at 30, 44, 70, and 100 GHz. Table 1 below summarizes the characteristics of the LFI receivers. The total number of detectors is a compromise between the high sensitivity desired, and the requirement to cool the enclosure passively to $\sim$60 K (the LFI amplifiers are the dominant heat input into the focal plane). The number of detectors in each band was chosen as a trade-off between the instrumental sensitivity (which varies with frequency), and the desire to obtain the highest final sensitivity at $\sim$100 GHz. Table 1 illustrates the instantaneous and final (i.e. after achieving two full coverages of the sky) sensitivity expected to be achieved with state-of-the-art receivers.

 

Figure 1: (a) A unit-level block diagram of the LFI. Each box in this diagram represents a single physical unit in the instrument. (FEU = Front End Unit (FH = feed horns, OMT = Ortho Mode T ransducers, FEM = Front End Modules, WG = Waveguides, BCCE =Sorption Cooler Cold End); BEU = Back End Unit (BEM = Back End Modules, DAE = Data Acquisition El ectronics); SCP = Sorption Cooler Piping; SCS = Sorption Cooler System (SCC = Sorption Cooler Compressors, SCE = Sorption Cooler Electronics); REBA = Radiometer Electronics Box Assembly (DPU = Data Processing Unit , SPU = Signal Processing Unit, PSU = Power Supply Unit).)

 

Figure 1: (b) An isometric view of the LFI Front End Unit, showing the horn/receiver arrangement. The brown-colored unit in the center of the ring formed by the horns is a schematic representation of the HFI Front End Unit.

 

Figure 2: A conceptual description of the ``continuous-comparison" receiver, two of which are fed by each (dual-polarization) horn. The Front End Unit includes all elements between the feed-horns and the phase switches. Thereafter waveguides (or coaxial cables at the lowest frequencies) will conduct the radiation to the Back-End Unit, which contains the second hybrid and the detection devices. For each detection chain, all of the elements in the Front End Unit can be packaged into a single MMIC (called the Front End Module) only a few cm³ in volume, with one input (radiation) connected to the feedhorn, and one output (the detected signal).

 

Table 1: Goal Characteristics and Sensitivity of the LFI

Center Frequency (GHz)	30	44	70	100
Wavelength (mm)	10.0	6.8	4.3	3.0
Detector Temperature (K)	20	20	20	20
Bandwidth (%)	0.2	0.2	0.2	0.2
Bandwidth (GHz)	6.0	8.8	13.0	20.0
Number of Detectors	4	6	12	34
Angular Res. (FWHM, arcmin)	33	23	14	10
Tsys (K) (Diff. Rec.)	9	15	22	31
Array $\Delta$# (mK$\sqrt{s}$)	0.12	0.16	0.19	0.22
Polarized detector	Yes	Yes	Yes	Yes
Nominal mission* (1$^\star$ sensitivities, per pixel$^\ast$)
Average integ. time per det.(sec.)	213	103	38	20
$\sqrt{2}$ per pixel$^\dagger$ ($\mu$K)	4	7	10	12
$\Delta T$ Sensitivity$^\dagger$ ($\times$ 10-6)	1.6	2.4	3.6	4.3
Flux Sensit. per pixel (mJy)	13	19	25	27
* 12 months of observations, or two full sky coverages
$^\ast$ A pixel is a square whose side is the FWHM extent of the beam
# Antenna temperature
$^\dagger$ Thermodynamic temperature.

Radiation will be coupled from the telescope to the detectors via conical corrugated feedhorns exploiting the two orthogonal polarizations at a given frequency. Thus, each horn feeds two receivers. The design of the feedhorns, though straightforward, must be optimized to provide low sidelobe level, low loss, and good stability. The feedhorns will be oriented so as to minimize beam pattern aberrations: a good first approximation places the feed apertures on a spherical surface centered on the subreflector apex.

Each receiver consists of a pair of amplification/detection chains connected in parallel via so-called hybrid rings (Figure 2), and constitutes a ``continuous-comparison" device (Bersanelli et al.996). In this scheme, the difference between the inputs to each of the chains (the signal from the telescope and that from a reference blackbody respectively) is continuously being observed. To achieve this, it is necessary to modulate periodically the sign of the inputs via solid-state phase-shifters. This design was chosen over a much simpler total-power scheme (consisting of one of the two parallel chains) because the latter exhibits inadequate gain stability at time scales larger than a few seconds. The differencing receiver improves the stability considerably if the two input signals are almost equal (at a cost of a factor of $\Delta T/T$ in sensitivity). The optimal case is achieved when the reference is a blackbody at a temperature of $\sim$4 K while the signal is at $\sim$3-5 K, Any rsidual differences between the sky and reference signals are effectively equalized by modulating the gain (g). The blackbody reference itself must remain at a very stable temperature ( $\sqrt{2}$ 10 $\mu$K).

Each amplification stage will be provided by High Electron Mobility Transistors (HEMTs). This technology offers at present the best compromise between sensitivity and ease of implementation in the frequency range of the LFI. Total power receivers based on Monolithic Microwave Integrated Circuit (MMIC) technology, in which the whole receiver is packed into a device a few centimeters across, are state-of-the-art but already available commercially with characteristics close to those required by Planck. The next step, consisting of building a complete continuous-comparison receiver into a single MMIC is currently in progress, and is relatively straightforward.

As noted before, the dominant heat input into the focal plane is due to power dissipation in the MMICs; since this fact drives the thermal design of the payload module significantly, it is desirable to reduce it as far as possible. During the course of the development of the instrument significant improvements have already been made in this direction (partly by switching from GaAs to InP technology, but mostly by splitting the low noise amplifiers into Front End and Back End Units), converting the initial assumptions into extremely conservative ones.

The FEU is cooled to 20 K by a hydrogen sorption cooler, which works via the Joule-Thomson principle, i.e. the expansion of hydrogen through an orifice. The cooler also provides 18 K precooling to the 4 K cooler of the HFI. The gas is driven by passive compressors, containing a metal hydride material which are cyclically heated (thereby releasing gas at high pressure) and allowed to cool (thereby resorbing gas into the material). Stainless steel piping connects the compressors to the Front End Unit, where liquid reservoirs act as heat exchangers.

 

Figure 3: Schematic of the 20 K sorption cooler which serves both LFI and HFI.

Beyond the signal detection level, the LFI consists of electronics, data processing and interfacing software, all of which are packed into a single room-temperature box which is placed outside the optical enclosure.