This section contains status reports of current activities, milestones or problems from time to time.

last UPDATE May 23 2000:

The X-Y Stage for the errorsensor is going well and should be ready for delivery this week. Below are two images of the X-Y stage.

TOP VIEW

ISO VIEW

Mirror Post Polihing Activity May 8 2000

In this report we invoke the well known concept that a picture is worth a thousand words. So here are 6000 word equivalents.

Optician A. Camacho and Proj. Scientist C. Claver sucessfully complete alignment of M3 Mirrors with Surround prior to post Poishing.pix 2 (below): M3 Mirror Assy on Air Bearing, Metrology and Data Acquistion/Analysis. Mirror position determined within 0.5 microns.

pix 4: Optician G. Poczulp checks figure of small Error Sensor IMA Paraboloid Parent with interferometer prior to aspherizing.(below)

optician H. Yarborough (not shown) has worked figure and finish of lightweight nickel plated Tip-Tilt Mirror to near acceptable value.

---

Beam Splitter status Report:

As noted in the 2/24 status report below fabrication of 10 Beam Splitter substrates is proceeding in the NOAO Optical shop. The first 4 substrates will be used to provide the following beam splitters as part of the delivered instrument.

(1) Mirror with a hole

(2) reflection over transmission of 80/20)

(3) reflection over transmission of 98/2

(4) a Dichroic pass band that transmitts R and reflects V and I or if

that is not satisfactory two dichroics, one reflecting longward of

the R-filter the other Shortward. Potental vendors have been contacted.

The scientific rational of these four beam splitters are briefly described below and comments were solicited during the WTTM report given at the March WIYN Board Meeting.

The first beam splitter (1) is meant for maximum efficiency for the error sensor, while providing broad wavelength coverage to the science beam. It's use will be in field where only faint guide stars are available or when the fastest tilt correction rates are desired. It is limited by the fact that the guid star will not be centered in the science field and effects of limited isokinetic angle may be seen in the science field.

Beam splitters (2) and (3) are "grey" in that they have uniform characteristics across a broad wavelength range. These are useful for project requiring multi-color imaging in the science beam and where moderately bright (2) or very bright (3) guide stars are avaialble.

The dichroic beam splitter (4) is optimised for two color, Harris V and I, imaging in the science beam and maximum efficiency in the error sensor. More details on this Dichroic Beam Splitter at this link.

                   

Friday March 3 a successfull mini-design review was held for the WTTM Error Sensor design.

The review which was presented by Ron Price described the optical layout and methods of achieving the required Alignment and focus. Figure 1 identifies the optical components and their layout. Figure 2 shows the Error Sensor Mechanical structure. This assembly mounts on the X-Y stage. The X-Y stage mechanical/electrical design had previously been reviewed and fabrication of the stage has been started at Balfrey.

          Based on the WTTM Error Sensor design review the WTTM Project manger and Project Scientist have given the go ahead to proceed towards fabrication.

                   

The following status report was prepared 24 February 2000

          Project Management:

Schedule: The Project schedule has been adjusted to accomadate Spring 01 prime North galactic Pole dark time imaging. The IAS will be removed from WIYN about 1/8/01 and returned 3/5/01; intially this was scheduled for 4/6/01. This change looks like a realistic schedule.

Meetings: There are seven WIYN papers being presented at the upcoming SPIE conference on Astronomical Telescopes and Instrumentation 2000 to be held in Munich, March 27-31, 2000 among the presentations is "Four Quadrant Error Sensor which Yields Postion and Focus utilizing an Internal Mirrorlett Array" D. Vaughnn, C.F. Claver, M.H. Richardson, "Real time Linux and LabView as a control environment for the WIYN tip-tilt module" P.N Daly, C.F Claver NOAO. Other related papers connected with optimizing performance prior to the implementation of tip-tilt are "Optimizing the delivered image quality at the WIYN 3.5 meter telescope " D.G. Sawyer, C. Corson . A. Saha " Seeing experiments with the WIYN 3.5 m telescope" D. Blanco, EOS Technologies Inc., C. Corson/NOAO, D.G. Sawyer/WIYN and "Closed loop focus control at the WIYN 3.5 m telescope" D.G. Sawyer/WIYN, C. Corson, D. Mills/NOAO

          Optics:

The reimager optics M₁, M₂, M₃ off-axis reflectors were accepted for delivery for NOAO inspection and Q/C. As noted in the 1/00 Monthly Report a problem with the spherical mirrors M₁ and M₃ was found during in-house interferometric testing which showed a pinching at the corners of the mirrors. Apparently this was caused by stress due to the surround fixture which touched the mirrors and caused the pinch/stress. The fit of both surrounds has been releaved and the post polishing should remove the mirror distortions. The best way to mount the mirrors in the optical houseing is under study.

The M₂ mirror which was post polished at CMM showed on visual inspection and from Normarski microscopy that the post polishing was irregular, i.e. patchs showing surface roughness. Testing over a larger area of M₂ has been carried outat Zygo to quantify the surface roughness and surface figure in order to determine whether further polishing is required. Analysis of these measurements is underway.

Other optics include the tip-tilt miror. Three blanks have been prepared and ground in-house. Polishing of these mirrors is starting.

The paraboloid for the IMA (Internal Mirrorlettes Array) has been made but does is not curently acceptable. This is a problem receiving attention. Design of the IMA, Prism and other Error Sensor optics is complete

A mini design review was held for the Beam Splitter and the deliverables established. There will be 10 substrates made in the NOAO Optical shop. Work is already proceeding on these substrates. The following beam splitters, which must be change manually will be (1) Mirror with a hole, (2) reflection over transmission of 80/20. (3) 98/2 and (4) a Dichroic pass band that transmitts R and reflects V and I or if that is not satisfactory two dichroics, one reflecting longward of the R-filter the other Shortward. Potental vendors have been contacted.

          Error Sensor:

The X/Y stage received in December from Dyna-Optics proved to be unacceptable. The Project selected alternative vendor, Balfry Precison. After a site visit and careful review of the RFQ it has been determined that the mechanical design of the X/Y stage will meet specs and conform to the required space envelope. The electrical control system is still under evaluation but will probably prove acceptable.

The PI Tip-Tilt mirror PZT assymbly and the PI servo system has been delivered and is undergoing test. The as delivered response is not as fast as we had expected, however, adjustments being carried out in house to optimize response have resulted in improvement..

As of this date the APD modules have not been delivered from EG&G. The project is prepared to preceed with testing as soon as they are delivered. The mechanical layout of the Error Sensor is nearly complete and the software interface is well on schedule.

          Mechanical:

The Optical Housing design is essentially complete and detailing for shop drawings has started. The lathe assymby process and drawings is in progress and a decision on vendor or in-house fabrication is pending. This will entail visitations to possible vendors and review of in-house resources. The NOAO shop has been providing testing and assymbly fixtures as required. Mechanical work is on schedule.

          Software:

The following diagram shows the present status of the WTTM control loop. As can be seen the higher-level LabVIEW environment and the real time Linux core communicate through two Fifos and a common (shared memory structure) protocol. The real-time Fifo handler accepts both commands to stop and start tasks and reset parameters in a global memory buffer.

The acqTask samples the APD modules as some specified rate (or awaiting some interrupt to say that data is ready). Once acquired, 3 error corrections are derived and the data along with a time stamp are written to a (large) shared memory buffer. This buffer can hold nearly two million such data sets. Nothing, except the user, ever stops the data acquisition task as it has the highest priority in the system. The acqTask also counts datasets collected and issues a semaphore to the memTask when the required number of points are in memory.

Surely even this much memory (128 Mb) this is not enough? At the highest acquisition speeds the entire memory segment can be filled in 176 seconds! No. After the acquisition resets the memory pointer to the start of the segment, it will continue and overwrite previous data. This would be a bad idea if the ttmTask was generating position demands very slowly (i.e. slower than 0.006Hz) but we expect the tip-tilt mirror to be updated much faster than that. The ttmTask is autonomous and can be run at frequencies up to the sampling rate of the APDs although it will always lose out to the acquisition task when re-scheduling at competitive frequencies. The ttmTask reads the data from memory and issues demands to the mirror controller via a digitial I/O card.

The memTask sends this information along another Fifo back to the LabVIEW environment which reads the data into a suitable array for display as a fast Fourier transform, Lomb periodogram etc. LabVIEW is able to filter this data and send low-pass signal to the telescope via the generic WIYN client software. This part of the work is not yet fully integrated into the system. What is not shown in this diagram is a high-level LabVIEW status task that polls the PI-controller for status information to be fed-back to the user.

WTTM Software Map

This screenshot shows the present design of the WIYN error sensor LabVIEW interface. As can be seen there are three panels explained below.

The Top Panel is for Status

Although some values may be changed they should not be altered without knowledge of the effects of these widgets. The top line of boolean buttons should, in normal operation, be green indicating that the port, the stage and the sensor are all at nominal integrity. The messages bar contains status messages about the state of any command issued to the Delta Tau controller. The serial line cluster indicates the port, baud rate, number of data bits per word, number of stop bits and the parity of the serial communications. The default values should suffice although in the event of a failure at night the cable can be switched to another serial port and these values adjusted accordingly. The miscellaneous cluster contains information on various items as shown in the table. They should not be altered. The X/Y Deadbands chart relays information from a background status monitor on the present position of the X-Y stage with respect to the acceptable deadband. At present it is up to the telescope operator or observer to take appropriate action if the stage exceeds the deadband. Work, however, is afoot to automatically re-adjust the sensor when it exceeds the limits - this is the use of the LOCK button in the lower panel.

The Middle Panel is for Position

The demand motor position can be set with the ring buffer. The options are 0:null, 1:motor co-ordinates, 2:beam-splitter hole, 3:array centre, 4:CCD co-ordinates. The X and Y position boxes appear and disappear as required for input i.e. you should only see them as active when using position slector 1 or 4. To move the error sensor, click the GO button once. To abort a motion, click the ABORT button once.

The Bottom Panel is for Algorithm

The centroid algorithm can be set with the ring buffer. The options are 0:5-point search, 1:peak=pixel, 2:none. Clearly, if "none" is selected the error sensor simply accepts the demand position. If "Peak-Pixel" is selected, the error sensor reads the APDs once and determines the proper centroid position and moves there. The 5-point search moves the error sensors to five positions (centre, right, left,up and down) and takes data at each position. Once complete it can evaluate the correct centroid and move there. More important, it can evaluate a sky subtraction value from the outside pair of elements at the extreme positions. This screenshot shows the error sensor patter overlayed on a 0.8 arc second image on the array. The derived sky value can be fed back into the tip-tilt control loop for optimal signal-to-noise sampling. This work is ongoing.