Version 13 July 1999; Text: 5 May 1999

Note: This description is old but still accurate regarding the optics. The main changes are:

• The C band mixers has been replaced with tunerless B band mixers for use with the SMA.
• The D band mixers has been upgraded to tunerless mixers of HIFI design.
• The Water Vapour Monitor (WVM) which used to be mounted above RxW has been removed.

Optics

The receiver optics is built on an optical plate. The plate is mounted facing the front of the receiver and fills most of the receiver frame. In addition the plate is divided into a left and right part by another plate which is mounted vertically with the edge towards the front. This plate minimizes cross talk between the A channel to the right and the B channel to the left. In the picture components behind the optical plate are dashed while components that can be seen in front of the plate are drawn with solid lines.

The incoming radiation enters the receiver in the upper right corner. It goes through the optical plate and is reflected downwards and towards the center line behind the optical plate by M1. It passes the chopper Ch1. It then reaches a polarization splitting grid G1 that is placed on the vertical center line. The A channel radiation is reflected by the grid while the radiation the B channel sees goes through the grid. Thus the A channel is on the right side while the B channel is on the left side of the receiver frame. The grid is oriented so that, refereed to the receiver frame, the A channel sees vertical polarized radiation and the B channel horizontal. The two channels are mirror images of each other and are grid separated by a vertical plate on the center line of the receiver to minimize crosstalk. The following description applies to both channels. After the grid the radiation is reflected by the mirror M2, goes through the optical plate and enters the first Martin-Puplett (MP) interferometer. This MP interferometer is used for sideband separation. The fixed corner mirror of the MP is just in front of M2 but drawn with solid lines since it is in front of the optical plate. The moveable corner mirror of the MP is above the fixed corner mirror but still in front of the optical plate. The radiation leaves the MP going downwards in front of the optical plate. It passes a LO injection grid G3 that can be rotated so LO can be injected from the left (650 GHz band) or right (450 GHz band). The incoming radiation and the LO power continues downwards and enters the second Martin-Puplett interferometer. The LO injection MP aligns the polarization of the signal and LO. The radiation and LO leaves the LO MP going inwards through the optical plate and is reflected first by M3 which is hidden behind the fixed corner mirror of the LO injection MP. The radiation is further reflected by M4 which is behind the fixed corner mirror of the LO injection MP before it enters the dewar. Inside the dewar there is a final grid that separated the two mixers. Any radiation reflected by the grid enters the 450 GHz mixer while radiation that passes the grid enters the 650 GHz mixer. The sideband and LO MP interferometers has to be tuned so the right polarization comes into the dewar.

When the sky chopper Ch1 is out of the beam the receiver sees the sky while a calibration load is seen if the chopper is in the beam. The load chopper Ch2 selects between a ambient load on the side of the receiver frame or a cold load inside the dewar which is seen via reflection in the chopper and M5. The function of the side band chopper Ch3 is similar to the load chopper but selects if the unwanted sidebands sees a cold load or a ambient load. Of course in normal operation the cold load should be used.

The use of the Martin-Puplett interferometers are more intricate. If we look at channel A (to the right) the LO injection grid G3 is oriented with the wires perpendicular to the optical plate and horizontal. For the incoming radiation to pass this grid it must be polarized parallel to the optical plate. The polarization of the LO signals that is reflected into the beam will be the opposite i.e. perpendicular to the optical plate. Now the grid inside the dewar is oriented so the C band mixer sees horizontal polarization. Thus for the C band in the A channel the LO injection MP must rotate the signal so it emerges horizontal i.e. n * 180 degrees while the LO signals with a frequency 4 GHz away must be rotated n * 180 + 90 degrees. While if we used the D band the rotations would be reversed. The B channel is similar apart from the LO injection grid there is oriented with the wires parallel with the optical plate. The grid inside the dewar in the B channel is also mounted so the C band requires horizontal polarization. Thus there the C band signal must be rotated n * 180 + 90 degrees while the LO n * 180 degrees.

The sideband separating Martin-Puplett interferometers operates very similar. With a 4 GHz IF the separation of the sidebands are 8 GHz. The idea is to have frequencies separated with 8 GHz with the same polarization in come out with orthogonal polarization. This will make the unwanted sideband in the incoming sky emission to be reflected by the LO injection grid. By the nature of MP interferometers this implies that radiation 8 GHz away with perpendicular polarization will emerge from the MP with its polarization aligned with the signal. Thus in the A channel incoming radiation 8 GHz away from the signal but polarized vertical will be reflected by the grid G1 into the A channel as will the signal itself. The sideband filtering MP will turn the polarization of the unwanted sideband orthogonal to the signal so it is reflected out of the signal path by G3. On the other hand radiation 8 GHz away and polarized orthogonal to the signal will enter the signal path through G1 (for the A-channel). This radiation will come from one of the single sideband loads. The MP interferometer will turn the polarization of this radiation parallel to the sky signal and both will be seen by the mixer. Thus the mixer sees the sky at the signal frequency but a SSB load at the unwanted sideband.

The LO power coming from the C band source to the right or the D band source to the left is refocused by the mirror M6 and then reflected by G3 into the signal paths. The polarization of the LO power can be adjusted by a polarization rotator just before M6. By adjusting of the LO signal polarization we can equalize the LO power enter the A and B channel. By rotating the grids G3 we can inject C band or D band LO power into both mixers. In principle it is also possible to orient the grids so D band LO is seen by the B channel and C band LO by the A channel. In fact the G3 grids are oriented that way in the optical figure. However, since the receiver not has two complete phase lock systems we can not operate in a hybrid C/D band mode.

The dashed unit on the top left side is a laser used for alignment.

Physical Description

Receiver frame

The receiver frame is placed in bay 11 of the JCMT cabin. Below the receiver frame is the W phase lock system. Note the two wheels for rotating the LO injection grids just below the center of the receiver. The C and D band positions are marked. The grids reflects LO power into the signal path from the C band LO source to the right or the D band LO source to the left. They must be set to the correct band. The left and right grid injects LO into the B and A channel, respectively.

The two LO sources look identical. They both consist of a Carlstrom Gunn having two micrometers for tuning. The backshort tuner on the backside points downward and the frequency tuner on the top points sideways. Each turn of the micrometers are 25 divisions. After the Gunn follows a black LO power modulator. The modulators are controlled electronically. Next are two multipliers mounted together, each with two tuning micrometers. The micrometers are marked 1 - 4. On top of the multipliers are a polarization rotating turret. It is used for equalizing the LO power to the two mixers.

There are three choppers on receiver. The sky chopper up towards the top right corner selects between the receiver looking at the sky and a calibration load. The receiver is looking on a calibration load if the chopper blade is in the signal path and the sky if the chopper blade is out of the signal blade. Hence the marking in and out on the chopper control modules. Just below and to the right of the sky chopper is the load chopper. It selects between a cold load (in position) and an ambient load (out position).

Phase lock unit

The phase lock unit is placed below the receiver rack. It is divided between C and D band. The C band control is on the top and the D band on the bottom. The labels are to the left but a bit hard to see. Note the two sets of three LEDs right of the center. These indicates the status of switches and interlocks for the Gunns. One set of LEDs for C band and one for D band. All three must be on for the Gunn to be operating. The voltage and current through the Gunns can be read on the displays just to the right of the status LEDs. Even further to the right is another display with a selector switch to look at different test points. To the left is a level OK and a Lock OK LED and one analog meter for each Gunn. These indicates the state of the lock. The top set for C and the bottom set for D band. The Level OK LED is normally on after setting the backshort and frequency tuner on the Gunn according to the tuning table. Then tune the frequency tuner on the appropriate Gunn so the Lock OK light comes on and the meter needle stays in the center position. The tuning tables are good so there should be no need to tune far. The frequency tuner is sensitive, particular at D band. If it is impossible to get the Lock OK LED to stay on and the meter to stop in the middle the Gunn frequency might be on the wrong side of the reference. Just turn the tuner in a little bit and you should get a stable lock. Close to the edge of the tuning range the Level OK LED might not go on. If you can get the Lock light to go on all is still fine.

The tuning panel shown controls all the servo motors used on RxW. The buttons on the display selects a servo motor. However, not all servo motors were implemented in the final design. Some are "virtual".

The three buttons to the far right have special functions. If the light on the remote button on the micro is in remote mode and no servo can be controlled from the panel. Pressing any other button brings you into local mode. The VAX will force W into remote mode during observing. There is also an anchor switch. If the micro has problems with controlling a servo motor it will anchor the motor. The micro will not try to move an anchored motor - neither will it generate a fault. You can unanchor the motor by pressing the anchor switch with the appropriate servo motor selected. Pressing the switch when an unanchored servo motor is selected will anchor the motor - the switch works as an toggle. If needed you can anchor a servo motor. This stops the micro from moving that motor without generating a fault. Handy if something breaks and you want to move it yourself without micro interference. There is also the datum button. This starts the datum process for all four Martin-Puplett interferometers. The process takes several minutes during which nothing else can be done. This only has to be done if the micro has been rebooted not otherwise. It is best to let the VAX take care of the datum operation.

The rest of the buttons select servo motors - real or virtual. After a real servo has been selected you can adjust the servo with the big wheel to the lower right. The display in the upper left corner gives the current position of the servos. The letter to the right of the display gives the motor status: O - for OK, F - for fault, B - for backlash, A - for anchored ... See ledger on the panel itself for a complete list. The use of virtual servos is to give the tuning table value on the display when the virtual servo is selected. Just read the value and tune the device manually. So there is no need to write down any micrometer settings.

The four buttons to the left along the schematic incoming radiation line control the Martin-Puplett interferometers for SSB filtering and LO injection. The SSB filter should normally not be tuned, although it is OK to tune the LO diplexer Martin-Puplett interferometers. Just above and below the LO Diplexers are the Gunn buttons for the C and D band, respectively. These are virtual servos i.e. only there to give you the values for manually tune the Gunn. The backshort is the tuner in line with the output from the Gunn - i.e. horizontal on the panel. On the Gunn itself the backshort is pointing downwards but still in line with the output. The other button is for the frequency tuner on the Gunn. Further to the right on the top and bottom are the multiplier buttons. These are also virtual. All for buttons #1 - #4 are used fro both bands - the panel is from a time when it was thought that only two real servos would be needed to adjust on the multiplier. Nowadays the servos are virtual and four. The numbers #1 - #4 corresponds to the numbered micrometers on the multipliers. Following the multiplier buttons are the mixer tuners. These buttons control real servos. The first column controls the C band mixers and the second column the D band mixers. The servo in line with the input horn of the micro is the backshort tuner. The servo on the side of the mixers are the E-plan tuners. Note - THERE ARE NO E-PLAN TUNERS FOR D BAND. So selecting the top or bottom button in the second column is a null operation.

Trx monitor, IF control and chopper units

The Trx Monitor module is leftmost on the row of electronic modules below the tuning panel. The two analog meters shows the IF power or the S/N in channel A and B, respectively. You select between the option with the switch. The four red LEDs warns against to low or high IF power. Make sure you have the Total Power/Trx switch in the right position for what you are doing.

Next follows the IF control unit. With the rem/local toggle switch you can get local control but the micro can always regain control. In manual mode you can set the IF levels manually with the up/down switches. However, whenever the micro takes back control it will override your values. In manual mode the top center switch is important. Even if there are four mixers inside there are only two IF channels outside the dewar. This switch determines if your external IF electronics are connected to the C or D band mixers. Again the computer will take care of this but if you need to be in local mode this switch must be in the right position.

The three chopper controllers are all identical and controls the image, the sky and the load chopper, respectively. During S/N tuning you need to have the two rightmost chopper controllers in local mode. This is done by the rem/loc toggle switch. The sky chopper should then be set to the in position (receiver locking at calibration loads) while the load chopper should be in phase lock position (phk). This switches between the cold and ambient load in phase with an external signal from the Trx Monitor.

Mixer Control

The two rows with mixer control units each start with two mixer bias units. One for each channel. These units should normally be in computer control mode. A red LED will come on as a warning if the unit is in manual mode. The top LCD gives the mixer bias voltage in mV and the lower the mixer current in mA. Normal bias voltage values are around 2.2 mV for both bands. The C-band mixers are optimal around 0.016 mA while the D-band mixers need about 0.020 mA.

Next follows the HEMT bias supply units. These should be on. The selector switch allows you to measure the bias conditions on each of the three HEMT transistors with a DVM. There is a special cable in the W cabinet for the connector.

Finally, there are two coil drivers. The purpose of these is to generate a magnetic field to suppress the Josephson effect. The Josephson effect causes excess noise and IF power instability. The suppression is cyclic - there are several values that will cause a good suppression. But a to large magnetic field softens the knee on the IV curve and degrades the mixer performance. We usually have a current of 42-43 mA for the C band units and 60 - 100 mA for the D band units. The rate of change of the current is limited by the electronic so it is easy to get ahead when adjusting the magnetic field current. If the receiver warms up the coil drivers will trip and a red LED will light up. If the receivers is cold the supplies can be reseted with the reset switch.

LO Power modulator, temperature and vacuum gauges,

There is one module controlling two modulators. The top display and nob controls the C-band LO power modulator. The actual modulator is the black circular unit between the Gunn and the multipliers. This units should normally be in computer control mode but it might be necessary to be in manual control during tuning. A red LED warns you if you are in manual mode.

The temperature readout only gives approximate temperatures. The micro itself can give more accurate values based on a table.

The vacuum gauge should show less than 10**(-3) mbar. If not the receiver is warm, or warming up.

Oscilloscope

The oscilloscope is used to check the I-V curve and the Pwr-V curve. To use make sure the mixer bias units are in remote control. The turn on the seep switch to the left of the oscilloscope. You normally also have to increase the intensity of the beam. Then select the I-V or Pwr-V mode and mixer/channel. Note that for the Pwr-V mode only the channel that is selected on the IF control unit is displayed. The selector switch only selects between the channels. You can not look at the D band Pwr-V curve while tuning up the C band mixers. If you try you will only see a copy of the C band Pwr-V graph. Remember to turn down the intensity on the oscilloscope and to turn off the sweep before you leave.