First results of the IRAM 30-m telescope improved thermal control system

The structure of the 30-m telescope is made of steel and the design aimed from the beginning at the minimisation of its thermal deformations. For this purpose the antenna steel structure is covered with thermal insulation. In addition, the Reflector Backup Structure (BUS) is equipped with an active cooling/heating thermal control system (TCS) maintaining the temperature homogeneity of the BUS within 1 K.

  

Figure 5: The 30-m telescope and the position of the new Temperature Control System

Figure 5 shows a cross-section through the 30-m telescope. The strict conical design of the BUS is terminated in a flat circular steel surface, called the membrane. Below the membrane are the Yoke and both Counter Weight arms, one at each lateral side of the antenna.

The initial TCS controls only the temperature of the BUS. Five ventilators apply a tangential air circulation with a heating/cooling capacity of 30/22 kW. A temperature sensor installed in the upper part of the Yoke, below the membrane, measures the reference temperature to which the temperature of the BUS and the feed legs is slaved. The initial TCS does not apply temperature control to the Yoke and the Counter Weights. During the past years, 174 temperature sensors were installed in the BUS, Yoke and Counter Weights and also in the TCS fluids of the feed legs in order to monitor the temperature of the whole antenna. In a further step, a Finite Element Model (FEM) was developed to interpret the measured antenna temperatures with respect to the antenna performance. The predictions of the FEM have been confirmed from dedicated focus and holography measurements. The FEM permits a decomposition of the calculated temperature-induced reflector surface deformations into large-scale contributions represented in Zernike polynomials; from these we found that the dominant components of reflector surface deformations are astigmatism (of amplitude $\alpha _2$), and 3rd order defocus (of amplitude $\alpha_{4}$). It was recommended to improve the thermal homogeneity of the upper part of the Yoke, just below the membrane.

An independent forced ventilation system, consisting of 4 ventilators moving the air in a circular way, was installed in the upper part of the Yoke in autumn 1999. The temperature of this part of the telescope structure was improved. Nevertheless, the reflector astigmatism still showed a permanent high value though with less variation than previously observed, which was due to temperature inhomogeneities of the yoke.

  

Figure 6: Initial Temperature Control System during five days of good weather: Temperatures of the BUS and counterweights.

  

Figure: Initial Temperature Control System during five days of good weather: Temperature induced rms thermal deformations and amplitude $\alpha _2$ of the astigmatism.

A sample of the performance of the initial TCS, with the ventilators at the upper Yoke working, for a period of 5 days with good weather, clear sky, low wind and the antenna observing all the time is shown in Figs. 6 and 7. After analysing the information from the temperature sensors and the FEM we concluded the following:

a)

the initial TCS works well and the controlled temperature of the BUS is homogeneous within 1 K (Figure 6). This behaviour fulfils the design specifications.

b)

the Counter Weights are always approximately 3 K cooler than the reference temperature (Figure 6). The reason for this difference is the extra heating applied to the BUS and the upper Yoke, which comes mainly from the ventilators inside the BUS, the cooling compressor of the TCS, and the conducted heat from the receiver room inside the antenna structure. The Counter Weights are less strongly coupled to the internal heating sources and more to the open air. In general, the reference for regulation of the BUS is 4 to 7 K above the average temperature.

c)

Figure 7 shows the prediction from the FEM of the total thermal rms deformations of the reflector surface $\sigma_{\rm T}$ and the amplitude $\alpha _2$ of the temperature-induced astigmatism, derived from the decomposition into Zernike polynomials. The displayed amplitudes $\alpha _2$ are between 40-120 $\mu$m.

When using the FEM to analyse the reflector astigmatism versus the antenna temperature distribution for a long period of time we noticed that occasionally the amplitude of the astigmatism is small. This situation occurs when the air temperature drops rapidly due to some particular meteorological condition. The BUS and Yoke temperatures decrease faster than the temperature of the Counter Weights due to their lower inertia, and at a certain moment the temperatures of these three components are equal. At this moment the astigmatism disappears.

According to the FEM a certain temperature gradient in the Counter Weights has a smaller effect than a similar gradient in the Yoke. On the other hand, the temperature difference between the Counter Weights and the reference temperature is much higher than the temperature difference between the Yoke and the reference temperature. We decided to modify the TCS by applying heat to the Counter Weights in order to reproduce the noticed condition of small astigmatism, i.e. of equal temperatures in the BUS, Yoke and Counter Weights.

In September of this year (during the yearly maintenance) we have made the following changes in the antenna TCS (Figure 5):

1)

Two ventilators, one at each lateral side of the Yoke, were installed between the elevation axis and the Counter Weights. Each ventilator moves 4300m³/h of air around the elevation axis at the corresponding lateral side of the Yoke, equalising the temperature between the Counter Weights and the upper part of the Yoke. Both ventilators are permanently ON.

2)

Four sets of heaters, with 3 kW each, were installed in the Counter Weights, two sets per antenna lateral side. Each heater is installed in a closed compartment to optimise the heat transfer to the Counter Weights and from there, by conduction rather than convection, to the rest of the Yoke. The number of activated heaters depends on how cold the Counter Weights are with respect to the reference temperature.

The new TCS operates since 17 October 2002. We have analysed the data of 5 days with clear sky, low wind, and with the antenna observing all the time (last VLBI session) to understand its behaviour. Results are shown in Figures 8 and 9 which illustrate the improvement reached with the new TCS. Besides the 5 days analysed so far, the new TCS is working well since its beginning of operation.

  

Figure 8: New Temperature Control System during five days of good weather: Temperatures of the BUS and Counterweights.

  

Figure 9: New Temperature Control System during five days of good weather: same as Fig. 7

  

Figure 10: New Temperature Control System during five days of good weather: up to day 29 with the operation of the new TCS, from day 29 on operation of the initial TCS.

Table 2 compares the behaviour of the initial TCS and the new TCS, including the reflector performance predicted from the FEM. The values give the average and the standard deviation (rms-value).

 

Table 2: Comparison of the initial and new temperature control system

Parameter		Initial TCS	New TCS
Reference - Counter Weight Temperature	[K]	3.34 (0.57)	0.26 (0.14)
Total Reflector Surface Deformation (rms-value $\sigma_{\rm T}$)	[$\mu$m]	41.1 (6.9)	16.5 (4.2)
Component of Random Thermal Deformations (rms-value $\sigma_{\rm rd}$)	[$\mu$m]	16.0 (4.4)	12.5 (3.7)
Amplitude Astigmatism $\alpha _2$	[$\mu$m]	89.6 (17.7)	14.5 (8.5)
Amplitude 3rd Order Defocus $\alpha_{4}$	[$\mu$m]	18.9 (13.9)	13.3 (10.6)

Note that $\alpha _2$ and $\alpha_{4}$ is the amplitude of astigmatism and 3rd order defocus, respectively. The corresponding quasi rms-values of these systematic surface deformations are $\sigma_{2} \approx 1/3 \times \alpha_{2}$ and $\sigma_{4}\approx 1/3 \times \alpha_{4}$.

Figure 10 is similar to Figure 9 but including two more days. Between day 29.5 - 30.5 the new TCS was intentionally switched back to the initial TCS to facilitate the comparison of the performances.

With the operation of the new TCS the thermal behaviour of the antenna has improved. Several more months of data must confirm the preliminary results. However there is no reason to expect a degradation of the new TCS after working well under different meteorological conditions during 4 weeks. The goal to keep the antenna from the Reflector Backup Structure to the Counter Weight on the same temperature has been reached and so far the antenna shows a more stable thermal behaviour. Under this improved condition a new efficiency measurement as well as holography should be considered in the near future.

Juan Peñalver, Albert Greve and Michael Bremer