Group for Advanced Receiver Development
Onsala Space Observatory
Department of Radio and Space Science

Chalmers University of Technology


Swedish
Single-Pixel Heterodyne Facility Instrument
for
APEX Telescope

Documentation Package*

 

 

 

 

 

 

 

 

* This document is delivered as a hard copy upon request; the documents all available electronically on the supplementary CD with text files in pdf format and design/drawing files in format compatible to the AutoDesk Inventor Pro 11


Content

1.                  About this Document…………………………………………………   3

2.                  Introduction to Swedish Single-Pixel Heterodyne Facility Instrument
(SHFI) for APEX telescope.  General description of the SHFI……….  4

3.                  Description of the SHFI Optics ……………………………………..     10

4.                  SHFI Cryogenics & Dewar/mechanical interface …………………..      13

5.                  SHFI Band 1 211-275 GHz mixer assembly ………………………..     20

6.                  SHFI Band 2 275-370 GHz mixer assembly ………………………..     27

7.                  SHFI Band 3 385-500 GHz mixer assembly ………………………..     34

8.                  SHFI Band T2 1250-1390 GHz mixer assembly ……………………     38

9.                  SHFI IF Diagram and Cold IF amplifiers ……………………………   44

10.              SHFI LO Diagram ……………………………………………………   47

11.              SHFI Mixer DC Bias …………………………………………………   49

12.              SHFI Wiring Diagram ………………………………………………..   66

13.              SHFI Control System.

A.    Hardware and Cabling …………………………………………….. 88

B.     Software and User Interface ………………………………………. 88

14.              SHFI Maintenance Procedures

A.    SRDK-3ST Maintenance …………………………………………   89

B.     Warming up and Venting Procedure ……………………………..   89

C.     Pumping down and Cooling Procedure …………………………..   91

D.    SHFI Dewar Opening Procedure ..………………………………..  92


1. About this Document

The purpose of this document is to provide reference information on the design, components and operation of the Swedish Single-Pixel Heterodyne Facility Instrument (SHFI) for APEX telescope.  The information is presented according to the content above with further reference into respective files for more details on the subject.  Externally purchased components are represented mostly by an information on the producer companies, web-links, manuals when available and invoices for reference.  Every chapter has reference names of GARD members who carry the main responsibility and could be contacted to provide further information.  For every chapter of this document we have created a folder with reference files and other relevant information.

The document editing is done by V. Belitsky (Victor.Belitsky(at)chalmers.se) with contribution from persons in the contact lists for every particular chapter.

It is assumed that maintenance and service operations described in this document are carried out by highly qualified staff with appropriate level of knowledge in cryogenics, vacuum technology, microwaves, electronics and receiver technology.


2.  Introduction to Swedish Single-Pixel Heterodyne Facility Instrument (SHFI) for APEX telescope.  General description of the SHFI

The Swedish single-pixel heterodyne facility receiver (SHFI) is part of OSO contribution to Atacama Pathfinder Experiment, APEX, and to be placed in the telescope Nasmyth cabin A. The receiver is built based on integrated design concept where no plugged modules or cartridges are used.  Instead, the receiver provides tight integration of all components, which saves space and reduces the mass.  For example, 6-channel SHFI occupies about 20% of ALMA 10 ‑ channel receiver volume and about the same 20% in mass.  Integrated design allows to share many components, e.g., the second stage IF amplifier is common for Band 1, Band 2 and Band 3, those bands using SIS junction based mixers, while Band T2 with HEB mixer has separate IF chain due to the difference in the SIS and HEB mixer IF bands.  The receivers are coupled to the antenna via relay optics providing possibility to operate either one of the two different PI-type instruments or SHFI; optics allow to cover 211 – 1500 GHz frequency range.  Out of 6 positions in the SHFI receiver GARD OSO is responsible to populate four: APEX Band 1, 211 ‑ 275 GHz, using sideband separation technology (2SB), Band 2, 275 – 370 GHz, 2SB, Band 3, 385 – 500 GHz, 2SB,  and Band T2, 1250 – 1390 GHz, HEB waveguide balanced mixer.  Due to SHFI placement in Nasmyth cabin A that is about 6.5 m away from the antenna focal plane, the optical solution implemented allows observation with one frequency channel of SHFI at a given time.

In order to ease installation and maintenance of the SHFI, its design provides well defined interface between the warm optics and the dewar with cryogenic parts of the receiver, see Figure 1.  The upper part of the installed receiver comprises the room temperature optics, LOs for all 4 channels and interfaces to the dewar with over-defined mechanical interface with 4 guiding pins.  Over-definition is justified due to longer base lines, around 1 m, and thus degraded manufacturing accuracy. The warm optics and LOs are installed on a separate platform and once installed and aligned are kept while to lower part with cryogenic electronics could be removed for maintenance and easily returned back without a need for optical alignment.  The split interface is marked on the Figure 1 by the blue balloon.  The entire structure is integrated with earlier provided mechanical support frame for Cabin A.

In order to reduce the amount of sensitive operations during the maintenance, e.g., plugging the SIS junction, HEB and amplifier DC bias cables, and the box with DC biasing circuitry is mounted directly on the SHFI support frame.


Figure 1.  SHFI Integration into the support structure installed in the APEX Cabin A, view from the Cabin A door-side.  Figures 1a shows more details from different view angle.


Figure 2.  SHFI Integration into the support structure installed in the APEX Cabin A, view from the Cabin A Nasmyth tube.

Vibration produced by Sumitomo SRDK-3ST is substantially lower as compared to the cryo-cooler used in APEX2a receiver.  Nevertheless, in order to provide efficient isolation of the vibrations from the RF part of the receiver the cryo-cooler in SHFI is mechanically isolated from the receiver by suspending it on vibration dampers ND50 A vibration isolator,  NEWPORT Inc. while the vacuum sealing is done via flexible bellow of 160 mm in diameter.  Figure 3 displaces the details of the cryo-cooler suspension.  Additional support structures preclude the bellow collapse under atmospheric pressure from outside and interconnected via rubber sheets.  The latter support structures could be completely removed but in order to preclude the bellow collapse the cryo-cooler should be loaded by additional about 14 kg of counterweight (at APEX altitude).  All cryo-connections with the Sumitomo SRDK-3ST inside the dewar at 70k, 12 K and 4K stages were made using flexible thermal links made of oxygen-free copper twisted wire in order to reduce possible vibration leaks.

Figure 3.  The entire cryo-cooler is placed on a vibration isolated platform.   For that we mounted the cryo-cooler with a vacuum sealing via a bellow with diameter of 160 mm and 4 vibration damping elements providing vibration isolation of the entire platform.

Figure 4.  All IF and DC interfaces of the receiver are located at the bottom plate of the receiver dewar while all RF interfaces are placed on the top plate of the receiver.

The mixer design for each channel had in focus the specifications. Namely, the specifications call for single-sideband mixers and completely tuneless design of the mixers and LO sources. The only technology available to fulfill all these requirements is to employ mixers in quadrature layout with 90-degree signal phase shift between the mixers and further 90-degree hybrid at IF to recombine components for upper and low side bands. The illustration below presents the concept of the sideband separation mixer scheme, which was chosen for APEX Band 1, Band 2 and Band 3 channels using all SIS junction mixers.

Figure 5.  The sideband rejection mixer explained.  Vectors at points C, D, E, F show USB and LSB components with their respective phases illustration sideband separation technology.

The most critical issue for APEX Band T2 mixer was the availability of a LO source and how much attainable LO power. The LO source from VDI has an available power (according to documentation from VDI) across the entire band that does not exceed 10 mW and falls below 2 mW at the higher end of the tuning band. With such low attainable power there can be no solution with a beam-splitter as it will be harmful for the mixer noise performance and introduce too large losses in the signal path. Any technical solutions employing a diplexer based on interferometer requires mechanical tuning for every observational frequency and is ruled out by the specifications. For APEX T2 we have chosen a balanced waveguide mixers scheme as the only providing virtually no loss for LO signal and thus allowing us to tackle the problem of extremely small available LO power, and additionally with advantage of rejecting the amplitude/additive component of the noise leaking via LO mixer port. Figure 6 illustrates the layout of the balanced HEB mixer implemented in APEX Band T2 channel.

Figure 6.  Illustration of the balanced HEB mixer layout implemented in APEX Band T2 channel.

3.  Description of the SHFI Optics

Two main considerations governed the design of the APEX Nasmyth Cabin A optics. First, in order to cover the required frequency band, the optical solution for the “common” path in Cassegrain Cabin and inside Nasmyth Cabin A (NCA) should be frequency independent and ease the design of the optical components while could still fit the Nasmyth layout with 150 mm limited aperture (by the elevation encoder in the Nasmyth tube of the NCA).  Second, taking into account extremely wideband receiver channels of the facility receiver having the RF band up to 30% of the central frequency, we would like to provide illumination of the secondary at the level of -12 dB edge taper frequency independent over every partial RF band in order to avoid optical loss at the receiver bands’ edges.  This is not the case for APEX T2 as its operational band is only 10% of the central frequency and optical loss are negligible at the band edges. The optical components of the relay optics in the Cassegrain Cabin should be placed to circumvent interference with multi-pixel bolometric instruments and their respective optics residing in the Cassegrain Cabin.

In order to provide relaying of the sky signal to PI instruments and the multi-channel facility receiver we used optical switch, a flat mirror driven by precision rotation stage placed at 45o to the telescope elevation axis (also the axis of the incoming beam). Similar solution is used for switching between the channels of the multi-channel facility receiver: a long focal distance mirror mounted on the same type of rotation stage is used for coupling any of the six receiver channels to the antenna. Figure 5 presents the schematics of the relay optics for NCA and Figure 7 shows the optical components placement inside the telescope. During the design of the optics we use Gaussian beam technique.

Figure 7.  Optical layout for the APEX Nasmyth Cabin A receivers.  Every channel of the facility heterodyne receiver has 8 mirrors in the signal coupling path (Mirror F10 was abandoned).  In order to optimize coupling to the antenna, the optics for every receiver channel has been designed to provide frequency independent illumination of the secondary over the operating frequency band with edge tape -12 dB.

The optics optimization for each channel, in general, followed the concept presented in [1] and further developed in [2, 3]. In contrast to relatively simple optical circuitry discussed in [2, 3], we have to achieve frequency independent behavior of the 9 mirror chain relaying the antenna beam to the mixer. In order to simplify this task, common path optical components, M3, F4, F5, M6 and Fs7, were chosen based on requirements of frequency-independent waist positions with M3 and M6 forming a Gaussian telescope [4] separated by 2500 mm with focal length of 700 and 1800 mm respectively. Mirrors F4, F5 are flat and used to fold the beam in the Cassegrain cabin to avoid interference with optical components and supporting structure of the multi-pixel receivers. Flat mirror Fs7 is the optical switch between PIs and the facility receiver placed in the NCA thus providing the same optical interface to all instruments placed in the cabin. The remaining mirrors in the chain, M8s, F9, M11, should be designed in order to achieve the required edge taper on the secondary (-12 dB) with as weak as possible frequency dependence. However, the primary function of the M8s as channel switch is to provide receiver band selection of the multi-channel facility receiver. M8s serves all the receivers and therefore the mirror should guide the Gaussian beams of extremely wide frequency band, 210 – 1500 GHz.  This pushes for the mirror with relatively long focal distance, it was chosen for M8s to be 530 mm.  The last consideration is connected to the convenience of alignment and mounting of the optical components: for this, we decided to have mirror F9 to be flat and placed at the same height from the cryostat top plate for all receiver channels.  Finally, the choice of the M11 and the corrugated horn design were optimized in the manner described in [3] with the difference that all 10 optical components including the secondary and the corrugated horn, with distances between them, their respective focal distances were put into an analytical equation describing the system in terms of geometrical optics. To achieve frequency-independent illumination of the secondary we should re-image the secondary onto the horn aperture with the required beam size and, at the same time, the focal point is re-imaged into the horn apex [3].  After obtaining all optical component parameters and positions from the geometrical optics, we run Gaussian beam analysis to ensure that all solutions would provide desirable performance.  Aberrations on the mirrors are kept small using small opening angle between the incoming and the reflected beams.  More details on optics for every particular band could be found n the corresponding chapter for each of APEX Receiver Bands.

References

1.      T.-S. Chu, “An imaging beam waveguide feed,” IEEE Trans Antennas and Propagations, vol. AP-31, no. 4, pp. 614–619, Jul. 1983.

2.      J. W. Lamb, “Optical study for ALMA receivers,” ALMA Memo Series No. 359, Mar. 2001

3.      A. Baryshev, W. Wild, “ALMA Band 9 Optical Layout,” ALMA Memo Series No. 394, Sept. 2001

4.      S. Heyminck, “APEX Cabin A Common Optics”, APEX Technical Note, Dec. 2003.


4.  SHFI Cryogenics & Dewar/mechanical interface

SHFI uses cryo-cooler with 3 stages to provide operational environment for SIS and HEB mixers. The cryo-cooler of model SRDK-3ST is produced by Sumitomo Heavy Industries Inc. and purchased via Janis Inc., USA.  The compressor must be connected to a close cycle water cooling system for normal operation. Without additional load the cryo-cooler third stage temperature measured in the SHFI dewar is around 3.6 K with normal sea-level atmospheric pressure and room temperature around 24 C. When the room temperature is colder, around +15 C as in the Cabin A, we expect the unloaded temperature would be around 3.4-3,5 K. The SRDK-3ST has a liquid He pot integrated into the 3d stage for additional temperature stabilization. The factory manual, maintenance and debugging documentation is following in Appendixes.

In order to minimize influx of thermal radiation loading the 4 K stage of the cryo-cooler, SHFI employs a shutter which is placed between the dewar vacuum shield and 70 K thermal shield. The shutter has two openings for LO and RF signal windows and its position is synchronized with rotation of the channel switching mirror M6s (see Figure 5) that allows automatically move the shutter into desirable position. The cooler in the receiver is placed vertically (see Fig. 3) which is one of positions allowed for regular operations by the manufacturer, SHI. SRDK-3ST is known to have very low vibration level, however to further isolate the vibrations we designed the dewar in a such a way that the cryo-cooler is suspended in the system via vibration-damping elements, see Chapter 2, above.  Vibration isolation requires use of a special type of thermal link connecting the cooler stages and the receiver components; the thermal links should be flexible enough to preclude vibration leak towards receiver components they are connected to.  Figure 8 presents details of cryo-mechanical layout of the dewar.

Mixer assembly interface plate provides mechanical interface (usual combination of precision steering pins and M3 stainless steel screws) for fixing mixer assemblies in the dewar.  Figure 7 shows how this looks in reality with all mixer assemblies installed, while Figure 9 shows CAD generated view when all mixer assemblies installed into their respective slots.

 

Figure 8.  Computer simulated CAD drawing showing details of the SHFI dewar inner structure. The 70K and 12 K base-plates support is not shown; it is done by using the same fiberglass pipes as for 4 K mixer assembly interface plate, however as 70 and 12 K cooling capacity is much greater than that for 4 K stage (0.98 W (at) 4.18 K), the fiberglass support tubes are much shorter and more stiff.  For 4K the support structure is designed to minimize the amount of heat coming from the 12 K stage and required long fiberglass pipes to be used; to enforce the structure stiffness, 3 diagonal bindings have been added.

Every integrated band has individual signal and LO input optics and corresponding pare of windows ob the dewar top plate. All windows were individually produced by QMC Inc. made from crystal quartz with special anti-reflection coating providing minimum RF loss and very little reflection over each frequency band.

 

Figure 9.  Computer simulated CAD drawing showing details of the SHFI dewar inner structure. The mixer assemblies are shown integrated into their respective slots.

Due to different area of window aperture of each band and slightly different mixer assembly configuration, the total amount of heat during operation of each channel is different.  This results in different equilibrium physical / ambient temperature for each frequency band mixer under operation. In laboratory conditions with constant environment and given atmospheric pressure around 1000 mB, the receiver cold plate temperature stability is not a problem. However, in order to address the fact that on the mountain the physical temperature could be substantially lower, as for example happened with APEX2a, we built in temperature stabilization system that uses de-fluxing heater resistors to stabilize each mixer operating temperature. The system takes use of a computer-based control system with software feedback loop: 12 bit resolution over the temperature measurements and the heater current. Figure 11 shows measured temperature on Mixer 2 of the Mixer Assembly B2 with the temperature stabilization engaged, as captured from a LabView program.


Figure 10. Details of the SHFI dewar inner structure with all 4 mixer assemblies installed onto the interface plate; additional flexible thermal links at 4 K stage and at 12 stage are clearly visible.

Figure 11.  Temperature stabilization system: current temperature vs. time. Plots show temperature variations in [K].  With the temperature stabilization system engaged, a maximum of ±2 mK with a sampling period of 250 ms.

Figure 12.  Temperature stabilization over a 30 s interval.


Figure 13. Standard cooling process; the data were taken in the laboratory conditions (+23-24o C) in Gothenburg.


All mechanical drawings including input RF signal Gaussian beam tracing trough the Cassegrain Cabin and LO injection beams have been carried out using AutoDesk Inventor 11 Pro and the entire package is included in the supplementary CD disk.  Warning: total size of the drawing with sub-assemblies and components is close to 150 MB; in order to handle this drawing file using Inventor Pro 11 the computer must be at least Intel C2D 2 GHz, minimum 2 GB RAM and high-end graphic card produced not later than 2006.  Autodesk Inventor 11 Viewer free software should provide possibilities to review the files and was included on the supplementary DVD disk (English and Spanish versions).  Disclaimer: GARD OSO has no responsibility and does not provide any support for AutoDesk products.

Contacts:

Mechanical and cryogenic design: Igor Lapkin lapkin(at)chalmers.se, Victor Belitsky Victor.Belitsky(at)chalmers.se;
Optics: Igor Lapkin lapkin(at)chalmers.se,
Temperature stabilization system: Magnus Svensson magnus.svensson(at)chalmers.se, Doug Henke, doug.henke(at)chalmers.se, Erik Sundin erik.sundin(at)chalmers.se .


5.  SHFI Band 1 211-275 GHz mixer assembly

The SHFI APEX Band 1 mixer is based on SIS mixers and uses a sideband rejection (2SB) scheme.  In contrast with many 2SB SIS mixers built using modular design, SHFI B1 mixer makes use of an integrated design whereby both SIS mixers are placed on a single substrate, which also comprises the LO division scheme using a double-probe structure and the LO injection directional couplers.  All this provides optimum layout with minimum loss between the mixer and the corrugated horn.  Mixer details could be found in published papers collected on the supplied DVD under SHFI_APEX_Band_1 folder.  Figure 14 shows details of the SFHI B1 Mixer assembly consisting of the mixer, the IF amplifiers, 3 dB IF hybrid, IF isolators, IF amplifiers (2-stage GaAs HEMT, designed by GARD), LO and RF optical components all integrated on a single bracket.

Figure 14.  CAD generated drawing of the SHFI Band 1 Mixer assembly.

When installed in the dewar, SHFI Band 1 mixer assembly is also completed with wiring, cabling, thermal straps, etc.  Figure 15 a, b, c demonstrate SHFI B1 mixer assembly installed in the dewar, please note the DC interface contact is disconnected.  For reference and part identification please refer Figure 14.

Figure 15a.  Picture of the SHFI Band 1 Mixer assembly inside SFHI dewar.  Note changed IF interface, no 90o bend, as compared to the CAD picture, Figure 14.

Figure 15b, c.  Pictures of the SHFI Band 1 Mixer assembly.

Figures 16. The measured SSB system noise temperature of the SHFI B1 mixer for both sidebands and two LO sources – SHIFI multiplication chain (circle/cross) and Gunn oscillator plus tripler (rectangle/star). The temperatures are measured over the full 4-8 GHz IF band and include the contribution from the image sideband, optical losses and IF noise. The measured sideband rejection ratios for 18 LO frequencies (36 sidebands) between 219 and 270 GHz taken with 0.5 GHz resolution and 4-8GHz per sideband.

Figure 16 illustrates the typical performance of band 1. The measured and averaged over the full IF band noise temperature for two different LO sources shows that the SHIFI multiplication chain gives higher noise temperature at high LO frequencies, partly, due to a problem with spectral impurity. This problem can be overcome by using band 2 mixer which covers RF frequencies between 265-275 GHz, with less noise temperature. The SSB rejection ratio is typically above 10 dB.  In order to characterize/verify the quasi-optical design and to align the receiver beam of the APEX-facility receiver, vector-field measurements are performed by scanning a transverse plane with respect to the axis of propagation and measuring the complex field distribution. The characterization includes determining of the propagation direction, the size and position of the beam waist, as well as the shape of the beam.  The measured data is fitted to the fundamental Gaussian beam mode by optimizing the power coupling coefficient. In order to achieve optimum coupling between two beams, the simulated and the measured beam, not only the axis of propagation needs to be the same, but also the size of the beam radius and the radius of curvature. Figure 17 shows a 3D plot of the measured beam at a frequency of 221GHz at a distance about 400 mm from the waist position. Figure 18 and Figure 19 show power- and phase- distribution at a cross-section of the measured beam pattern together with the values for the targeted Gaussian beam. It can be seen that Gaussian beam approximation agrees very well with the measured data and a power coupling coefficient of about 99% was obtained when a fitting down to an on-axis relative power level of 20 dB.

Figure 17.  APEX-Band1 beam pattern measured at 221 GHz at scan distance approximately 400 mm from the waist position.

Figure 18.  APEX-Band1 beam pattern measured at 221 GHz: the beam cross-section with measured values (-o) and fitted Gaussian beam approximation (-∆).

Figure 19.  APEX-Band1 beam pattern measured at 221 GHz: the beam phase cross-section with measured values (-o) and fitted Gaussian beam approximation (-∆).

Figure 20a. Allan variance (vertical axis) vs. time (horizontal axis, log scale, sec) measurement result for SHFI B1 taken with power meter and band-pass filter of 500 MHz centered at 6 GHz, at the LO frequency 235 GHz. The sampling time and IF bandwidth were chosen such that the 1 Hz gain variation is best visible. The mixer settings are according to the tuning table, the sampling time is 50 ms.

Figure 20a shows a plot of total power stability measurement without (left) and with (right) thermal stabilization engaged. The result shows clearly the improvement in the stability due to reducing the 1 Hz thermal variation caused by the cooling machine.

Figure 20b Spectroscopic Allan variance at IF frequencies 4, 4.5, 6 and 8 GHz taken at LO 235 GHz, sampling time 150 ms and IF bandwidth 3 MHz.

Figure 20b shows the corresponding spectroscopic stability measured with spectrum analyzer for four different IF frequencies. Despite that the plots do not show the minimum in the measured variance (due to the time required for the measurement) they demonstrate that the stability is consistent along the IF band and does not vary with the IF. The temperature stabilization of the mixer was not engaged during this measurement.

 

Contacts:

Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se
Mixer design, performance measurements: Vessen Vassilev vessen.vassilev(at)chalmers.se, Doug Henke doug.henke(at)chalmers.se
Beam measurements: Olle Nystrφm olle.nystrom(at)chalmers.se


6.  SHFI Band 2 275-370 GHz mixer assembly

SHFI APEX Band 2 mixer is based on SIS mixers and used sideband rejection (2SB) scheme.  Following success of the APEX2a receiver we use the same type of mixer design, developed by Dr. C. Risacher, and implemented it in 2SB configuration building SHFI B2 mixer by using more traditional modular design.  In this design the middle piece made of two symmetric parts (split-block technique) and comprises the RF 90o waveguide hybrid and LO division scheme with LO injection directional waveguide couplers using novel on-substrate coupling elements.  Mixer details could be found in published papers collected on the supplied DVD under SHFI_APEX_Band_2 folder.

Figure 21.  CAD generated drawing of the SHFI Band 2 Mixer assembly.

Figure 21 shows details of the SFHI B2 Mixer assembly consisting of the mixers, the middle piece, the IF amplifiers, 3 dB IF hybrid, IF isolators, IF amplifiers (2- stage GaAs HEMT, designed by GARD), LO and RF optical components all integrated on a single bracket and, provided it is necessary, could be removed from the dewar as a single package.  When installed in the dewar, SHFI Band 2 mixer assembly is also completed with wiring, cabling, thermal straps, etc.  Figure 22 a, b demonstrate SHFI B2 mixer assembly installed in the dewar.  For reference and part identification, please refer Figure 20.

Figure 22a.  Picture of the SHFI Band 2 Mixer assembly inside the SFHI dewar; equivalent to the CAD picture, Figure 21.

Figure 22b.  Picture of the SHFI Band 2 Mixer assembly inside SFHI dewar; equivalent to the CAD picture, Figure 21.

In order to characterize/verify the quasi-optical design and to align the receiver beam of the APEX-facility receiver, vector-field measurements are performed by scanning a transverse plane with respect to the axis of propagation and measuring the complex field distribution. The characterization includes determining of the propagation direction, the size and position of the beam waist, as well as the shape of the beam.  The measured data is fitted to the fundamental Gaussian beam mode by optimizing the power coupling coefficient. In order to achieve optimum coupling between two beams, the simulated and the measured beam, not only the axis of propagation needs to be the same, but also the size of the beam radius and the radius of curvature. Figure 23 shows a 3D plot of the measured beam at a frequency of 319 GHz at a distance about 400 mm from the waist position. Figure 24 and Figure 25 show power- and phase- distribution at a cross-section of the measured beam pattern together with the values for the targeted Gaussian beam. It can be seen that Gaussian beam approximation agrees very well with the measured data and a power coupling coefficient of about 99% was obtained when a fitting down to an on-axis relative power level of 20 dB.

Figure 23.  APEX-Band2 beam pattern measured at 319 GHz at scan distance approximately 400 mm from the waist position.

Figure 24.  APEX-Band1 beam pattern measured at 319 GHz: the beam cross-section with measured values (-o) and fitted Gaussian beam approximation (x).

Figure 25.  APEX-Band2 beam pattern measured at 319 GHz: the beam phase cross-section with measured values (-o) and fitted Gaussian beam approximation (x).

Important note:  Performance data presented in the figure below for the SHFI B2 are indicative and will be replaced later with the final performance date from the tuning tables.

Figures 26. Measured performance of the SHFI B2 mixer.

 

Figure 27a. Total power Allan variance (vertical axis) vs. integration time (horizontal axis, log scale, sec) measurement result for SHFI B2 taken at the LO frequency 363 GHz IF bandwidth of 500 MHz and sampling time 50ms.

Figure 27a. Spectroscopic Allan variance vs. integration time measurement result for SHFI B2 taken at the LO frequency 361 GHz IF bandwidth of 3 MHz and sampling time 150ms.

Figure 27b shows the corresponding spectroscopic stability measured with spectrum analyzer for four different IF frequencies – 4, 6, 7.5 and 8GHz. The temperature stabilization of the mixer was not engaged during this measurement

 

Contacts:

Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se
Mixer design, performance measurements: Vessen Vassilev vessen.vassilev(at)chalmers.se, Raquel Monje raquel.monje(at)chalmers.se, Doug Henke doug.henke(at)chalmers.se
Beam measurements: Olle Nystrφm olle.nystrom(at)chalmers.se

7.  SHFI Band 3 385-500 GHz mixer assembly

Multiple attempts of GARD to make the APEX 3 mixer with LO injection on the chip were unsuccessful; these attempts were focused on getting 2SB configuration.

Taking into account the above circumstances, the main focus of the work was shifted on developing a DSB mixer in a simplest possible configuration, to provide solid, swift and compatible solution to be used for APEX Band 3 channel.  The description below presents results of the development for the APEX Band 3 DSB mixer as by December 2009.

In order to keep compatibility with the SHeFI cryostat, the mixer block has been designed to ensure careful fit with the original Band 3 optics, IF chain and mechanical interface.  Such design was tested in the SHeFI cryostat prior to the pre-shipment review and, thus, the installation of the APEX Band 3 mixer on the site should not be a concern.

Figure 28.  CAD generated drawing of the SHFI Band 3 Mixer assembly.  Here the drawing shows 2SB configuration.  In the DSB, the IF chain has only one circulator and amplifier, see Fig. 29.

 

Figure 29a.  Picture of the SHFI Band 3 Mixer assembly inside SFHI dewar in DSB version; compare to the CAD picture, Figure 28.

Figure 29b.  Picture of the SHFI Band 3 Mixer assembly inside SFHI dewar in DSB version; compare to the CAD picture, Figure 28.

APEX Band 3 DSB Mixer Performance

The mixer chip was fabricated using GARD Nb SIS technology resulting in the high quality SIS junctions with average yield of better than 90% ensuring reliability of the SIS mixer.  The SIS mixer was mounted the same way as already installed SHeFI mixers and the electrical contacts were made using wire-bonding with gold wire of 18 um diameter.  The assembly technology ensures reliable and stable electrical performance of the mixer.  For Josephson current suppression, a Cu wire compact coil was used.  The coil and the SIS junction bias-T are integrated with the mixer block facilitating the integration of the mixer into the SHeFI cryostat.

Figure 30.  Unpumped IV curve of APEX Band 3 (left) with applied optimal coil current. Inset shows zoomed -2mV to 2 mV range of the IV curve. The Josephson super-current is completely suppressed.   IV curve (right), black, and the corresponding IF output power as a function of the DC bias voltage, red curve. LO frequency is 444 GHz. A room temperature load (297 K) is placed at the cryostat window.

The noise temperature of the mixer was measured in the test cryostat using liquid He.  This, in contrast to SHeFI, set the physical temperature of the mixer operation to approximately 4.3K (for typical atmospheric pressure in Gothenburg).  The Y-factor measurements were performed with the cold load instantly removed from liquid Nitrogen and hold for necessary time (roughly 30-40 sec) in front of the dewar window.  To address the fact that the cold load is exposed to 300K atmospheric air, the effective brightness temperature of the cold load were assumed to be 80K.  Using a foam container to keep the cold load immersed in the liquid Nitrogen has no advantage as the RF losses in the foam material and reflections from the surface of the liquid Nitrogen add to the brightness temperature of the cold load.  The test cryostat input window and the optics were not the same as in the SHeFI dewar, in particular we have to use the Teflon lens, which cause additional optical loss, that should not be present in the SHeFI cryostat.  A Teflon plastic window was used in the test cryostat as compared to the QMC fabricate crystal quartz window with matching cover in the SHeFI dewar.  These considerations lead to the suggestion that the measured APEX Band 3 DSB mixer noise temperature should be treated as a pessimistic estimate of the performance in the SHeFI cryostat also providing the fact that the mixer operational physical temperature could be about 0.5 K lower.

Figure 31. DSB noise temperature as a function of LO frequency averaged over IF 4 - 8 GHz. At the left: measurements are performed in a test liquid helium cryostat, as described above.  Please note that the RF band coverage is 382 – 506 GHz.  The green line shows ALMA Band 8 specifications.  At the right: Resulting change in the mixer noise temperature clearly correlated to the absorptive properties of the air (according our calculation 70% correlation).  The mixer is operated as DSB and thus averages the noise from the two sidebands (3-8.5 GHz) around each LO frequency.

Figure 31a.  Comparing of the measurements performed at the telescope and in the Lab. The Band 3 mixer is integrated into SHeFI cryostat and uses the facility optics (Y-factor measurements performed with a black body placed at the dewar window).

Variation of the IF power vs. IF frequency across the band is +/- 3.5 dB.  Yet, the ripple depends on the configuration of the components inside the test cryostat and is expected to be less in the SHeFI cryostat with extra cold amplifier and the additional circulator damping possible standing waves.

The noise temperature is reasonably flat across the IF band of 3.5 – 8.25 GHz, while the specifications are 4 – 8 GHz IF.

Figures 32.  Output IF power as a function of IF frequency. LO frequency is 444 GHz (left picture). DSB noise temperature as a function of IF frequency performed at different LO frequencies (right picture).

Figure 33. Total power Allan variance (vertical axis) vs. time (horizontal axis, log scale, sec) Measurements at the APEX Telescope, SHeFI dewar. Left: Allan variance of Channel 200 MHz of the FFTS (1 MHz BW), RF 498 GHz (LSB) spectroscopic Allan variance of Channels 200-800 MHz of the FFTS (1 MHz BW), 498 GHz (LSB).

 

Example:

Contacts:

Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se
Mixer design, performance measurements: Denis Meledin denis.meledin(at)chalmers.se, Victor Belitsky victor.belitsky(at)chalmers.se;
Beam measurements: Olle Nystrφm olle.nystrom(at)chalmers.se


8.  SHFI Band T2 12-1250 – 1390 GHz mixer assembly

SHFI APEX Band T2 mixer is based on HEB mixers and used balanced scheme.  HEB mixers do not need any magnetic field as they use completely different physics; they also operate at typically lower than SIS mixer bias voltages. Mixer details could be found in published papers collected on the supplied DVD under SHFI_APEX_Band_T2 folder.

Figure 32.  CAD generated drawing of the SHFI Band T2 Mixer assembly.

Figure 32 shows details of the SFHI BT2 Mixer assembly consisting of the mixers, the middle piece, IF amplifier (2- stage GaAs HEMT, 2-4 GHz, designed by GARD), 3 dB 180o IF hybrid, LO and RF optical components all integrated on a single bracket and, provided it is necessary, could be removed from the dewar as a single package. When installed in the dewar, SHFI Band T2 mixer assembly is also completed with wiring, cabling, etc. Figure 33 a, b demonstrates SHFI BT2 mixer assembly installed in the dewar. For reference and part identification, please refer Figure 32.

 

Figure 33a.  Picture of the SHFI Band T2 Mixer assembly mounted inside SFHI dewar; compare to the CAD picture, Figure 32.

Figure 33b.  Picture of the SHFI Band 2 Mixer assembly inside SFHI dewar; compare to the CAD picture, Figure 32.

In order to characterize/verify the quasi-optical design and to align the receiver beam of the APEX-facility receiver, amplitude measurements are performed by scanning a transverse plane with respect to the axis of propagation and by several such scans at different distances determine the propagation direction, the size and position of the beam waist, as well as the shape of the beam. By fitting a line through the amplitude centers any tilt of the beam can be determined. Figure 34 shows a 3D plot of the measured beam at a frequency of 1335 GHz at a distance about 400 mm from the receiver window. Figure 35 shows power distribution at a cross-section of the measured beam pattern together with the values for the modeled Gaussian beam. In Figure 36, a contour plot of the beam at 1335 GHz can be seen.

Figure 34.  APEX-BandT2 beam pattern measured at 1335 GHz at scan distance approximately 400 mm from the window.

Figure 35.  APEX-BandT2 beam pattern measured at 1335 GHz: the beam cross-section with measured values (-o) and fitted Gaussian beam approximation (x).

Figure 36.  APEX-BandT2 beam pattern measured at 1335 GHz measured contour plot.

Figure 37 Allan variance (vertical axis) vs. time (horizontal axis, log scale, sec) measurement result for SHFI BT2 taken at the LO frequency of 1294 GHz and at 6 GHz IF within 500MHz bandwidth using Agilent power meter. Sampling rate is 50 ms.

Figure 39. Noise performance vs. RF for SHFI BT2 measured in the SHFI dewar taken at IF frequency of 3 GHz within 100 MHz bandwidth (without 5-7 GHz up-converter)

Figure 38 Noise performance vs. IF (at)Flo=1320 GHz for SHFI BT2 measured in the SHFI dewar (without 5-7 GHz up-converter).

 

Contacts:

Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se;
Mixer design, performance measurements: Denis Meledin denis.meledin(at)chalmers.se;
Beam measurements: Olle Nystrφm olle.nystrom(at)chalmers.se


9.  SHFI IF System Diagram and Cold IF Amplifiers

All SIS mixers of SHFI have an IF band of 4 – 8 GHz.  All SIS mixer assemblies have a common output and are merged via the power combiner installed prior to the second IF amplifier installed at 12 K cooling stage. Idle channel IF amplifiers at 4 K stage normally are un-powered and do not contribute to the output of the power combiner.  The SHFI Band T2 HEB mixer has IF band 2-4 GHz and uses separate output at the SHFI dewar bottom plate; Band T2 signal is upper-converted to 5 - 7 GHz band (note the spectrum frequency inversion 2 GHz->7 GHz and 4 GHz->5 GHz). Coaxial switch connects Band T2 and Band 1 to Band 3 IF output signals to one output IF port. Figure 39 shows block diagram of SHFI IF system.

APEX users guide for cryogenic 4-8 GHz low noise amplifiers.

 

This chapter describes the use and handling of the cryogenic low-noise amplifiers used in the SHFI. There are three types of amplifiers in use: 2 and 3-stage 4-8 GHz and 2-stage 2-4 GHz. The 2-4 GHz version is only used for band T2.  To distinguish between the different versions, the overall dimensions (without connectors) can be measured and compared with those in the Table 1. The 2-4 GHz amplifiers may have different height, depending on if the SMA connector is a 2-screw or 4-screw version.

 

Type

Width [mm]

Depth [mm]

4-8 GHz 2-stage

34.5

32

4-8 GHz 3-stage

31.5

27

2-4 GHz 2-stage

43

40

Table 1 - Amplifier dimensions

 

Figure 9.1 and Figure 9.2 show a visual comparison of the dimensions for the 4-8 GHz versions. Please, note that the connectors’ gender may differ from those on the pictures.

 

Figure 9.1 - 2-stage and 3-stage side by side, gender of SMA‑connectors may differ.

Figure 9.2 - 2-stage and 3-stage, gender of SMA‑connectors may differ.

 

The following abbreviations for the bias are used:

Vd – Drain Voltage;
Id – Drain Current;
Vg – Gate Voltage,
where, e.g., Vd2 is the drain voltage for the 2nd stage of a given amplifier.

The HEMT has three terminals, gate, drain and source as in Figure 3. The operation of a HEMT is such that a negative voltage is applied on the gate, and a positive voltage is applied on the drain. The drain current, Id, is then adjusted by the negative gate voltage, where a voltage approaching -0V results in a larger current. In an ideal device, no current is drawn through the gate. The source is connected to the box through bond wires and then to pin nr 1 on the bias connector.

Figure 3 - HEMT schematic

 

The bias connectors for the different cryogenic amplifiers are compatible, with the exception that the Vd3 and Vg3 are not used on the 2-stage version as in Table 2 and Figure 4. The same cable can be used, even with bias applied to Vd3 and Vg3 when connected to a 2-stage device without any potential risk of damage.

 

Table 2 – Bias connector pin-out

Figure 4 - Bias connector pin-out

 

Amplifier Type

Drain Voltage, Vd [V]

Drain Current, Id [mA]

4-8 GHz 2-stage (all stages)

1.4

4.5

4-8 GHz 3-stage (all stages)

1.4

4.5

2-4 GHz 2-stage (stage 1)

1.8

5

2-4 GHz 2-stage (stage 2)

1.5

5

Absolute maximum Vd

2

 

Table 3 - Recommended bias settings

 

Bias operation:

The biasing is normally done automatically by the control system but may allow manual biasing. For manual operation through the control system, the following procedure and recommended bias setting (Table 3) should be used:

To enable a stage:

To turn a stage off:

The reason for this procedure is that if Vg is set to 0 V and Vd is brought positive, the transistor will operate at an unsafe bias where it consumes a lot of current and will be damaged.

Contacts:

Amplifiers: Erik Sundin erik.sundin(at)chalmers.se;
System: Victor Belitsky Victor.Belitsky(at)chalmers.se


10.  SHFI LO System Diagram

All SHFI LO sources have been ordered from VDI Inc. and based on VDI’s “frequency extension module” (FEM) consisting of a series of Schottky diode multipliers and a power amplifier stage.  The power amplifiers for each channel have separate DC bias input for the drain voltage of the last stage; this voltage, called on the LO diagram “control” is used for altering the output power of the amplifiers and adjusting the final LO power level.  Please note, that in certain cases, like for example for SHFI BT2, the control voltage is quite high and the consumed current could exceed 2 amps.  All power amplifiers are operated in the deep saturation to reduce amplitude noise.  After receiving the sources from VDI they have been re-engineered to be usable within APEX project, namely, the layout was changes to fit available room at the common warm optic plate, diagonal horns were added (except for T2 LO with built-in diagonal horn), focusing lenses and support brackets were designed.  GARD has added extensive cooling radiators to cool down the power amplifiers.  The fans installed around the room-temperature optical plate with LO must be always going during the receiver operation. APEX B1 source went trough more thorough redesign as it exhibited a higher level of noise as compared to Gunn- based source.  At the moment of creating this document the origin of this extra noise in SHFI B1 LO is still under investigation.

The FEM used as the SHFI BT2 LO needs a DC bias of the first two doublers after the power amplifier. In order to provide remote operation of this LO source, GARD has built a dedicated DC bias unit, which provides all the necessary voltages and is operated via the control system.  Detailed description of this DC bias unit can be found in the LO_docs folder together with operation manual for VDI FEM T2.

As all the LO sources are of a direct multiplication type, there is no need for a PLL system; the reference source signal from APEX facility synthesizer is directly multiplied by the necessary factor to get the desirable LO frequency and power is provided by adjusting the corresponding control voltage through the control system.  In order to bring the input signal up to the required power level, the LO system input uses the MITEQ power amplifier MITEQ AFS3-00502000-18P-4 with an average 25 dB gain; the required input power is in the range of -5 ..-9 dBm (TBC).  All input power settings must be done via the control system; exceeding the input power level will destroy the input circuits of FEM.  All original manuals for VDI FEMs are placed in the folder LO_docs.

 



11.  SHFI Mixer DC BIAS

All mixers of SHFI adopt the same bias circuitry depicted below.  The only difference is that all SIS mixers have the resistor part with capacitors integrated inside the mixer block, while the HEB mixers for APEX Band T2 have the resistor circuitry on a separate PCB inside a bias box and a coaxial NARDA bias-T.

 

The SIS/HEB bias supplies are very similar to what was used with the APEX2a receiver, however the electronics and the layout have been somewhat changed in order to fix some circuit and performance problems.

 

 


 

SIS Bias Supply

Operation Manual

 

 

 

 

 

 

 

 

 

1      Overview.. 51

1.1       Block Diagram.. 51

2      Computer Control and Power Requirements. 52

2.1       Connections and Cables. 53

3      Analog PCB.. 54

3.1       Additional Information. 56

4      Digital Board. 57

4.1       Additional Information. 58

4.2       Filter Board. 58

5      Front Panel and PCB Layout 59

6      Tuning. 62

7      Grounding. 63

8      Pictures. 64

 

 

 

 

 

 

Updated: Oct 29, 2007


1.         Overview

The bias supply provides DC bias to the junction over the range of ±10 mV with a 5΅V resolution.  The voltage is set via a 12-bit DAC which is optically isolated from the computer control.  A voltage feedback operational amplifier is driven by the DAC to stabilize the junction voltage.  All biasing circuitry is powered from isolated “floating” power supplies.  In order to match the ground potential of the mixer block (cryostat chassis), another feedback opamp is used to equalize the floating ground with respect to the chassis ground (or non-floating ground).

 

The user may select between internal and mixer.  With the switch depressed (in mixer mode) the SIS mixer will be biased through an external cable.  If the switch is in the outward position, the internal “dummy” circuit is used to emulate the mixer and the user may use this position for verification purposes.

 

The voltage and current are each sensed using a instrumentation amplifier which has a gain of 100.  The voltage is measured across the mixer junction and the current is determined via a voltage drop across a 10Ohm resistor.  In general, the junction voltage should stay within the range of ±10mV resulting in ±1V to the control system analog input.  The current may be as high as 2 or 3 mA resulting in a voltage of ±3V to the analog input to the control system.

1.1.       Block Diagram


2.         Computer Control and Power Requirements

The bias box requires 4 digital outputs (0-5V), 2 analog inputs, and 2 power supplies.  As an example, the APEX control system uses compact Fieldpoint for control and monitoring; the cFP-DO-403 module is used for the Enable, SetToZero, and UpDown digital outputs, the cFP-AI-118 module is used for the Vmon and Imon analog inputs, and the cFP-PG-522 module is used for the pulse generation.

 

The analog inputs should be capable of measuring ±5V with a resolution of 5΅V x 100 = 500΅V.  The cFP-AI-118 module is 16-bit which exceeds this requirement.

 

A schroff PSK 215 may be used to supply the non-float power and a PSK315 for the float supplies.  The following table summarizes the current requirements.

 

Bias power

Maximum Current

Schroff Supply

Rated Current

+5V Float

20mA

PSK315

0.5A

+15V Float

25mA

0.2A

-15V Float

25mA

0.2A

 

 

 

 

+15V NFloat

10mA

PSK215

0.4A

-15V NFloat

10mA

0.4A

 

Note that the APEX control system powers 8 bias boxes off the same NFloat Schroff supply.

 

2.1.       Connections and Cables


3.         Analog PCB


3.1.       Additional Information

The output voltage from the DAC (±5V) is divided by a factor of 500 to give ±10mV across the junction.

 

Using two OP27E operational amplifiers, the voltage across the junction is established.  Both opamps have feedbacks; one fixing the voltage and the other matching the floating ground to the chassis ground.  It is important to note that the voltage feedback wire, VI, is shared with the inamps.  Furthermore, each wire that goes to the mixer through the 6-pin Fischer connector has EMI filtering at the cryostat wiring terminal and extra impedances due to the cryo-wire.  With regard to the voltage feedback, there is a feedback capacitance C5 in parallel with feedback loop.

 

Any apparent noise that is measured at the output is a function of the analog voltage circuitry and the readout circuitry.  Changing the feedback capacitance of the Vin opamp helps to limit the bandwidth of the sourcing voltage (along with the capacitors at the junction).  The readout circuitry herein includes the DAC of the control system, but also the inamps (AMP01) of the bias box.  Since these inamps share the negative feedback terminals of the opamps, noise may be present on the readout that is actually not present across the junction. 

 

The maximum voltage is a function of the current passing through the junction.  Remember that the junction is in series with a 10Ohm resistor and that these 2 components are in parallel with a 10Ohm resistor.  Both opamps have a 100+500Ohm resistance on the output which drops some of the available voltage.  Furthermore, on high currents, there is some drop across the supply resistors, which drops the rail voltage.

 

The switch on the analog board is an 8 position switch that allows the user to switch between external connection and internal emulation.  The two extra switches are wired up to the Vin and GndRET (the outputs of the opamps) in a de-bouncing fashion.  When the switch is not depressed, the internal circuit is measured, and also the Vin and GndRET wires (with respect to the external SIS circuit) are grounded to chassis ground.  When the switch is depressed, the de-bounce time constant prevents transients across these wires.  There are 2 insulating washers on the front panel; one for the 6-pin Fischer mixer connector and one for the control cable.  The red “NFG” wire on the front panel provides the solid non-floating ground between the cryostat and each bias box (refer to section on grounding below).  This grounding provides lower noise for the readout circuitry (the mixer bias is still floating and should not be affected).

 

Each bias box must be tuned using an external mixer emulation circuit that uses precision 10Ohm resistors.  The gains and offsets should be adjusted according to the procedure outlined below.

 


4.         Digital Board

4.1.       Additional Information

The digital output control signals are optically isolated from the floating supply circuitry.  Depending on the type of outputs from the control system, the forward current through the photocouplers should be determined to maximize the current transfer ratio (consult data sheets).  Note that the pulses opto-coupler has different requirements.

 

The digital circuitry has been routed such that if the control signals Enable or SetToZero are 0 or 1 respectively, the bias box will be reset to zero voltage.  If Enable is high and SetToZero is low, then the counters are active and change with the applied Pulse cycle.  The counters are connected to D0-D11 of the DAC to give 12 bits of resolution to the output voltage. 

4.2.       Filter Board

 

5.         Front Panel and PCB Layout

 

 


6.         Tuning

  1. Zero the Multimeter and let boxes warm up for 30 minutes with bias of 9mV.
  2. Zero the in-amps

a.                   Remove J3 & J12

b.                  Jumper the inputs (J4 & J9)

c.                   Jumper the Rg inputs (J17 & J16)

d.                  Measure output at J8 & J10 (using wires for PCB measurement)

e.                   Tune input offset (VR4 & VR7) for zero

f.                   Remove Rg jumpers & tune output offset (VR5 & VR8) for zero

  1. Zero the op-amps:

a.                   Remove J4 & J9

b.                  Jumper the inputs (J2 & J11)

c.                   Measure output at J3 and J12

d.                  Tune offset output at VR2 and VR9 for zero

e.                   Note, there will probably be some remaining offset, because the load is sensed through high value resistors and the input bias current of the opamps cannot be tuned out.

  1. Compensate for remaining offset

a.                   Install J3 & J12 and J8 & J10

b.                  Use the tuning software  to set voltage to zero

c.                   Attach external dummy to bias box and depress front panel button to activate the “mixer” setting

d.                  Measure the Voltage offset of the external dummy with a BNC cable

e.                   Tune VR9 (GndRET offset) to give 0.0mV

f.                   Looking at the software readout, tune VR4 & VR7 for 0.0mV

  1. Tune the gain:

a.                   Using the bias box tuning software to send 1843 pulses (set to zero first).  This will give 9.0mV

b.                  Measure the voltage on the external dummy with the multimeter and tune VR1 to give 9.0 mV

c.                   Also check with 600 pulses (2.93mV)

d.                  Adjust gain of In-amps: repeat a-c but look at the software readout and adjust VR3 & VR6

  1. Verify the tuning by setting to zero first and then ±9mV, and ±3mV

 


7.         Grounding

8.       Pictures

Analog PCB

 

 


 

     

Internal “Dummy” Emulator                                                            Switch depressed in “Mixer” position

 

Contacts:

Doug Henke doug.henke(at)chalmers.se

 

 

 


12.  SHFI Wiring Diagram

Internal SHFI wiring is presented in this chapter in the form of wiring diagrams.  The SHFI external cabling including connections to the control system and power supplies are listed in the separate document for the SHFI Control system.

All wiring inside the cryostat employs Lake Shore cryogenic twisted pair wires; in case of wire replacement it is extremely important to keep original arrangement of paired wires to minimize influence of EMI and specifically magnetic field interference.

Contacts:

Cryogenic cabling and wiring: Mathias Fredrixon, mathias.fredrixon(at)chalmers.se
Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se;


 

 

 

 

 

 

 

 

13.  SHFI Control System

A.  SHFI Control System: Hardware and Cabling

Due to the volume of the current document, this entire chapter is delivered as a separate document. This separate document describes the hardware solutions for the entire control system, including data acquisition and measurement, digital and analogue control interfaces, power supplies and LO power control units, and cabling.  The document SHFI_Control_System.pdf is in the root of the DVD and supplementary docs in Control_Hardware folder.

Contacts:

SHFI Control System: Magnus Svensson magnus.svensson(at)chalmers.se;
Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se;
Cryogenic cabling and wiring: Mathias Fredrixon, mathias.fredrixon(at)chalmers.se

B.  SHFI Control System: Software and User Interface

Due to the volume of the current document, this entire chapter is delivered as a separate document. That document describes the software control system, including GUI and provides reference data on the computer used.  The document SHFI_Control_Soft_linux.pdf is in the root of the DVD and the computer documentation in the directory Comp_docs.

Contacts:

SHFI Control System: Michael Olberg michael.olberg(at)chalmers.se;


13.  SHFI Maintenance Procedures

A.  SRDK-3ST Maintenance

Sumitomo SRDK-3ST, three – stage 4K, 1W cryocooler have been purchased via Janis Research Company, Ltd., USA, contact person Mr. William R. Shields, bshields(at)janis.com.  For maintenance refer Sumitomo HI documentation in the folder Sumitomo_SRDK-3ST-A61D-A in the DVD, for repair contact Janis Inc., USA.

Contacts:

Vacuum: Alexey Pavolotsky alexey.pavolotsky(at)chalmers.se;
Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se;

B.  Warming up and Venting Procedure

In order to proceed with warming up and venting of SHFI dewar the following procedure must be followed strictly:

Warming up

1. Ensure through the control system that all SHFI bands are “off” together with other facility components, like IF amplifier (s) at 12 K stage, and are set to DC bias zero. If necessary, the entire SHFI control system with all respective power supplier rack mounted units should be put off; in this case the user looses possibility to read out the temperatures of the components inside SHFI dewar.

2. Switch the compressor off and wait until all thermal sensor show similar temperature around 288 – 290 K (it is assumed that the temperature stabilization in the Cabin A supports the temperature at 15 C !).  The time period between switching off the compressor and the complete warm up depends on residual vacuum; normally the warming up takes 7 or more hours.  If there are no plans to open SHFI dewar, the pumping could be engaged during warm up process (see for instructions chapter C); in case of engaging the pumping, the complete warm up time could be as long as 48 hours.  If no temperature sensors are used during the warm up to monitor the temperature inside the SHFI dewar, the wait time should be doubled.

Venting

1. Ensure that warming up procedure has been completed and all receiver components have the same temperature as the ambient temperature in the room (see Chapter above); if some components are still cold, during the venting water will be condensing and damage those components.

2.  Ensure that the pumping station is off at least 15 minutes prior to the moment you plan to vent the dewar. If turbo is running, follow the switching off pumping procedure below.

2a. In order to switch off pumping, halt turbo-pump and watch its deceleration over the display of the turbo-pump control unit. When turbo-pump’s rotation speed falls under 20 Hz (within 5 minute usually), the backing pump should be switched off.

3. Open the pumping gate valve (see the pumping down procedure) and after that the venting valve.  Venting flow is intentionally made very small to ease recovery for the foam material used for IF filters.  Quicker venting will destroy IR filters used for all LO and RF windows, except for SHFI Band T2.

4. Always minimize the time during which the SHFI is exposed for atmosphere.

 

Contacts:

Vacuum: Alexey Pavolotsky, alexey.pavolotsky(at)chalmers.se;
Mechanical, cryogenic and optical design: Igor Lapkin, lapkin(at)chalmers.se;


C.  Pumping down and Cooling Procedure

In order to proceed with bringing the SHFI into operation the following procedure(s) must be followed strictly:

Pumping procedure could be used for bringing the SHFI into the operation after opening of the dewar for example, for repair.  In that case, one should ensure that all necessary mechanical connections have been made, all vacuum interfaces are tightened and O-rings are not damaged during the disassembly.

1.  Ensure that the venting valve is closed.

2.  Ensure that the pumping gate valve is open.

3.  Switch on the backing pump and observe the SHFI internal pressure reaches the level below of 20 mBar (TBC!);

4.  Engage the built-in turbo-molecular pump;

5.  Wait until the pressure inside the dewar reaches lower than 5x10-5 mBar (TBC!) (pumping time depends on the time duration the dewar was exposed to the atmospheric pressure and can be more than 8-10 hours); for reference, the ultimate vacuum reached in the lab conditions was 4x10-7 mBar (cold). Note: it is advisable to start the cooling (step 6) after reaching the ultimate pressure.

6.  Switch “on” the compressor to start up the cooling process.  For the typical cooling curve see Figure 13, page 18.  After approximately 8 hours the SHFI cryogenic parts should reach its operating temperatures.  Control the temperatures via SHFI Control system.

7.  After the receiver temperature has reached the desirable level, close the pumping gate valve.  Watch the vacuum level is stable and is not degrading.

Note:   If the vacuum level in the dewar does not reach the level of 5x10-5 mBar after extensive pumping before engaging compressor; and/or

            If the vacuum level in the dewar does not reach the level of 1x x10-6 mBar after SHFI reaches its operating temperature; and/or

            If the vacuum level is unstable and degrades after pumping gate valve is closed,
these points out to the sufficient vacuum leak. In such case, normal SHFI operation is not possible. The system should be warmed up (see the section 13
B above for the reference) and searched for leak.

8.  Switch off the turbo pump;

9.  Wait 5 minutes (TBC!);

10.  Switch off the backing pump.

Contacts:

Vacuum: Alexey Pavolotsky alexey.pavolotsky(at)chalmers.se;
Mechanical, cryogenic and optical design: Igor Lapkin lapkin(at)chalmers.se;

D.  SHFI Dewar Opening Procedure

Removing SHFI from the supporting frame

1.  Ensure that the described above Warming up and Venting Procedure has been completed;

2.  Ensure that all SHFI room temperature electronics is switched of completely;

3.  Disconnect powers supply cables going to the DC bias rack box mounted at the SHFI receiver frame (Figure 1); leave all cables for DC mixer bias connected; disconnect all DC cables for the IF amplifier power supplies, thermometers and defluxing resistors and Band T2 IF up-converter; disconnect the cables from the rotating stage (channel switch);  depending on available length of the helium lines inside Cabin A it may be required to disconnect them from the cooler; disconnect the turbo-pump exhaust line (install blinders / covers on both sides) see Figure 4;

4.  Disconnect SMA contacts of the receiver IF output (from the IF coaxial switch) and the reference signal for T2 IF up-converter; the SMA cable providing reference signal to SHFI LO reference signal should not be disconnected.

5.  Move away all lines or cables precluding removing the SHFI from the instrument supporting frame of the Cabin A.

6.  Release 8 bolts M10 keeping SHFI connected to the plate with receiver warm optics and LOs (see Figure 1, 2 and the Figure below).

7.  Land SHFI dewar on the wheels by adjusting the length of the four support pads under the SHFI frame corners.

8.  Roll the SHFI dewar from under the Cabin A instrument support frame structure.  Note: carefully avoid damaging of the turbo-pump integrated at the back side of the SHFI dewar.

Important Note: all following steps must be done in a controllable environment, free of any dust, dirt or other contaminations alike.

Opening SHFI Dewar

9.  Release and remove the claw grips, which secure the top plate to the vessel. Lift up carefully the top plate and put it aside. Both the top and the bottom surface of the top plate must be protected with bubble-film against scratching of the vacuum sealing surface and breaking of the quartz windows. Watch the centering ring with viton O-ring not to be damaged and contaminated.  It is important to avoid damaging of the shutter which is connected to the top plate via vacuum feed-through and made of very thin aluminum.

9a.  Release and remove the claw grips connecting the SHFI outer vessel and the base plate (see Figure 4; note that the claw grips are not presented on the figure above as they were not used in the Laboratory environment).

Figure capture: disassembling of 70 K shield of SHFI dewar.

10.  Slide strictly vertically the outer SHFI vessel upwards until the 70 K shield will be completely cleared and after that put it aside. While sliding the outer SHFI vessel up, it is recommended to look from the top side to avoid crashing it into the 70K shield wall. Note, that due to total weight and shifted center of gravity, at least 3 persons (one supporting the turbo-pump) must be involved in the removing of the outer vessel of SHFI.

11.  Place hard plastic protection onto the SHFI base plate. Remove the 70 K shield by releasing the screws connecting it to the 70 K cold plate (see Figure below);

Figure capture: disassembling of 12 K shield of SHFI dewar.

12.  Remove the 12 K shield be releasing the screws connecting it to the 12 K cold plate (see Figure below);

SHFI Cryostat assembly should follow the steps above in reverse order. Pay attention to the proper aligning outer SHFI vessel and the top plate, the 70K shield and the 12 K shield, following the relevant alignment marks. While sliding down the outer SHFI vessel, it is recommended to look from the top side to avoid crashing it into the 70K shield wall. Even though all threaded holes in the thermal shields have Helycoil inserts, persons assembling the SHFI cryostat should use controllable tightening for all screws.

Contacts:

Vacuum: Alexey Pavolotsky alexey.pavolotsky(at)chalmers.se;
Mechanical, cryogenic and optical design: Igor Lapkin, lapkin(at)chalmers.se;
Cryogenic cabling and wiring: Mathias Fredrixon, mathias.fredrixon(at)chalmers.se