Action Items
From NOAOWiki
Torrent Action Items
Return to the Torrent home page
Active action items by person responsible
David Sprayberry (D2)
From Torrent Meeting 20110930
A discussion followed concerning the detector, Dewar, controller integration schedule. 1st integration package for KOSMOS starts 1st November. At that point science grade can be delivered on test Dewar. David asked if the two Dewars will be required to complete the work. Ron replied that having two Dewars will reduce the number of times the detectors will need to be swapped out and so expedite work and enhance safety.
ACTION: David S. is to make a decision on 17th October as to whether or not to build up the second
"green' Dewar for integration work.
It was reported that the one missing part is currently in the mechanical workshop awaiting a cost estimation. The plan is to build several of these parts to be used on other Dewars in the future.
ACTION: This week, cost to be considered and an order placed to either proceed with one
or several parts to be made to the workshop. See also action items for Ron.
Dave Sawyer (D3)
Nothing Pending !
Dee
From Torrent Meeting 20111019
ACTION: Dee is responsible for the procurement of the Torrent kits. David and
Peter (and everyone) will help out with whatever is necessary for Dee
to complete this task.
From Torrent Meeting 20111109
ACTION: Dee is to contact the indicated person and ask for a quote to perform the incoming
inspection on 40 AFE boards.
Guillermo
Nothing Pending !
Kathie
Nothing Pending !
Nick
ACTION: Nick to pass document source for the assimilate (and sysConfig) documentation to
Dee - The document keeper.
Peter
From Torrent Meeting 20111109
ACTION: Peter to finish off the temperature servo firmware in time for Thanksgiving.
Ron
From Torrent Meeting 20110930
It was reported that the one missing part is currently in the mechanical workshop awaiting a cost estimation. The plan is to build several of these parts to be used on other Dewars in the future.
ACTION: This week, cost to be considered and an order placed to either proceed with one
or several parts to be made to the workshop.
ON hold until further notice. R.G. 10/26/11
From Torrent Meeting 20111102
ACTION: There was discussion about the heat exchanger fan after Dee raised the issue.
Ron and Peter have decided that an AC fan may be the best solution.
Ron is to spec and order a couple of AC fans to try out in this application.
Torrent Meeting Minutes
Return to the Torrent home page
Torrent Meeting minutes 20110615
1. Stock inventory: Kathie currently working on producing two of each cable assembly, ECO work, etc. Some (rough) ideas about how much material is in hand and what is available. Postponed any decisions about how many systems need to be assembled until more information is available on next Fridays meeting.
Action: Dee and Kathie to produce account of stock in hand, in production, and with estimates on time required to complete tasks on ECO's. cable building etc. This to allow us to proceed with the plan to build as many Torrent Kits as possible in an efficient manner.
Action: Peter to send last PSM ECO asap to complete the required modification set on this module.
2. Inspection of rev -C- AFE boards: Preliminary inspection carried out last Friday by Ron. Boards are functional with regards to bias setting, video offsets, and clocks but seems to have telemetry channel assignment changes and constant (mid-range) data value returned from the video ADC channels.
Action: No action until Ron has finished with the test system hardware details and/or Mark returns.
3. Test system triumphs: Work on mechanical, electrical, and software aspects continue. Estimates for start of testing the test set are for early July.
Action: Ron and Dee given green light to continue full steam on this project. Nick to continue working hard to be ready for hardware availability.
Action: Peter to send current firmware version 2.07 with Mezzanine board power enable register to allow integration into test system software.
Action: Peter to check that Vana+/- limits reflect the ECO that allow operation up to +/-12v.
4. Performance test schedule: The impact of Rons (approved) holidays on the proposed schedule was discussed. It was decided that there wasn't too much impact and that some of the work planned to be shared by Ron and Peter together during that time should already have been completed before the arrival of Peter in Tucson which reduces the work load on this point.
5. Software production: Nick submitted a report that shows that all major aspects of the operational software required to put Torrent into the observatory are available but that these software products require testing before they can be judged fit for service.
Action: None at present; wait for first test set to be operational and an opportunity to access real (and hopefully production ready) Torrent DHE hardware.
6. Documentation: Dee has received a detailed handover package from Roy and has a list of incomplete documentation that requires finishing.
Action: None .. lower priority and await opportunity to continue with documentation effort.
7. Configuration control: Effort has begun with meeting to define the inputs and products of configuration control and hopefully allow implementation before Torrent hardware deployment. Ron and Dee are to meet tomorrow to discuss workshop area and layout to allow Torrent kits to be born and natured in the Tucson labs.
Action: Everyone is invited to submit Torrent application requirements scavenge parameters to Nick who will produce the first online configuration control scheme. These parameters will be used to generate a set of documenattion and work orders that will define each Torrent system.
8. Dee asked if deware connectors should be part of the Torrent master parts list. We define these connectors as those used to connect the Torrent TSM module to the Dewar vacuum connectors.
Action: Ron to pass part number to Dee for ordering five sets of parts for the five Torrents assigned to KPNO projects.
Action: Peter to pass part number to Dee for the SOAR Dewar connectors.
Action: Ron to check for availability of back nuts for the KPNO and CTIO SOAR connectors.
9. Dee asked about the availability of the GIGe port module. Peter stated that the capability for using Torrent with an Ethernet interface is in the requirement documents and as such we must test this functionality. We decided to use the material we have to do the testing at this time.
Action: Peter to pass part number to Dee for the flex cable part required for testing.
Action: Dee will try to purchase enough of these flex cables for the production run.
10. Ron asked for clarification on: a). What was to be the paradigm for testing the two systems during performance testing ? answer - combining our efforts on testing both systems to the same level. When work required to fix a problem, one person to concentrate on this, generate an ECO, and then the second system to be brought to the same level for verification.
b). How much hardware and what type of hardware will be required for the performance testing to be carried out ? answer - Not much and mainly small pieces of jury rigged hardware to allow routing of signals, measurement points, etc.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110617
1. We set the number of Torrent kits to be produced on 2011 budget at 20 units. This means we are required to procure enough hardware materials to assemble 20 complete Torrent systems.
Action: David to see what the funds are that remain of the 2011 budget for material purchase.
Completed: All original funds are intact and David is identifying and collecting the funds under one budget in preparation for parts procurement.
Action: Dee to complete stock take to asses what is required to build 20 kits.
2. Dee has furnished a list of materials and processes that need to be applied to the materials for procurement. This serves as an important foundation to the process of production planning.
Action: Dee and Ron to continue this work to setup a production workshop environment suitable for the production of Torrent.
3. A raw list of parameters that will be needed to define a new Torrent application was generated by brainstorming around the table. This list is reproduced at the end of this report. Apart from enlarging and/or optimizing the list, work must now be started to define which parameters that we collect are used to define the products we require to build a Torrent system e.g. what collected parameters are used to define a work task to build the TSM module, etc.
Action: Nick to organize the list and begin (after test system work priority) to organize into the 'data base' needed to generate the Torrent build process.
Action: All, to further define the list of parameters.
Action: Dee, identify the official name for this document so that common language can be used to identify it.
4. A short report on progress for the work task was presented:
Kathie - Continues to work on applying upgrades to the in-house hardware material. It was decided that when parts needed to be removed from stock and/or when the work was more efficient to be done in serial fashon, all modules in-house would be modified to the same standard.
Ron - Finishing off the hardware and mechanical aspects of the test system. Will continue to define test procedures for the test set with Nick and cooperate with Dee to setup the manufacturing area for Torrent kits.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110622
1. The 2010/2011 FY budget has funds to allow us to purchase the hardware to make 20 Torrent kits so we agree to make 20 Torrent kits in the first batch. Dee reports that the procurement time for the major hardware parts is approximately four weeks which includes incoming inspection. Dee also reports that the documentation is close to being up to date for the purposes of ordering. We need a deadline date that tells us when we need the hardware parts in our hands to build the kits. We would like to allow Mark to contribute to this decision so allowing for a decision to be made at the Friday 1st July meeting would make hardware available for kitting by early August - while performance tests are being carried out. These decisions were adopted i.e. build enough hardware for twenty kits, make an informed decision to go ahead with purchase at the July 1st meeting.
Action: Dee to attempt to finish the details of the list of materials that will be required to be ordered. Action: All we postpone the AFE production until we have assurance that the hardware is functional (end of July).
2. Change orders on the power supplies seems likely to be finished except for two outstanding boards. ECO work is to progress until complete. We need to make sure we have at least one, ideally two, AFE Rev -B- hardware boards ready for Mark when he returns.
Action: Ron to make sure that the Rev -B- AFE hardware boards go through incoming inspection ready for June 27th.
3. We discussed whether to send the production hardware at NOAO South back to Tucson for ECO work or to attempt to carry out the work here. It was decided to send the hardware (one LCB Rev -A- + two PSM Rev -A- + one TSM_UTIL rev -A- + one Vcb Flex cable + one Video Flex cable) back to Tucson as early as possible but without hindering the firmware work that currently requires this hardware in the south.
Action: Send south production hardware back to Tucson asap.
4. A decision was made to officially call the worked up document presented by Nick as the "Torrent System Parameter User Survey" document. The contents and purpose of the document was lightly discussed and the definition of the 'requirement' and 'goal' levels of compliance was clarified. The list is sufficiently complete to allow further work on the configuration control process to proceed.
Action: Dee and Nick - To (begin to) identify the products that need to be derived from the "Torrent System Parameter User Survey" that allow work tasks to be generated and responsibility for the work assigned.
5. The Mezzanine test routines where lightly discussed as presented by Ron. Nick will know turn these into software tasks for the test system.
Action: Ron, Nick to include LCB_MEZZ current telemetry tests to pass / fail abnormal telemetry scaling values. This requires a small addition to the test suite.
6. Nick reports that during the process to 'assimilate' firmware version 2.07 several of the GUI attribute names do not fit (match) the previous attribute mapping.
Action: Peter to change Telemetry attribute names by removing the 'V' in the name to match 2.04 GUI names.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110706
Wednesdays meeting agenda had only one item: Formation of a plan to take the Torrent project from Development status to Production status. To help in the determination the concept of 'Production status' was defined as:
x When we have Torrent Kits(*) on the shelf x When all hardware material in the Torrent Kits are at the latest revision level and tested. x New torrent material can be ordered from existing documentation x The hardware / software / firmware components of Torrent are under configuration control (**)
(*) Torrent kits are defined as generic kits of parts that are used to 'build a system' that is for a specific application. (**) Configuration control is defined as the book keeping that allows you to know what you have in stock, what systems have been build, where they are deployed, and maintain records of revision level compliance.
To achieve the results from the agenda item we defined three steps that needed to be completed at the meeting:
1. Define what conditions need to be met to meet production requirements.
2. Define a list of what we need build or things that need to be done to get to the conditions
defined in step 1.
3. Assign people to do what will be required in item 2
For item 1 (conditions that need to be met to meet production requirements) we defined:
1a. Functional test system needs to be in place 1b. Design verification needs to be successfully completed 1c. Support documentation needs to be available (at least for tier 1 and tier 2 support). 1d. Configuration control process in place 1e. Hardware / Software / Firmware available at the correct revision level 1f. Readiness milestone approved and signed off.
For item 2 (what we need build or things that need to be done to get to the production status) we resolved that for each item 1:
2a.i. PSM, MEZ, AFE, LCB, TSM_PRE, test fixtures need to be operational
2a.ii. PSM, MEZ, AFE, LCB, TSM_PRE, test software need to be operational
2a.iii. Test system verification needs to have occurred
2a.iv. Test system documentation needs to be available
2b.i. Test procedure documentation needs to be finalized
2b.ii. 2 complete Torrent systems need to be built from available Torrent kits (CHILE1, CHILE2)
and shipped to Chile
2b.iii. CCD in a test dewar needs to be available and shipped to Chile
2b.iv. Test fixture hardware (and system wiring) needs to be completed
2b.v. Software for auto-configuration (eeprom system) needs to be finished and tested
2b.vi. System configuration needs to be built
2c.i. Quick start guide (est. 90 % complete)
2c.ii. Operations manual (est. 50% complete)
2c.iii. Technical manual (est. 50% complete)
2d.i. Dee needs to be consulted for a more comprehensive overview and progress status evaluation
on this item.
2e.i. Prototype mechanical heat exchanger is near completion,
Afe board needs respin but kludge will be used for 1b.
2f.i. A report is to be generated and passed to first level users (KPNO / CTIO operations
and instrumentation group heads)
For item 3 (Assign people to do what will be required):
2a.i. Dee + Ron 2a.ii. Nick 2a.iii. Not specified (*) 2a.iv. Not specified (*) 2b.i. Peter 2b.ii. Mark 2b.iii. Ron + Mark 2b.iv. Peter 2b.v. Nick + Peter 2b.vi. Peter + Nick 2c.i. Peter + Dee 2c.ii. Peter + Dee 2c.iii. Peter + Dee 2d.i. Dee + Ron + Nick 2e.i. Mark + Ed + MT + ET 2f.i. Peter + Mark + Ron
The reference time frame for completing this work i.e. to transition from Torrent project development status to Torrent project production status is two Months .. Work to be completed by September 2nd
ACTION: Dave Sawyer to update detector group schedule - requires input from Mark and Dee to complete. ACTION: Mark to be responsible for the project management and leading the group to achieve the change
in project status
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110708
Item 1. A decision was made to down size the number of Torrent Kits to be prepared from the value 20 kits to a value of 4 kits. This complicates the configuration control process but reduces the time /resources that will be required to prepare the two systems for Chile and two systems for Tucson. 2. The two systems for Chile will be called CHILE1 and CHILE2. These will be built from 'virgin' Torrent kits. Both systems will be used for design verification testing then one returned to Tucson for use as the 'golden standard' required to complete 'Test system verification' (item 2a.iii. above). 3. To achieve this value of 4 kits, 4 more AFE boards need to be modified - Kathie to complete this work next week. 4. Shutter cables are to be fabricated in series (4 units) as soon as parts are available. 5. An accounting on flex cable assemblies need to be done and a total of 4 per type assembled. 6. Peter emphasized that the configuration control process _must_ be initiated now at the initialization of these activities. He further stressed that all of the possible and available methods of configuration must be used to assure the correct initialization of these records. The key (at least to Peters mind) to effective configuration control at this point is to have the Torrent Kits be the point at which all material is put under configuration control. This may be achieved at the Tucson labs by using a single key access system to access the material being stored as Torrent Kits in the IR Laboratory. The current hardware that is on the bench in Tucson must be returned to these storage locations as kits and then released under the appropriate documentation and record keeping.
ACTION: Mark to re-assign Torrent hardware to a system serial number (using the START form). This refers to the 1 1/2 system we have outside control right now. ACTION: Dee to assign the material (from above and from stock) to build four systems as CHILE1, CHILE2, TUCSON1, TUCSON2 ACTION: Peter to fill out 'user survey' to guide the building of the CHILE1 and CHILE2 systems.
7. A discussion was held as to the suitability of the current E2V system configuration (for KOSMOS) as the design verification test system configuration. The main restriction is the fact that the current KOSMOS configuration only uses one AFE board. However, the test Dewar will not have two detectors available for it and so there is a shortage of detector amplifiers to fully test the AFE hardware. ACTION: Peter to evaluate the requirement and build, if necessary, a specific verification configuration.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110713
1. Dee reported that the efforts to bring material stock levels up to the requiremnt to build four Torrent kits continue. At this time 1 1/2 kits worth of material is available. The TSM documentation package is complete (Allowing a TSM to be kitted and built), the Controller package needs to be completed.
2. A discussion was held concerning the termination method for the shutter cable to the TSM-UTIL board signals. Two methods were suggested, the use of a small 0.1" 10-pin header wired to the Shutter connector via flat ribbon cable, and the use of an intermediate Winchester miniture connector using discrete wires. These will be evaluated in the lab to select the correct method.
ACTION: Kathie, Ron to evaluate the termination method of the shutter cable.
3. Kathie asked if what needs to be done to finish off the utility boards. The ECO's for the TSM_UTIL board need to be finished off and released.
ACTION: Dee to incorporate the the final details into the 'last' ECO for the TSM_UTIL board (i.e.
removal of C35 and the decision (or not) to remove the 0.1" header pins if the discrete
wire method is used for the shutter wiring termination)
4. Dee reported that there are currently four TSM parts in the production and deployment work flow.
5. Dee asked for confirmation concerning the work flow of the deployment process. It was determined that the assigned Detector group engineer would take the 'user survey' and progress the paperwork and design to the 'work order' document stage. At that time the Detector group manager needs to allocate resoures for the dispersement of the work tasks
6. Peter asked that the design for the prototype heat exchanger be tweaked to allow better efficiency before it is manufactured.
ACTION: Mark to identify resources to do this work.
7. Peter asked that in addition, a small, commercially available heat exchanger be purchased and submitted to evaluation for the case where a heat extraction plenum needs to be employed. This stems from work done by Guillermo who identified the commercial product.
ACTION: Guillermo to order the heat exchanger item and help in evaluating the performance.
8. Peter asked for clarification as to what power source the prototype heat exchanger should use for the fan. It was decided that the fan would be run from 110v AC power.
9. Ron reported on the power supply testing progress. Three PSM assemblies have passed testing. All the faults that had been reported in the email sent to Peter about four weeks ago are now resolved.
10. However, Ron reports of new problems related to component falures caused by the wrong part being installed during manufacture. The FET located in the feedback path of the negative Vcb rail have been discovered damaged in two PSM assemblies.
11. A short discussion was held to determine if the PSM pcb layout needs tweaking to cover the rework required in the ECOs for the PSM.
ACTION: Mark to assess whether to incorporate the three small layout changes required by
the PSM ECOs.
ACTION: Mark to produce an order for a PSM pcb board production run and the manufacture of four
more PSM assemblies through CIREX (not our previous vendor).
12. Ron gave a brief status report on the Test Station design. Board layout review in progress. Soon to go for production.
13. Nick reported that the Test Staion software is complete and awaiting hardware to test it on.
14. Nick currently working on getting the Configuration toolset software running
15. It was decided to use the generic calibration values available in the vhdl source code as a test set of values to load to the four systems currently in production / deployment.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110715
1. The actions list was evaluated and several work tasks cleared from the active list. Excellent work by everybody in their diligence to take care of the actions that they are responsible for. A few new details were discovered in the process of the action item status evaluation that generated new actions.
ACTION: Dee to evaluate availability and cost of stainless steel connector locking nuts for the
Dewar connectors. The bare aluminum versions bind to the connector barrel. Either stainless
steel _or_ anodized aluminum nuts need to be found.
ACTION: Ron & Dee. Evaluate the price and suitability of the chassis mount connector types which
incorporate a locking nut in the hardware kit as a possible replacemnt of the current
Dewar connector types.
ACTION: Ron and Dee (?): Move the assembly benches around in the Torrent production lab space
to provide more horizontal work space.
2. Ron gave a short status update on the PSM / MEZ Test system hardware. pcb in layout review and ready to be sent out. In house assembly contemplated so that the hardware can be evaluated and the software tried out at the end of August.
3. It was decided to give the 'non-Torrent' system exclusively to Nick to allow final testing of the 'Collector' software before design verification starts. This hardware is that which has been occupied in recent AFE functional testing and Firmware development. It was due to be disassembled and put back to stock as part of the Torrent Kits. However, the looming deadline for the design verification (of which this software is required) over-rides the configuration control exercise. The hardware will be turned over to Dee as soon as the first Tucson test system is delivered from stock.
ACTION: Nick to concentrate on testing and perfecting the collector software sub-system.
4. Nick asked that we clarify the where and how we document system features (such as system gain, shutter drive circuit, temperature sensor configuration, etc.) Nick asserts that this information needs tobe included in the system configuration data collected and organized by the sysConfig tool. To determine these points, a special (and short) meeting has been called for Monday 18th at 10:00 MST in the B2 lab. All those with a vested interest are invited to attend.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110720
No minutes recorded
Torrent Meeting minutes 20110722
No minutes recorded
Torrent Meeting minutes 20110727
No minutes recorded
Torrent Meeting minutes 20110729
No minutes recorded
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110803
1. A short discussion between Peter and Nick was held concerning the OS version installed on the PAN named "farlie'. It was concluded that the La Serena systems administrators will need to become involved to solve the OS incompatibility issue.
ACTION: Peter to seek help in upgrading the farlie machine to Centos 5.3
2. Dee initiated a discussion to determine the correct vocabulary to be used to identify a Torrent system that needs to be built and configured from a Torrent kit. Various names have been used to describe this identifier such as 'system name', 'focal plane name', 'generic name', 'project name', etc. Nick explained that the correct identifier to be used is the 'focal plane' name. This identifying name appears in the top section of the Torrent User Survey document (i.e. is filled out by the client) and thus serves to identify all aspects of the Torrent that are used to build the delivered Torrent system to the client. Mark et. al. asked if this can be easily changed when the system is re-assigned. Nick answered that the idea is not to change the characteristics of a focal plane system but to regenerate another focal plane name if the Torrent is re-assigned. It was suggested that more appropriate names (i.e. to follow the conventions used on Kitt Peak) for the focal plane will come into use as the systems become deployed.
ACTION: To All staff - The unique identifier for a Torrent System shall be the 'Focal Plane' name.
3. Dee asked that we agree also on the terminology to be used fro the state of hardware modules and system builds for the purpose of material tracking and configuration control. This is also in the same spirit as the previous point i.e. to normalize the vocabulary between the detector group members. after some discussion it was agreed that Dee (and maybe Kathie) would create a list of 'acceptable' vocabulary to circulate to the group.
ACTION: Dee (and maybe Kathie) to generate a list of words (a vocabulary) to describe the parts,
modules and status of the hardware for Torrent systems.
4. With special reference to the status of different parts of the Torrent system, Dee asked that we agree on the vocabulary to identify the specific status of hardware to allow configuration control documentation to be unambiguous. It was suggested that the status words, once defined, be 'hard wired' into drop down boxes on the Excel spreadsheet used for materials tracking.
5. A request to modify the Torrent TSM hardware by removing the bottom cross piece was presented to the table. The table agreed that it was a worthwhile modification to allow a much simpler build up of the TSM Preamp to Dewar Connector plate wiring. It was agreed to modify the 18 pieces in Tucson to accommodate the concept.
ACTION: Dee (once again) is to draw up the modifications and get approval from Mark to go ahead
an carry out the mods. This work to be carried out by the expected Torrent Deployment
Readiness (TDR) date of September 30th 2011.
ACTION: Someone (not specified at the meeting) must pass these modifications to a mechanical
draftsman to have the source documents updated.
6. A brief update was requested by Mark of Torrent activities:
Peter stated that the preparations being carried out for the Design Verification tests are progressing well and only details remain (which include sequencer code building, PAN configuration details). The test zone is setup and ready with tools and supplies.
Kathie reported that the building of Tucson1 and 2 is progressing with 1 preamp completed from existing parts and 1 utility board in progress.
Guillermo reported that a recent utility board build was tested noisy (unstable) which was probably due to water incursion into the setup potentiometers during the wash-down phase of manufacture. This board has since been vacuum baked is since less noisy. In addition it is unclear as to whether the gain potentiometers have sufficient scope to achieve the calibration values required. Mark suggests that this may be a tolerance buildup issue.
Nick reports that the collector pgm is functional (used to collect the calibration constants from the DHE hardware during PAN software boot). This program currently takes over 5 minutes to extract the data but so far no clues as to where this (excessive) time is being taken. It was suggested that the same problem is probably responsible for the slow response of the borg. Nick suspects that this may be true and that the problem may be in the way that python handles socket traffic. Nick suggested a work around to this problem. This would entail reading just the first two pages from each hardware module eeprom to determine if there are differences between the eeprom contents and the eeprom mirror files held on the PAN disk. This seems like a good approach that can work although the fix to the original problem (of slow communications) is preferable.
Nick also reports that more work is required on the sysConfig tool to include new features (TSM thermocouple wiring, etc.), to add connectors and facilitate multiple connector assignments.
Nick also reported that the wiring list printouts from this tool now contain added information to assist in the TSM construction.
7. Peter asked about the progress on the construction of the new Torrent heat exchanger. Mark stated that it has been sent out for construction and should be in house in two weeks time. Unfortunately this means that the thermal testing section of the Design verification tests will need to be carried out in Tucson by Ron on his return.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110810
A bit late...
This meeting had only to reduced Tucson contingent as Peter & Ron were hard at it in the lab.
I gave an update on the Tucson1 & 2 controllers - Tucson2 is delivered to Nick and is the new software development system. The old system was removed for recycling.
Tucson1 is in testing in the B2 lab. I was seeing flaky performance on the clocks and that was later confirmed by what is happening in Chile.
Dee and Nick went over more of the documents and who does what. The biggest thing is that they are not all complete, which is why I couldn't figure out how to get my paperwork done! Progress is ongoing and Nick is pretty much ready for beta tests of his sysconfig which includes more of the configuration for the TSM.
No real action items from this meeting.
Return to the top of the page Action Items#top
Torrent meeting minutes 20110813
1. A status update from the south: Not as much progress as one would have wished. Principle problem is getting a reliable power on reset signal to the DACs on the afe boards. This function is supposed to be inherent in the DAC devices themselves. However, due to current leakage through the AFE interface registers while the afe is powered down, the AFE 3.3v supply rail floats to approx. 1.9v. This is obviously above the power down threshold of the DACS and inhibits the on chip power on reset function. We tried modifying the firmware to use the AFE Digital Output Enable signal (/AFE_DIGITAL_OE) to place the AFE interface registers into high impedance mode when the AFE is powered down. This did not succeed since any leakage from the VPIFC side of the registers to the outputs drags the OE signal low (true). More detailed reading of the register data sheet shows that this is 'normal'. We then attempted to use a specific power on reset signal driven from the firmware and applied to the DACs when the AFE is powered on. This works but shows unreliability. The DAC reset signal is consistent before the latches but the output of the AFE register shows variable pulse width and sometimes missing pulses. We realize that these two DHE systems that we are testing are the first to be fully populated with production class hardware and the first systems to be complete. Therefore we expect to see problems that have been otherwise hidden by the less complex prototype systems. Testing has also revealed a problem with the setting of the clock register. The data values written to the clock regisetr do not always appear on the output of the AFE interface register. Nick and Mark confirm that this is seen on the third production DHE on test in Tucson. Approx. 18% of the design verification tests have been performed to date. These tests are mainly those that do not require the system to be functional to reply to.
2. Nick reports detecting a failure to write individual pages of an eeprom. The read back of these pages shows the normal erased condition (0xFF values) so it is thought that the eeprom failed to write the page correctly. Peter suggested that the reliability of the eeproms is probably better than the reliability of our code and that the problem may be caused by the firmware missing a write command from the PAN. Nick is going to implement code to read back and verify the eeprom data to eliminate this problem.
ACTION: Nick to implement the code to verify eeprom writes are successful.
3. Nick reports that the slowness of the borg response, previously thought to have been caused by the socket interface logic has been improved. The socket interface has been improved to provide a 14 X increase in through put. This allows much better updates to be seen on the borg interface. It also seems to have eliminated the lockup problem with the interface when the update button was double clicked.
ACTION: Nick to update the borg software on -farlie- and -ctiola- PAN machines to improve
performance.
4. Last note from Peter. Ricardo and Rodrigo have provided infrastructure to pump and cool the test Dewar here at ctio. This will hopefully be taking place early next week when we have resolved the problems with the DHE hardware and firmware.
Post Script: As of Sunday evening (20110814): Hardware and firmware fixes for the DAC power on reset have been implemented on both systems and so far, appear have resolved the aforementioned problems. Design verification testing has resumed. Tests were performed on Sunday to determine sequencer functionality and performance.
Torrent Meeting minutes 20110817
Present: Mark, Ron, Peter, Nick, Guillermo, D2.
1. Discussion of Video input tests and gain VS dwell. No conclusions have been drawn yet. Gain is off by a factor of 16 from what we expect. Dwell does not change the Gain. May be test setup.
2. Discussion of acceptance test and Ron's geographical location. D2 asked about estimate of time line to complete testing. Peter states about two weeks or so. This includes enviro testing in Tucson.
3. D2 asked: Is 98% good enough for DAC load? No it isn't per Peter. Must be corrected and is a part of the respin of the AFE. David discussed the layout changes and the more being change the more to be verified. Peter confirmed. Peter explained that it is really a function of just one signal. Currently is a passive pullup and needs to be driven in a strong manner to keep the part in tri-state mode. Actual problem is stated in the data sheet. Peter states the "surgery" is minimal and verification is relatively simple. Mark says it is a simple fix, Peter agrees. Signal is AFE_digital_OE. Peter described firmware fallback. Currently at 97% of slices occupied in FPGA. FW fix much more difficult.
4. D2 states that he wants an updated test doc at the end of the week before Ron leaves.
5. Nick – problem with sysconfig is incompatible file versions. He will update new stuff today. Update should fix both configs for Chile1 and Chile2. Peter asked if templates will have changes as well and requested the whole thing be updated. Nick says both configs have everything they need to run sysconfig and get the right answers. Also will update high speed borg. EEPstore – Nick fixed the other problems but still has two issues problem when writing To clock state reg (all F's) and unsigned integer issue. Nick described the problem in detail. Problem with data type conversion using cast into long integer. Has a fix but is reluctant to try it. The change would have to change every single conversion and there may be unforeseen side effects. Peter suggested that if it is only in one place the value could be tested before cast and specified as a constant rather be converted. EG test the float to see if it is 8f's and make it a constant. Nick says it's not just one single value, any number with msb =1 is a problem. Peter agrees, calculations could just be ignored when MSB set (sign bit). Nick says no guarantee of word size. Further discussion about having to deal with it if we go 64 bit. Only part of eepstore is crc calc because of the high bit problem.
Peter asked about changes and effect of format of eeproms. Nick says it will be the same unless we change addresses in FW. Nick explains further. Ver. 208 will have to be checked. Peter says it will be a good test of the process. Nick says it should show up. Peter states one assumes that by august end there should be one final version of firmware. Nick agrees unless we find a bug. This is to be at the end of the design validation. Peter asked if we should go to ver.208. Nick and Mark say go to 208 as the final version. Peter says we'll designate the first production FW as ver. 2.1.
Nick says all report files come out as .pdf and postscript format. Dee will be happy. Peter mentioned one detail that the video pins are not in the config output showing where they are connected. Nick says the reason is that they weren't connected. Nick will look to see if they are connected. Peter says they did get wired so it must have been in there at some point. Nick and Marks think they were wired from an spreadsheet version of this. Guillermo says there was a doc that Peter left for them. Nick will check to be sure. Nick says videos should be on the printout if they are connected. Guillermo says they were sent as email attachments early on. Guillermo to forward these to Nick.
6. Peter asks: Directors blog about being sick of the non-monsoon monsoon. Mark says it's the weather. Been hot and muggy so probably no reference to us. D2 says it was in reference to the weather also. Ron stated "maybe I will change my flight". Nick spoke about a crazy ride home in the hail.
7. No other business discussed.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110817
Present: Mark, Nick, Guillermo, Kathie, David Sprayberry, Dave Sawyer, Peter.
1. Status summary from the south concerning the design verification testing.
58% of tests are complete. Confidence that the remaining tests can be completed before the end of August.
Measured noise levels exceed the specification by a factor of approx. 2.3 times. Concern about meeting the requirements of KOSMOS. Peter explained that the noise figure can be improved without major rework through optimizing some of the video processor component values.
From experience during this testing phase it is shown that the system is reliable and repeatable - The DHE is described as being a 'happy chappy'.
2. Tucson status report: Efforts continue to prepare hardware and take a count of available stock in preparation of building the 20 kits.
3. A question was raised by Peter as to the advantages of acquiring some spare hardware boards. In the currently envisaged maintenance model, individual spare boards could prevent the age old custom of swapping boards between different controllers that would cause a configuration control nightmare. David replied that the current lot of 20 kits provides for 2 full spares plus 2 'development' systems per hemisphere that should be enough to handle immediate and medium term needs. Davis also explained that the procurement of Torrent controllers should be viewed as a process rather than an order and forget procurement exercise. That is that the procurement of hardware parts be optimized in lot sizes that maximize the cost benefit and the availability of material for new systems and spares.
A correction to this last paragraph by David: The 20 systems in the first full build includes:
--2 lab development systems per hemisphere, total = 4
--2 designated "ready spare" DHE's per hemisphere, total = 4
--4 presently uncommitted systems
Grand total of "undeployed" systems = 12
Your notes neglected to mention the 4 presently uncommitted systems. The rest of the information and discussion summary in parag 3 is spot on.
4. Mark reports that the new heat exchanger will be delivered on Monday. Expectations are high that this will meet the thermal requirements of the Torrent DHE. Peter reported that the measured power dissipation for the DHE is 37 watts in worst case situation. He also mentioned that with the covers in place, the internal FPGA temperature appeared to be almost stabilized (i.e. very low rate of temperature rise) at just above 39 Deg. C
5. Nick reports that he is currently fixing minor bugs (from Marks report) and proceeding with the Tucson 2 system documentation. Tucson 1 and Tucson 2 documentation should be finished by Monday.
6. Nick confirmed that the south PAN systems have been updated with the latest (and improved performance) borg code. Also that the south systems currently only read the TSM eeprom upon start up when the collector program interrogates the DHE. This will be fixed in the future and will increase the time for the boot process to complete by approx. 20 seconds from what it is currently.
7. David reported on KOSMOS and gave encouraging throughput figures for the recently delivered grating assembly - this brings pressure to bear upon the Torrent performance since the spectrograph performance may be determined by the detector read noise values - Every photon counts !
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110824
Present: Mark, Ron, Nick, Guillermo, Kathie (?), David Sprayberry, Peter.
1. Status update on the design verification testing. Some progress in testing made. Currently the detector is cold and waiting for section 7 testing to begin. This will commence once the configuration changeover from tom_pcm to chile1 is complete, stable, and tested for detector safety. Currently a static chile1 config is being worked on i..e. one that is not dynamically generated by the collect process on startup. This is to provide a basis to run the detector and establish base benchmarks. Then the corresponding attribute values will be transposed to the default set in the configuration data base and the collect configuration tested and compared to the base set.
2. Peter estimates the end of design verification testing still stands at the end of August.
3. Status update from Mark concerning progress from Tucson: Cable and connector work continues to be carried out. Stock taking to assess the materials that we have in hand for the production lot run is also continuing. A key item, the 78pin D connector for the TSM is missing half of the bought in lot. Maybe Dee will know where these items are.
ACTION: Ron, Dee Locate or otherwise account for the missing parts.
4. Nick reports that all work on the SysConfig tools is complete, however, there are several improvements that should be made. These are: a). the default Borg display attribute allocation needs cleaning up. b). a method to automate the generation of the [sysname]_SetVoltages.mod file c). tools to implement detector tuning - a plan viciously opposed by Peter.
ACTION: Mark. Approve or disapprove the extra work asked for by Nick.
5. Peter pointed out that if an automated method was introduced to generate the set voltage.mod file then a function must be available to back annotate these values in the configuration data base after a detector engineer has optimized the settings for a certain application. Nick agrees and will look into the implications; since this also implies re-writing the eeprom files and their associated checksums, etc.
6. Ron asked when the test Dewar and two DHE assembles currently being tested for design verification can be shipped back to Tucson. Peter replied that this will occur as soon as the testing has finished which is planned for the end of August.
7. Ron stated that preparations for testing the thermal performance of Torrent (according to the design verification requirements) should start later this week however, the new heat exchanger will not be delivered until the end of August. For these tests another DHE (Tucson1 ?) will be used as soon as it has been tested to show reliable and consistent performance similar to that shown on the chile1 and chile2 DHEs. For this to occur, the modified firmware (with the DAC reset mods, etc.) must be loaded to the Tucson DHE. Peter asked that before this firmware is emailed to Tucson, all existing versions of TorrentFpga_Ver207.msc and .bit must be erased from the computers in Tucson.
ACTION: Ron to be responsible for erasing these files in Tucson and then emailing
Peter requesting the firmware modules.
7. David suggested that an meeting be set up to specifically discuss the results of the design verification testing. This is set to be at 11:00 on September 6th. If there are delays that prevent this meeting from taking place (i.e. report not generated) then this will be advised on the Thursday or Friday before this meeting. David suggested that the completed testing document should be enough to discuss at such a meeting. Peter asked that we stick to the plan to prepare a report that summarizes the results and makes recommendations before it is reviewed.
ACTION: Peter and Mark: setup the meeting rooms for September 6th for thsi meeting.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110826
Present: Mark, Nick, Ron, Guillermo, Kathie, Peter.
1. Status from Tucson: Mark tells us that inventory counting and small piece construction continues to be the main items being worked on. In addition, assembly of the resources that are required to carry out the thermal tests associated with the design verification testing have begun.
2. Status from La Serena: Peter tells us that a stable configuration base has been achieved and a sequencer has been written which gives stable readout capabilities. The next step is to connect the DHE to the Dewar after one more checkout at which time the tests in section 7 can be carried out.
3. Mark tells Peter that you can tell if the test Dewar detector is alive due to the 2v offset that appears on the R/H amplifier output. Also that there are several features on the R/H side that need to freeze out at below -105 Deg. C before the detector is useful.
4. Peter asked what type of temperature sensor was installed in the Dewar since the calibration of the temperature sense channels is not reliable with the two point resistive divider method used for board test. Ron informed Peter that the temperature sensor is a calibrated diode model number DT470 which is actually obsolete. However, the calibration curve for the DT670 is identical for our practical purposes. This can be found on the Lakeshore web site.
5. When asked by Mark if there was anything else form the south, Peter took immediate advantage of the offer and asked for a new multimeter to replace the > 10 year old model currently in use (a Fluke 8024 model). This trusty rusty Fluke is now beginning to suffer from age. Ron immediately seconded the idea and made himself responsible for shipping one down.
ACTION: Mark, Ron, to ship a new trusty rusty multimeter to Peter.
6. Nick and Ron commented on the compatibility of the tom_pcm configuration with Borg ... It isn't and probably never will be.
7. A short discussion of the upcoming thermal tests ensued: Ron stated that the idea is to baseline the current DHE mechanical / thermal design in the environmental chamber and to then see the difference that the new heat exchanger fan cover will make by testing again. This will give a clearer picture as to the thermal properties and the accuracy of the thermal model than just testing the new design.
8. Nick reports that the PAN computer masterdon has been updated but that some files were missing during testing. This prompted Nick to make the suggestion that either a list or a script needs to be written to carry the MONSOON and Torrent software products to a new computer or to update an existing PAN computer.
ACTION: Nick to write either a check list or a runnable script to assist in loading
MONSOON / Torrent files to new or updated machines.
9. Kathie reports that she is counting hardware (connectors and such) to see what needs to be brought in to complete the next 16 Torrent kits. Kathie is also building the connector cables for these kits.
10. Guillermo reports that he is still counting sheep .. err, ICs (that have begun to look like sheep ?) for the stock inventory before the 16 Torrent kits are produced. This electronic parts stock take should be finished by Tuesday next.
11. Nick asked for the 'official' number of pins on the 'standard' CTIO Dewar .. The answer .. after some teasing, is 41. The Dewar mating connector part number is AMPHENOL PT06A-20-41S-SR
12. Peter asked for the average hold time that could be expected from a 'standard' KPNO Dewar and was politely told that this is about 12-14 hours. Peter so far has a recorded a wet hold time of 16 hours so the Dewar vacuum must be good.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110907
Present: Mark, D2, D3, Ron, Dee, Peter, Ricardo, Guillermo, Kathie, Nick
David Sprayberry – Addresses Action items,
Mark reports on Monsoon orange schedule for contingency. Assumes only one person working on the orange build.
Late Oct. optimization of orange with detector if there is a Dewar. Mark and D2 discuss this further including shop availability. Roger stated that there were many Torrent mods And they would have to be delayed in order to get the orange work done.
Mark also states we would be short a CCD transition board for an orange build.
David inquired with Peter concerning the hard orange chassis Peter reports it is minus an MCB. Mark says we have enough MCB's but not systran CMC modules. CMC costs $2500 or so.
D2 last question about spares for orange. Mark says the CCD acq. would be the shortage.
Also, The mosaic spares are what would be used for CCD acq boards.
D3 said that we should look into taking boards from Quota. Mark and Dave discuss board history of quota boards.
Discussion that Kathie would be the person to build the orange systems. Guillermo would help. More discussion about test and integration of these orange systems.
D2 states integration to Dewar as early as 10/25/11. Also asks about availability of a test Dewar to do the optimization.
D2 asks if we have all the parts, e.g. shop made parts. How much time to build the Dewar? Kathie says it can be done by the 13th or 14th including a detector installed.
Peter asks when we want the test Dewar sent back to Tucson. D2 states we'll decide that in the near future.
D2 states the desire to have test Dewar's in both hemispheres. Peter states the desire to have a SOI Dewar in the south so it is like everything else they have deployed.
Ricardo Seconds this desire and adds that it also works better to support Cosmos and its system. D2 responds with concerns about shop capacity and warns that if it can go on a Telescope it probably will.
D2 starts discussion of what the assignments for the remedial work package are and Peter responds with a description of the plan. Peter explains he will work on the noise and the component changes will be done in Tucson and verified. Ron will continue thermal testing until then. After first stage of changes, a decision will be made as to Peter making a trip north to complete the testing on the noise. Peter states we expect to have results by the end of next week with the first round of modifications. There are three possible outcomes: 1 – noise is acceptable 2 - noise is improved but not to spec. (Peter to Tucson) 3 - no or reverse progress. (Peter to Tucson)
D2 Asked Ron about the thermal testing. Ron answered that it is as complete as possible And the results need to be discussed to determine of further testing is needed.
D2 Asked what support in the North is required in order to get the above plan into motion. Ron responded with technician support.
D2 and Kathie discussed what could be started to do two orange systems as far as shop work etc…. D3 offered Larry Riddel as a resource for the machining.
Peter asks if this is the point we "pull the ripcord". He has also stated that we have spent the majority of the meeting discussing the parachute. Peter says that there is also a bit of optimism concerning the noise of the Torrent AFE. D2 states we are not committing to plan B but are getting everything in line so we can after Ron and Peter get results or a lack thereof with the noise.
Peter reports on the histograms and spreadsheet he sent. All data is in ADU's. Concept is to run the front end in two modes to show where some of the noise is coming from. The results are explained by Peter in detail (not recorded here).
Peter in summary suggests that the preamp and either the clamp or the integrator amplifier is oscillating based on the histogram data. States the point is that without the preamp the noise levels reduce to ~4ADU rms. Indicates the preamp is the first level to of the onion to peel. He thinks it will get us in spec. after correcting the preamp noise contribution.
D2 will await further updates as to the progress.
Peter will send the ucode so that Ron can verify any changes to the AFE circuits as per the Peter Ron plan.
Nick asks that all code and config be updated to big-boy so that it is all master copied and current. Nick gave Peter the command to do the big-boy update. Peter will give it a go.
Kathie asks if the official list of action items can be sent out to everybody. D2 will do this after D3's look over. D2 will send notes and action item list.
Peter and Ron will continue to attack the noise issues. Ron will send out a write up of the thermal testing results.
Dee asked if she should continue ordering parts for Torrent. Mark and D2 discussed what is left to order and what is left in inventory. D2 asks what it was Dee was going to order. She stated parts and then boards.
D2 states that the requisitions should be directed to him only. Also told Dee to hold off on the AFE board. D2 states that we can live with a 10 calendar day delay. Mark asked if it is 2011 monies and if it would carry over into FY12. Dee believes that the quotes she has will not change dramatically since it is fab work.
Peter asks if the metal work has been modified for the switch displacement for the TSM grounding switch.
Return to the top of the page Action Items#top
David Sprayberry notes from the meeting
Re Buildup of 2 MONSOON Orange units:
Email of 4:58pm Tue 6 Sep from Mark Hunten
---begin email--- We have most items needed for building the two Orange CCD systems.
We have: 3 - MCB 3 - CBB 2 - CCD Acq 2 - power supply chassis including all supplies (needing some machine work) 2 - enclosures (needing some machine work) 2 - 6 slot backplanes 2 - CBB transition boards 1 - CCD Transition board
We are short: 2 - Systran CMC modules, since we have the KOSMOS, COSMOS & CHIRON computers already with PCIe ends) - Front metals for MCB and CBB 4 - small fans for internal circulation - maybe a mil circ connector 1 - Transition board for one CCD Acq - Perhaps some metal parts for the enclosure - probably some fasteners that we get from McMaster Carr
There needs some machine shop work on the power supply enclosures as well as some of the chassis parts. ---end email---
My questions based on above email (answers from Wednesday meeting in italics): · Any way machine shop work can be sent out? Or is there anyone else we can add to shop temporarily (Reddell?) Mark has checked with Roger, our shop can do it in one week from signal to start as long as it comes ahead of (non-urgent) torrent mods he was given yesterday by Dee. Per Sawyer, Reddell could also be made available for a few days if needed · Is transition board something we build up in house? Who/how long? Order it from outside, usually; in schedule, line 37. · Cost and time to get Systran modules? Has question in to Systran person, no answer on delivery time yet. Cost in past has been $2.5k to $3k. · Making 2 would leave us with 1 spare MCB, 1 spare CBB (no transition board) and no other spare parts o “Extra” unit in Chile from old 1.5m spectrograph pre-CHIRON could be a joint n/s spare (apart from fact that it’s missing an MCB, but there is an extra MCB in N) o QUOTA system could be considered for spares per D Sawyer comments; need to check with Pat K about WIYN observatory plans for QUOTA. · Green test dewar; plan is to build it up by the dog & pony show by the 13th or 14th.
Re Remedial Work on Noise, Crosstalk, other problems:
Ron would continue with thermal tests. Those are essential to understand the suitability of TORRENT for the environment. · Ron: essentially done, but until noise is down can’t do last test which is bias stability over 20 deg C change. · Ron: testing essentially done for now, remaining write-up and discussion can be done in parallel with noise work outlined below · Ron: has preliminary results, not entirely happy with them but wants to go over them with Peter before discussing in front of the group.
Peter should place all emphasis on killing the noise. He will send up modifications for Ron to implement in Tucson.
No numbers on people needed for all other tasks. Peter did not feel comfortable assigning documentation and other tasks to people in Tucson at present, but Ron and Peter will be focused on the hardware and firmware plus testing.
Agreed that first installment of mods for noise abatement would be sent to Tucson tomorrow and test results would be as soon as possible after that. Successive mods would come up until they had gone through each of the 4 sections of the front end.
Status. Assuming that we can complete first run of mods to decrease noise by end of next week, either we meet specs or don’t. If don’t, but there is some improvement, we will then evaluate at the end of next week, when to have Peter come up and how to attack it.
Peter: today’s spreadsheet shows that preamp is oscillating (a lot) and one other amplifier (maybe clamp?) is oscillating differently at a lower amplitude. Shows where the problem is; not necessarily how to solve it, but gives important ideas about where to start.
Action Items:
1) Mark to complete information on parts costs for buildup of two MONSOON Orange systems:
a.Get cost and delivery time from Systran for fiber modules
b.Get costs and delivery times for CCD Acq transition board(s) from fab house(s).
c.Get complete inventory with costs of remaining small parts (fans, connectors, etc)
2) Mark to update schedule for MONSOON fab to reflect 19 Sept start date and send electronic version to Sprayberry & Sawyer.
3) Mark to launch instrument shop on mechanical work needed for two MONSOON Orange units; target is to have work completed by Monday 19 September. Coordinate with Repp and Sawyer, obtain assistance in shop from Larry Reddell if needed.
4) Kathie to complete buildup of green test dewar as currently planned; Kathie to be ready to begin assembly of MONSOON Orange units no later than Monday 19 September.
5) Ron to write up results to-date from thermal testing, review with Peter and rest of Torrent team, then present to Sprayberry & Sawyer.
6) Noise abatement effort to proceed under leadership of Peter
a.Peter will send recommended AFE mods to Tucson;
b.Ron will implement mods, carry out tests to evaluate noise, and report results back to Peter and Mark.
c.Guillermo will support Ron in implementing the mods.
d.Peter, in consultation with Mark, will determine mods for next iteration; return to step (a) and repeat as necessary until either:
i.Noise meets specs, or
ii.Noon CDT on Friday 16 September.
7) Mark to complete the workup of costs for parts + board orders for the Torrent production run, assuming 20 units. It is understood that this cannot be completed for the AFE board until the noise abatement effort has succeeded. Mark is to send the complete workup of costs for the PSM and LCB – net of existing inventory – to Sprayberry and Sawyer as soon as possible but not later than 9am MST Friday 16 September.
8) Torrent Group + Sprayberry and Sawyer to reconvene at TBD time on Friday 16 September (default is 1100 MST / 1500 CDT). Agenda:
a.Status report on Noise Abatement efforts: Peter, Ron et al
b.Status report on readiness to build up two MONSOON Orange systems: Mark, Kathie, et al
c.Discussions of remaining obstacles
d.Decisions on next step
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110914
Present in the meeting: Mark Hunten, Peter Moore, Ricardo Schmidt, Kathie Zelaya, Dee Stover, Guillermo Montijo, Nick Buchholz, David Sprayberry Marco Bonati, Ron George.
Noise testing,
Peter Summarized the noise testing to date. Both Ron and Peter are getting repeatable results of ~6 ADU noise on both the Chile and Tucson Torrent DHE's. Ron and Peter both feel the preamp is twitchy and that the noise reductions to date are a result of a combination of bandwidth limiting and grounding changes at the preamp. 2 potential routes being explored one that moves the preamp right up to the AFE as an experiment and another that selects potentially better performing amplifiers. Mark and Peter had an exchange and discussion about the previous choices in amplifiers for the analog signal chain circuitry. Peter stated that the goal of the noise reduction would be to get the noise down to 3-4 ADU so that it is not the dominant noise contributor to the DHE detector system. He went on to describe a suspect oscillation with the AD829 in the signal chain and that is what he is currently working on. He is also concerned that a difference between the Tucson and Chile DHE is the bias gradient seen. The bias gradient is seen on the Tucson hardware for the first couple of frames and then goes away. On the Chile hardware it is changing as the controller warms up. Mark discussed that there is some possibility that the bias gradient is a result of the ucode Peter is running. Ron agreed to send Peter a copy of the .asm code that he has been running to do the Tucson noise testing. Peter believes the gradient comes from the integrator circuit. Needs to be looked into in more detail. End noise report.
Nick reported on his efforts to implement ROI (region of interest) within the Torrent structure. He detailed different potential options for ROI and questioned how the excess or ghost data would be handled. Mark and Nick discussed how Mosaic does file extensions and the impact of this on ROI. Peter commented on some of the options and hypothesized that in certain options presented by Nick the possibility that some of the data would be lost existed. Nick explained in further detail how it works and the other ROI options were discussed. Ricardo was asked how it is done at CTIO and Ricardo deferred to Marco who was called into the meeting at this point. Mark stated that we should be consistent in the ROI implementation between both hemispheres and that we should follow the CTIO model in order to maintain the consistency. Peter stated he believes the observer should get the exact amount of data requested in the ROI with no other ghost or excess data. Marco arrives to the meeting. Peter explains the question to Marco who then begins explaining in detail how ROI has been implemented at CTIO. He states the observer is allowed to specify any ROI but the detector must be read out symmetrically. He continued explaining that it is only the ROI data the observer specified and the extra data is not delivered with the final image. Marco explained that this was done through fits header keywords that identify the ROI to be delivered. David Sprayberry arrives. Mark inquired about the overscan data and Marco replied that the overscan is always saved for the ROI expounding with an explanation that the readout for the rest of a row is skipped after the ROI until the overscan and then readout commences again for the overscan. Ricardo then inquired about how binning is handled and furthermore about how binning is handled in the overscan. Marco states that the ROI (including the overscan) is always explicitly stated as non binned geometry and that the binning is then done after that. Nick decided he will revise his ROI scheme to work implicitly un-binned. Marco reiterates how much easier it is to do ROI un-binned. Peter then explains how the Monsoon sequencer loop registers work for binning and ROI, Nick elaborates even further on Peters explanation. Marco is dismissed but decides to stay for the rest of the meeting. Nick asked David Sprayberry what account to charge the ROI work to. David replied that it should all go on a Monsoon account. David Sprayberry briefly spoke about the OMB visit that took place in Tucson this morning. David asked about noise in just a few words or less and Ron responded with just one, the number "six".
Mark adjourned the meeting.
Notes taken by Ron George
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110921
Present in the meeting: Nick Buchholz, Ron George, Guillermo Montijo, Peter Moore, Ricardo Schmidt, Dee Stover, Kathie Zelaya.
1. A discussion on Torrent noise abatement progress was initiated. Ron reported that the AD4841 amplifier module built by Guillermo demonstrates 4.2 ADU noise when coupled directly to the AFE. Other amplifier types (AD829, OPA1641, and OPA827) were tried with slightly worse results (5.1, 4.6, 5.2 ADU respectively). It was decided to progress with the AD4841 design and look for the cause of the extra noise at the preamp layout and flex circuit. Ron and Guillermo also reported that the 10 Ohm power supply filter series resistors on the preamp had also been tried on the experimental board with very little effect. It was decided to progress with the idea that either (a) - the actual layout of the preamp board or (b) - the interaction between multiple amplifiers or (c) - the capacitance effect of the flex circuit is causing the extra noise we see (A later communication from Ron reported that the flex circuit added an additional 0.2 ADU noise component). It was agreed that Ron and Guillermo would continue the testing to identify and rectify the noise problem of the preamplifier.
2. Ron stated that the parts requested by Peter are on order and should be in the Tucson labs by early next week.
3. Peter reported that travel arrangements have been made to put Peter in the Tucson labs on the morning of Monday 26th (ETA 11:00 at NOAO).
ACTION: Tasks for Peter in Tucson:
1. Fix low-temperature high-voltage bias startup problem
2. Bring bias stability into specification.
3. Supervise AFE modifications for above items
+ AFE_OE signal drive strength
+ U128 dissipation problem.
4. Bring AFE and if necessary TSM_PRE board rework to readiness for
re-order for manufacture
4. A discussion ensued concerning the AFE noise abatement issues and the bias stability problem. Peter explained that the first approach is to try out a LT6202 amplifier in a SOT23-5 package in the buffer position since the alternative, that of making the buffer amplifier an 8-pin SOIC device, is out of the question due to layout complications. This amplifier should provide an improvement in the AFE noise figure. In addition we have identified another amplifier (LTC6240) that should be tried out in the integrator position. The problem with the bias stability seems to be coming from the charge injection of the analog switch that is driving the integrator. The amount of charge injection is not a linear function of time between the switch delays which tends to indicate a charge accumulation effect that is known to be very sensitive to time and therefore could well be sensitive to temperature as well. Increasing the value of the integration capacitor will proportionally decrease the effect and capacitors have been ordered to try out this remedy to minimize the bias drift to below the noise level.
5. Nick asked that we all review the specifications as written on up the Torrent wiki page concerning the proposed system for implementing ROI readouts. Nick asked that we consider modifying the current use of the loop registers to enable a normalized readout scheme when the ROI code is implemented. This proposal makes use of a set of registers that describe the horizontal over scan region and is used for both normal and ROI readouts i.e. the 'seqColCount' registers will _not_ include the over scan region which will be set in a different loop register.
ACTION: Everyone who has an interest (Especially those involved in
sequencer code development) must read the ROI on the NOAO wiki
/ Torrent sequencer page to comment before yesterday on the ROI implementation
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110928
Present in the meeting: Jay, David S, Kathie, Ron, Nick, Mark, Ricardo, Guillermo, Peter.
1. Ron gave a summary of the work being carried out to reduce the noise of the Torrent preamplifier. So far the best noise figure we have for the complete video signal processor is 5.5 ADU rms using the AD4841 amplifier. This figure is obtained when only one amplifier is enabled on the preamp. Many experiments have been used in the efforts to identify the principle noise contribution mechanism. In addition several different amplifier types have been tried in an effort to reduce the KTC noise of the preamplifier itself. It has been decided that the AD4841 amplifier will be used (with corresponding modifications to the ground and input circuit) for the preamplifier and future efforts will be aimed at reducing the KTC noise contribution of the amplifiers on the AFE board.
2. Ron also gave an account of the work being carried out on the bias stability / charge injection problem. This problem concerns an unusually high charge step contribution on the AFE integrator stage when the analog switches controlling the integrator are turned on. The MONSOON Orange system uses the same architecture but a different type of analog switch which does not show this marked behavior. The amount of charge injection directly affects the dc bias value of the pixel data value. The amount of charge injected from each switch is proportional to the temperature and the time between the last activation of the switch (i.e. there is an exponential change of the charge injected dependent on the time since the switch was switched off). The temperature dependence this signal means that the detector bias values along the column axis direction during a readout shows a gradient.
ACTION: Peter to have a status report and prediction of the resource cost of further efforts
to reduce noise and improve bias stability.
3. Nick reported that the software task list from the design review report document is complete. This includes the availability of ROI functionality. The ROI function will be / can be back propagated to the MONSOON Orange systems.
4. A short discussion was held to determine if the Y axis over scan functionality is required. It was determined that the functionality is not required for normal observatory procedures and that the same functionality can be emulated by extending the Y dimension size declaration to the software.
ACTION: Nick, Peter to remove the Y over scan functionality from software and sequencer code.
5. Kathie gave a concise report on the work being carried out to prepare the two MONSOON Orange controllers. The first power supply is complete except for some fuses which are to be delivered. The second power supply should be finished today. The mechanical work on the chassis is in progress in the mechanical workshop. This should be complete by the end of the week. Overall, the Orange production is "moving very well".
6. As a side bar to Kathie's summary, Ron asked if we should move the laboratory stock to the stock room control so as to facilitate control and restocking. It was agreed that this will be a good idea. Ron, Mark, and David commented that the work currently being carried out by Beth in identifying old stock and maintaining the control of the stockroom is showing benefit and is much appreciated.
ACTION: Ron is to facilitate the transfer of the lab stock to the stockroom control. Existing lab
stock cost will be zero if the parts have been paid for by a previous project.
7. Kathie asked if work should continue on the "Green Dewar". There is one part that needs to be fabricated to complete the parts list. Ron said that he is awaiting drawings from Dee to allow the parts to be either made from scratch or made from modified existing parts. The mechanical resource to carry this work out is estimated to be 1 week. In addition, a small PC board needs to be made which would take a few hours to complete. Jay told us that this work needs to be placed behind the fabrication of the MONSOON Orange controllers destined for KOSMOS and COSMOS. However, the "green Dewar" should be completed before the COSMOS Dewar integration work. It was further clarified that the "green Dewar" is destined to become the E2V test Dewar for CTIO (and is not a replacement for the MINIMO Dewar).
8. A short discussion ensued on the requirement to have the science grade E2V detectors, Dewars, and cameras functional and available to Ohio State for both KOSMOS (required early February 2012) and COSMOS (required late April 2012). Allowing six to eight weeks for integration means that the controllers and Dewars must be ready on mid-November for KOSMOS and late February for COSMOS.
ACTION: Kathie to keep David S. and Jay updated as to the status of the MONSOON Orange controllers.
9. Ricardo asked about the Dewar fabrication for the two spectrographs. David s. Replied that there are two Dewars for COSMOS being fabricated in Tucson from plans made available from Patricio at NOAO South. These Dewars are to be used for the E2V and LBNL detectors and are based on the STELES design.
10. Ricardo asked about the development work to be done on the LBNL detectors for the two spectrographs. Jay stated that the development and implementation of these detectors is planned to be done in 2012 but after we have the E2V science detectors operational. The development site (NOAO North or South) would probably depend on who has the resources available first. David S. reported that LBNL currently has two packaged science detectors ready to be shipped. a third detector is to be selected soon for packaging. There is a mechanical sample in house in Tucson. Peter suggested that the mechanical sample be sent to Chile for evaluation while there is still time before the detector engineering needs to be done. Jay told us that the sample will stay in Tucson until the engineering work for the mount is completed here. It was agreed to send the LBNL detector manual to Ricardo. Mark stated that there are several corrections that need to be made to the manual to make it correct.
ACTION: Mark to send corrections to the manual to LBNL and Ricardo and Peter.
11. A short discussion ensued concerning the need to clear the shipment of the LBNL detector to Chile via ITAR for possible integration there. It was agreed that official procedure channels (information available from Cliff Aldrich) needs to be used concerning this.
Return to the top of the page Action Items#top
Torrent Work Effort Status sent 20110929
Gentlemen,
For consideration in tomorrows meeting; a short status report on the Torrent DHE noise reduction work carried out this last week here in Tucson.
Torrent noise reduction task status - 26th - 29th September 2011 (*)
At this time we can report that the Torrent noise contribution of the Torrent controller is 3.6 ADU rms at an input sensitivity of 2.5uv/ADU. Using the 'normal' E2V detector sensitivity of 5uv/e- and noise figure of 3e- rms for the detector, the Torrent controller contribution becomes 0.47e- rms for a total expected noise value of 3.47e- rms at 111 KPix/chan/sec.
During design validation testing we measured a total noise figure of 12 ADU rms (i.e. 6e- rms). The Torrent DHE was at that time contributing 7 ADU. We have been able to reduce this by approx. 50% to levels which will satisfy the goal for the Torrent design. This result needs to be confirmed by measuring the noise figure on the E2V detector in the test Dewar again i.e. repeating the design verification test 7.2.1 next week. Further work in this area will be suspended pending verification of the noise performance.
The bias gradient seen in the column direction when taking a raw image has been improved by 50% to a 28 ADU delta across 4100 pixels read. This has been achieved by reducing the effects of charge injection form the analog switch that controls the signal integrator in the video signal processing circuit.
Unfortunately the temperature sensitivity of the signal processing circuit does not appear to have been improved at this time. If replacement of the analog switch type (programmed for tomorrow) does not reduce the temperature sensitivity (by reducing the charge injected from the switch) then further investigation will need to be performed next week to reduce the frame bias gradient to acceptable values. When we have the frame gradient below the requirement (<= 2 ADU) we will concentrate on reducing the temperature effects on bias stability.
3. We will endeavor to mount the Torrent DHE (SN CT204) to the E2V Test Dewar early next week and begin noise tests using the modified AFE board (SN014) and preamp circuits. At the same time we will attempt to use the alternate Torrent DHE with a modified AFE (SN017) on the bench to analyze the bias stability problem. This same system will later (mid-week ?) be used in the environmental chamber to characterize the bias stability over temperature and to begin looking at the high voltage bias cold temperature start up problem.
4. We expect that work on bring the bias stability into specification will occupy most of next week (Oct 3 - Oct 7). During this time we should also be documenting the rework that is to be carried out on the layout of the AFE so that Dee may begin this task.
5. At the end of next week we should have the bias stability problem solved, The AFE rework documented and, hopefully, a handle on the high voltage cold start problem i.e. know what the problem is.
(*) Please note that this report summarizes the results of work that has been carried out over the last three weeks at both NOAO North and South facilities. The date range only indicates the period in which the culmination of these efforts by the team has come together to yield the results that are contained in this report.
We look forward to discussing these results and refining the plan of action for the immediate future at tomorrows meeting.
Cheers,
Peter.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20110930
Present in the meeting: Jay, David S (D2), Dave (D3), Kathie, Ron, Nick, Mark, Guillermo, Peter.
Report on Orange assembly progress: Kathie was not available so Mark and Ron gave a short report: Both power supplies have been completed, Transition boards on order, essentially on schedule, expected completion date as per original project plan. David suggested that the probable delays will become apparent once testing of the boards and system begins. Kathie, now present, asked that the system configuration documention be completed and passed to her.
ACTION: Mark is to complete the system configuration documentation for the two Orange spectrograph
controllers and pass to Kathie asap.
ACTION: Dee to keep tabs on transition board fabrication to make sure delivery is on time.
A discussion followed concerning the detector, Dewar, controller integration schedule. 1st integration package for KOSMOS starts 1st November. At that point science grade can be delivered on test Dewar. David asked if the two Dewars will be required to complete the work. Ron replied that having two Dewars will reduce the number of times the detectors will need to be swapped out and so expedite work and enhance safety.
ACTION: David S. is to make a decision on 17th October as to whether or not to build up the second
"green' Dewar for integration work.
It was reported that the one missing part is currently in the mechanical workshop awaiting a cost estimation. The plan is to build several of these parts to be used on other Dewars in the future.
ACTION: This week, cost to be considered and an order placed to either proceed with one or several
parts to be made to the workshop.
A report on the Torrent work efforts to reduce noise and improve bias stability was circulated Thursday. This report was discussed and more details given. The plan (contained in the status report) was approved for next weeks work to continue.
ACTION: By Friday 7th October, another status report is to be given to David and Jay for evaluation.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20111019
Present in the meeting: David S (D2), Dee, Guillermo, Kathie, Peter.
Dee asked for clarification on the number of Torrent Kits to be built up and if these should be built serially or in parallel. David answered that we need 5 or 6 configured Torrents to fit to the time schedule which suggests that the kits be built in quasi serial fashion to meet the deadlines. 16 More kits need to be produced and Dee can count on the dedicated participation of both Kathie and Guillermo to help in the procurement of these systems. The Torrent configurations needed to fulfill obligations in the short term are:
1st December - KOSMOS E2V Science detector characterization. 12th December - NOAO South - Chiron E2V controller replacement. 1st March 2012 - NOAO South - COSMOS E2V science (?) detector characterization. asap - NOAO North - Torrent laboratory standard asap - NOAO South - Torrent laboratory standard 1st May - NOAO North - KOSMOS LBNL detector characterization. 1st May - NOAO North - COSMOS LBNL detector characterization.
ACTION: Dee is responsible for the procurement of the Torrent kits. David and Peter
(and everyone) will help out with whatever is necessary for Dee to complete
this task.
Dee and David asked for clarification as to whether the preamp will require a re-layout. Peter replied that no, the work done by Guillermo and Ron shows that the defined ECO used to bring the first Preamp within specification can be reliably applied to the existing Preamp layout to meet specifications. A review of the Preamp layout will be probably be done when we attempt to use the Preamp for the P-Channel (LBNL) application.
ACTION: Peter to add Preamp safety switch wiring modification to Preamp ECO.
David asked if the recently ordered parts will be delivered on time to meet the schedule for configured Torrent systems. Dee replied that most parts are in house and the rest will be delivered by end of October. In the meantime, board manufacture goes ahead and part kits delivered to the manufacturer in a timely manner to allow board delivery by the end of November. This allows approx. 3 days to inspect and test the boards before they are required for the first system (KOSMOS Science detector characterization). This assumes a 25 day turn around for the AFE boards with the order being cut this Friday. David said he would be standing by to approve any such order that went out on Friday. This schedule looks tight (is tight) but no roadblocks are foreseen so it should be do'able.
The heat exchanger part is in process of a RFQ and not expected to unduly delay the procurement of the first kits and systems.
The mechanical workshop should have capacity for Torrent after Christmas for the expected 80 Hours of mechanical work required for new kits.
ACTION: Peter to talk with Dee concerning the fix for the TSM presence switch sensitivity
ACTION: Peter to send back the two production systems from NOAO South. The south prototype
system will be used to develop and test the production firmware version.
David suggested that he write up a brief news item for this weeks internal newsletter to tell the rest of NOAO of the rollover of Torrent to production status. Peter suggested that if possible, this should include the main characteristics of the system i.e. that it is built to normalize the technology at the observatories, be reliable and easily maintained, and, through the interchangeability of the controller section, be extremely quick to replace to minimize downtime should it fail.
ACTION Kathie is to procure some suitable pictures to give to Beth who will put them on
the MONSOON / Torrent website so that a link can be put in the news item.
Kathie asked if the priority to build and test the MONSOON Orange systems has changed in light of the progress that Torrent has made. David replied that NOAO North still requires one complete system to be used as a backup for MOSAIC II and work should continue on construction. However, the deadline to finish this work has been removed and the highest priority is to help Dee procure the Torrent Kits. Kathie reported that while waiting for mechanical parts, the current advance on building one Orange still allows delivery as per the schedule.
Kathie asked what is the priority of the mechanical workshop with respect to the remaining work required to machine the plug plate for the Orange. David replied that the MONSOON priority for the mechanical workshop remains the same i.e. high.
David mentioned that besides helping Dee with the Torrent kit procurement, and working on the Orange system, Kathie should be available if Ron requires backup with the test system work. Guillermo is currently working with Ron on this but Kathie should also be prepared to help if required.
Kathie asked about the Green Dewar - David replied that this will still be built but there is no urgency on this now.
David asked Guillermo for an approximate date when the first test set for Torrent would be operational. Guillermo ducked and allowed this question to pass overhead towards the ears of Ron. Dee, Kathie, and Guillermo confirmed that the first test set hardware had been powered up without incident and described Ron's reaction to the event as 'positively bubbley'; First tests are assumed to be encouraging.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20111026
Present in the meeting: David S (D2), Jay, Mike, Ron, Dee, Guillermo, Kathie, Nick, Peter.
Torrent meeting minutes taken by R. George
1. Peter opened the meeting with a walk around the table identifying each individuals responsibilities and contribution to the Torrent project. Each member agreed with the identified responsibilities. Peter concluded that there adequate resources to move forward. Jay and D2 reiterated that Torrent is the priority over the Monsoon orange build in the event of conflict. Ron mentioned D3's wishes of using the eng. grade that is in Ohio to do the science grade optimization. D2 and Jay opposed the idea since the device is being used to align the optics.
2. Dee reported all parts (except for 3) are ordered. AFE's are ordered and started. LCB and PSM are ordered but not started. PSM changes finished and ready for basic review. Dee is building a spreadsheet to organize what is where and at what state of build. This will identify the priorities and shortages.
3. No kits are being built yet due to mods of parts and arrival of production boards. Dee is concerned about resources. Peter stated there may be some electronic tech time available in the south.
4. Parts must go back to the shop in order to address this issue properly. D2 is going to address this in the shop planning to have the work done. We will use a jumper to bypass the TSM present switch in the meantime.
Ricardo inquired about the team of Tucson people traveling south to assist with DECAM efforts. D2 elaborated on the personnel, the work and the timeframe.
Peter asked about the heat exchangers and D2 stated he approved the req. with a promised delivery of Nov. 30.
5. Kathie reported that the effort to complete the Monsoon orange systems is making forward progress with a small hiccup. She has the parts in the shop and will get them back in 2 weeks. She is taking care of the details. Peter asked how much time on orange. She replied 75% on monsoon, 25% on other stuff for Ron. Peter asked about the final disposition of the two systems. Ron spoke that one may be planned by D3 as a KP spare. Peter asked Kathie how the test system performed. Kathie said all was fine except for a clock channel with a slight offset.
6. Peter reported on the cold start fix. He found a problem with one of the components on the PSM that is also contributing to the problem. He would like the ECO to be implemented and tested in the enviro chamber. He feels problem is solved, and verifications is low priority. Ron is going to use Kathie to do this testing under his guidance. Peter discussed the specific changes to the release version of firmware he sent for the fix.
7. Ron spoke up about the test set status. Moving forward but needs the cables to be completed.
8. Nick reported on the standard config, ROI code etc... We have tested the ROI code on a detector and it does work. We will be implementing the ROI on the KP STA2 system in the next weeks. Configuration is not valid for Kosmos's. Nick is going to copy the Tucson2 so that the clocks and biases left/right are broken out completely for detector optimization. Nick is also addressing some minor issues with the sysConfig software in terms of Dewar temp control and status monitoring. Peter requests that the config tools be complete prior to turn over of the Kosmos science controllers in Dec. Nick thinks this deadline will be fine. Peter asked if there would be a release of the final software (pan etc..) that coincides with the first production Torrent? Nick replied yes and went into much detail about what he intends to do. I can't type fast enough to catch it all. Peter asked if this could also be done by early December. Nick wants to focus on the ROI general detector library first.
Ricardo and D2 conversed about DECAM, Chiron priorities and schedule including the effects of possible schedule slips.
Meeting adjourned.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20111102
Present in the meeting: Dave Sawyer, Peter Moore, Ricardo Schmidt, Dee Stover, Nick Buchholz, Kathie Zelaya, Guillermo Montijo.
Torrent meeting minutes taken by R. George
Note - Items were discussed out of order from the agenda.
1. LCB Xilinx device....
Dee and Peter discussed the market availability of the Xilinx Virtex5 FPGA used on the Torrent LCB. Peter stated that the parts should not be that hard to get and are available from many places. Dee identified many of the vendors that come up in a search as "gray market" vendors. Ron asked if anyone had contacted Xilinx about their distributors. Dee replied that she had and that Avnet is the only legitimate authorized dist. for Zilinx. Peter decided it would be best to source the parts now and have them mounted to waiting boards when they come in. Dee agreed and the team believes getting these on order is a priority.
2. AFE, LCB, PSM production progress...
Dee reported on the status of the production runs for the subject boards. She believes the AFE's will be a week late from the original date (which I failed to capture). The PSM's are being part kitted and will go out by Friday. She estimates the PSM's to be complete by Nov. 29th. Peter asked Ron if he thought the PSM test set would be ready by then. Ron replied that he hoped so but not if the Dac init problem isn't resolved. Dee and Peter also discussed the newly identified LCB shortage, which also plays into #1 above. The quantity needed was discussed and the general consensus was that we were at least 6 LCB's short due to the test systems and fallout. The price difference was brought up and Dee stated there is no real cost difference for fab between 6 and 10 boards, it is the parts cost that is the greatest difference.
Update to Items #1 and 2 from David Sprayberry: I just spoke with Ron and then Dee about the Xilinxen*. After consultation, I have instructed Dee to order 9 more FPGA's from Avnet immediately. This will support increasing our LCB order from 11 units to 20 units. Dee will also obtain quotes for the board fab at the higher quantity, and will give me a careful estimate for the cost of the remaining parts that need to be bought to increase from 11 to 20 LCB's. Overall goal is to get everything in place for enlarging the order from 11 -> 20 ASAP. Onward. David *my made-up plural for "Xilinx"
Update to Items #1 and 2 from Dee:I just have some corrections additions on Item 2: board production progress. AFE Part are late going out we received some wrong parts and purchasing errors. Once the vendor has received and audited the parts we will get a new dock date. I am estimating Nov 29th or 30th. Power supply board fabrication is under way, I must box the parts and send them off. This will be top priority after ordering stuff for the increased LCB quantity. They are currently on schedule with a dock date of 11/29 LCB - as stated by David (D2) and Peter for increased quantity. Only thing I will add is instead of a 2 week assembly time it might possibly need to go to a 3 week, hopefully not. I will have the $ figures today or tomorrow.
3. Kit production progress...
Peter and Dee discussed the status of the mechanical mods that are currently in the shop. The Kits are being held up because of the missing parts that are getting modified. Once these are in, the kitting will commence in full.
4. Kosmos configuration progress...
Peter asked Dave Sawyer when the Kosmos e2V science grade 44-82 detector would be optimized. Dave replied that he was planning to do it the week of Thanksgiving, late November. Nick went into detail about the new Kosmos config and tabled an idea for a grounding change in regards to the video returns. Peter commented about the ground currents and the unpredictability of the results by making this change. Ron chimed in with a repeated description of the star ground that is currently employed on the preamp between acom and vid rtn to get the low noise performance. After much discussion between Dave, Peter, Ron & Nick the team decided to leave the grounds as they are. Nick then went into an explanation of how the J1 and J2 connectors were configured and how they would conform to the standard that is adopted for all universal Dewars. Nick stated that the initial voltages for the config would be a clone of the Mosaic 1.1 voltages. Peter requested that Nick add 3 attributes to his config, the "cdsChanSlctReg", "clkPortConfigReg", and the "cdsPortConfigReg". The second two registers are used to mirror AFE2 to AFE1 or vice versa. Dave Sawyer inquired about vhtr and asked if the noise testing was done with or without it up and running. Both Peter and Ron responded that it had not been done while driving power, but is installed in the DHE the noise tests were done on. Peter further elaborated that the heater power is off when reading out in order to decrease noise anyhow. Ron and Peter agree that the current ucode (slated for use on Kosmos) runs 111Kpix/sec. Dave thinks this is fine for Kosmos.
5. Test system progress...
Ron gave a brief status of the tester debug and explained that a major problem he has run into is the AFE DAC initialization. Ron described the nature of the init problem and asked Peter for advice and or ways to defeat a possible lockout of the AFE interface by the firmware. Peter responded and advised Ron to look into the power status override register since the dac init may be due to a bad status of power.
6. Peter spoke about his efforts with the control servo for PSM voltage output regulation. There are no results to report yet. He also inquired with Dee as to if the documentation was still to be submitted in .pdf format. She responded with "yes".
Action items:
1. There was discussion about the heat exchanger fan after Dee raised the issue.
Ron and Peter have decided that an AC fan may be the best solution.
Ron is to spec and order a couple of AC fans to try out in this application.
2. Peter to send back the Chile1 and Chile2 DHE's by Friday before he commences
his holiday.
Nothing further - meeting adjourned.
Return to the top of the page Action Items#top
Torrent Meeting minutes 20111109
Present in the meeting: David Sprayberry, Dave Sawyer, Ron George, Dee Stover, Nick Buchholz, Kathie Zelaya, Guillermo Montijo, Peter Moore, Ricardo Schmidt.
Torrent meeting minutes taken by Peter Moore
1. Production progress: Good progress made. No big problems. AFE part (that was incorrect) is now in stock and will be sent to manufacturer. Memory component supply problem now solved. In summary, all parts required for Torrent are now on order or have been received. Focus is now to bring parts into kits and send out to board manufacturer.
2. Bottleneck for pcb incoming inspection and test. A short discussion was held to determine methods to alleviate the bottlneck of inspection and test tasks that will be required as the manufactured boards come in from the board houses. This task has been estimated at 1/2 day per board and there are approx. 90 boards to come in December. It was suggested that an independent and experienced outside contractor be hired to relieve the inspection work load. Some training in our procedures would be required but this is not deemed a significant problem if the person (recommended by Dee ?) is already experienced. Testing will still be a problem although the PSM test set should be operational by the time the PSM boards have been received and inspected. It was recognized that not all boards need be inspected at once and then there are resources available (providing these resources do not become re-assigned) to complete the inspection and test work to meet our commitments. It was agreed to look for an outsource for the inspection of the AFE boards to free up our technical personnel to concentrate on the rev -A- board ECO work.
ACTION: Dee is to contact the indicated person and ask for a quote to perform the incoming
inspection on 40 AFE boards.
3. Report on the PSM / MEZ test set. Good progress. The power loads work as designed (applause). Next week looking forward to first tests on a real test board and first run up with automated software.
4. KOSMOS. E2V test Dewar has been tested out from socket to DHE using Nicks configuration documentation for the Tucson 2 sys config (two detectors). Will run the KOSMOS science characterization using the Kosmos E2V system configuration (single detector). The Dewar is closed, the engineering detector is mounted, the FE55 source installed. Missing the molecular sieve that will be installed latter today. Ron reports that the second E2V detector connector was found 'banging around' i.e. unsecured in the Dewar when opened. This could have caused a short between the detector substrate and the Dewar metal work on previous attempts to use this Dewar. This Dewar is to be ready for the Thanksgiving weekend. The Torrent temperature control will be used to characterize the detector.
ACTION: Peter to finish off the temperature servo firmware in time for Thanksgiving.
5. Software. Nick reports that the system configurations for Tucson 1, Tucson 2, and Kosmos E2V have been regenerated using version 2.1 firmware sources. These are untested / unverified since no DHE exists for testing and so Collector cannot be run to test the configuration data base. The attributes specific to version 2.1 firmware are now visible and operational. ctiola has also been updated with this configuration - eepStore needs to be run on this machine to complete the process. Assimilate is now documented and awaits a review of grammar and spelling (?). Next progress i son sysConfig and write / update the documentation for this tool. STA2 system needs work to be done to incorporate the ROI code. This work is not needed until February. An array description file has been generated for the LBNL P-Channel detectors that where recently delivered. Peter asks to be included in the configuration process for these detectors.
ACTION: Nick to pass document source for the assimilate (and sysConfig) documentation to
Dee - The document keeper.
6. Ron reports that the modifications to the mezzanine board to allow the high voltages to be switched on at low temperature are being applied in Tucson and will be tested in the environmental chamber soon.
7. Chiron. Peter reports that the current plan is to swap out the existing Chile 1 DHE in NOAO south with a re-manufactured production DHE sometime in mid December. This will probably not happen although there is room for this to slip and still get Chiron working with the Torrent controller.
ACTION: Peter to send Chiron wiring list to Nick so that the system configuration process
can be started.
8. Peter asked David what was the best method for Juan Estada to begin negotiating for the a 'couple' of Torrent controllers for the (non-NOAO) high resolution beam imager project. David asked that Juan begin by contacting him directly. Meanwhile David will determine if we have sufficient Torrent systems to consider a request by Juan and what would be the correct acquisition procedure.
ACTION: David Sprayberry to update the list of Torrent controller assignments so everybody knows
what are our commitments.
UPDATE from David Sprayberry with regards to Torrent system assignments:
Destination Num Need Date
KOSMOS e2v 1 Nov 2011
CHIRON e2v 1 Dec 2011
COSMOS e2v 1 Feb 2012
Mini-MO 1 Feb 2012 what kind of CCDs?
KOSMOS LBNL 1 Mar 2012
COSMOS LBNL 1 Mar 2012
GoldCam 1 Mar 2012 what kind of CCD?
Mosaic-1.1 2 May 2012
Hydra-S LBNL 1 Jun 2012
Lab systems N 2
Lab systems S 2
Spares N 2
Spares S 2
Total known needs 18
FY11-12 Budget Funded 20
Return to the top of the page Action Items#top
