DAQ-962 Cost and Schedule Estimates ---------------------------------------- Original Rev. 12-JUNE-2007 Current Rev. 27-JUNE-2007 ***************************** ***** ***** ***** DRAFT Version ***** ***** ***** ***************************** This note provides the cost and schedule estimates for the T962 DAQ system. These estimates cover all of the equipment from the signal feedthrough port to writing event files on disk. In preparing this estimate we have made a number of assumptions which are based on discussions that took place in the T962 meeting on 7-June-2007. Specifically: - There are approximately 480 readout channels in T962. - The Wire Bias Voltage final filters and distribution resistors along with the readout signal DC blocking capacitors will all be inside the cryostat. A schematic of this is shown on the web page: www.pa.msu.edu/~edmunds/LArTPC/T962/bvdc_card.pdf - The DAQ system does not need to provide any signals for generating triggers for T962. - The DAQ system does not need to provide, in hardware, a zero suppressed readout data format. - Readout of the full binary data, i.e. about 2 megabytes per event, is what will be expected from this DAQ system. - The trigger rate will be about 0.5 Hz. - Major firmware modifications for the ADF-2 card are not needed. That is, firmware for the ADF-2 cards that provides an ADC sample once every 198 nsec with a total record length of 2048 samples, (405 usec) is adequate for the T962 experiment. - Data from the T962 DAQ system does not need to be merged with the data from other detector systems at the time of readout. It only needs to have each event time stamped so that all of the data from the various detector systems for a given event can be put together offline. - All equipment, firmware, and software must be ready for this system by the end of December 2007. The cost and schedule estimates are provided in the following sections: - Cables - Cards - Boxes, Crates, and Power Supplies - Computer and VME Interface: - Software and Firmware Cables: ------- - Readout Signal Feedthrough to PFC-16 Cards Theses cables which carry the low level signals to the input of the preamplifiers. They are standard 17 pair twist and flat cables with standard 2x17 0.1"x0.1" mass termination connectors. They are located within the inner shield box. Quantity: 30 Input twist-flat cables Materials Cost for each cable: $8.26 total 2 connectors at $2.12 each 2 18" sections of cable $4.02 Assembly time about 10 minutes for each cable Additional work: engineering time to verify cable lengths, make labels, design and make brackets that are necessary to hold the cables from moving - PFC-16 Output to ATC Cards (ADF-2 analog input) These cables carry the analog output signals from the PFC-16 cards to the ATC cards. The ATC cards connect these signals to the ADF-2 card analog inputs. These are 3M Pleated Foil Cables with 3M type MDR connectors. These cables must be long enough to run from the PFC-16 cards in the shield boxes on the cryostat to the location of the VME readout crate(s). From the electrical point of view any reasonable cable length (e.g. up to 100 ft) is OK. The easiest way to obtain these cables is to have them assembled by an outside company. Their cost will be a function of their length. Quantity: 30 Pleated Foil Cables Materials Cost for each cable: $ ? total 2 3M MDR connectors at $11.50 each 36 conductor PFC cable at $5.00 a foot Assembly cost at an outside vendor $ ? each 10' 3M Pleated Foil Cable assemblies using: 90211/36 cable, 10136-6000EC connectors, 10336-1230-00 shells coolamp silver plating on copper foil Quotation 03212005-1 qty 192 @ $89.69 ea PO#562752 Vendor Subsem Inc 5/31/2005 Additional work: engineering time to verify cable lengths, make labels, design and make brackets to hold the cables in place - Preamp Power Supply to Outer Shield Box DC Power Cables These cables carry the DC power from the Preamp Power Supplies to receptacles on the outer shield boxes. For this estimate I will assume that there are 2 outer shield boxes and 2 Preamp Power Supplies. There are 2 cables from each power supply: one for the preamp power and one for the filter power. I will assume that these cables are 20 ft long, i.e. that the power supplies can be mounted within 20 ft of the cryostat. Quantity: 4 Preamp Power Cables Materials Cost for each cable: $64.38 total 4 conductor AWG 14 cable Belden #8627 at $2.53 a foot 2 Cinch Jones series 2400 connectors at $6.89 each Assembly time about 40 minutes for each cable Additional work: engineering time to verify cable lengths, make labels, install standard cable clamps to hold the cables in place - Power Distribution Cable Assemblies These are the cable assemblies that take care of distributing the preamp and filter power within the outer shield boxes. These cables assemblies route DC power to the individual PFC-16 cards. These assemblies include: the DC power receptacles for power from the power supplies, fuses, toroidal ferrite common mode noise and ground isolation chokes, and the individual feed lines for each PFC-16 card. For this estimate I will assume that there is a separate power distribution cable assemblies for the preamp power and for the filter power in each shield box. I will assume that power distribution for up to 20 PFC-16 cards in each shield box should be available. It may end up that it is easier and faster to make part of this cable assembly as a printed circuit board. Quantity: 4 Preamp Power Distribution Cable Assemblies Materials Cost for each cable assembly: $183 total 1 Cinch Jones series 2400 connectors at $6.89 5 ft of 4 conductor AWG 14 cable $12.50 120 ft of AWG 22 wire $30 40 fuse blocks at $0.74 each 40 2 Amp fuses at $0.50 each 40 1/2" ID toroid ferrite choke cores $0.87 each 6 3/4" ID toroid ferrite choke cores $1.12 each 20 6 pin connectors with contacts $2.12 Assembly time about 16 hours per cable assembly Additional work: engineering time to design this cable assembly, make labels, design and make the structure to hold this cable assembly within the outer shield box Cards: ------ - PFC-16 16 Channel Preamplifier and Filter Card The PFC-16 card is a combination of the FET preamp section of the D-Zero dual FET preamp hybrid and the filter section of the PMB-16 card that is used on the PAB test LArTPC. The circuit board for this card needs to be designed, the raw boards need to be manufactured, all of the components for this card need to be purchased, and the cards needs to be assembled. A 480 channel system will require 30 of the PFC-16 cards. Before doing the layout for the PFC-16 I want to test the overall design by hand modifying a few channels on a PMB-16 card so that it carries the same functions and design implementations as will be used on the PFC-16 cards. In that way I propose to skip all "test run" stages during the manufacture of the raw PFC-16 cards and in the assembly of these cards. The layout of the circuit board for the PFC-16 can be heavily leveraged from the design of the PMB-16. A test layout of the dual FET preamp section has already been done and it fits within the same 12 mm pitch of the channels on the PMB-16 cards. I expect the full layout of the PFC-16 to take 3 weeks of office time. Work on the actual layout of the PFC-16 will not start until everyone has had a chance to review a written specification for this card. There is the issue of selecting the rc values to determine the characteristics of the filters on the PFC-16 cards. My hope is that this can be guided by work on the PAB Bo LArTPC system. The time and cost to manufacturing the raw circuit boards for the PFC-16 can be estimated from other cards that we have built recently. Most of component costs for the PFC-16 can be found in catalogs. The actual component cost is often dominated by the need to purchase full reels of tape and reel surface mount components. Balancing the cost of needing to purchase full reels are the partial reels of components that we have left over from other assembly runs. Components left from the PMB-16 assembly run will be the most useful. The time and cost to assemble the PFC-16 cards is estimated from building other cards is similar size and complexity. An issue for the schedule of the assembly work is having this project run into the holidays at the end of the year. Quantity: 30 PFC-16 cards for the running system plus 5 or 10 spare cards Cost: $511 per card based on 35 cards raw circuit boards $70 each passive components: resistors 17 values $40 per reel ceramic capacitors 10 values $60 per reel Tantalum capacitors 4 per channel $20 per card electrolytic capacitors 4 per card $2 per card semiconductors (per channel) FET Preamp $ 2.00 Filter $12.75 connectors $6.50 per card assembly cost per card $120 (plus $650 for stencils) Schedule: 3 full weeks in the office of design time for the board 4 calendar weeks to manufacture the raw circuit boards 2 full weeks in the office to purchase parts and setup the build 2 calendar months for assembly - ATC Adf-2 Transition Card The ATC card is a passive card (connectors and traces only) that routes the analog input signals from the Pleated Foil Cables into the input of the ADF-2 cards. The ATC cards mount in the back of the backplane. For 480 channels you need 15 ATC cards to feed the analog signals to the back of the ADF-2 crate backplane. This many spare ATC cards are available because a number of these cards were built with a defective a LVDS connector section that is not used in the LArTPC application. I saved these cards which are useless for running the D-Zero application. Quantity: 15 ATC cards Cost: $0 unofficial loan from D-Zero Schedule: available now - ADF-2 Adc and Digital Filter version 2 There are a total of 20 spare ADF-2 cards. In the LArTPC application this card provides: the ADC, the 2048 stage circular buffer, and the VME readout functions. I must keep 8 spare ADF-2 cards at D-Zero to have a spare of each species of ADF-2 card ready to use if there is trouble with the running D-Zero system. The 8 species come from 4 different analog input ranges and in each range either the "maestro" type or "normal" type (this difference is just the installation of jumpers). There are 3 ADF-2 cards in the PAB Bo LArTPC system. They have talked about wanting an additional ADF-2 card to readout some more stuff, e.g. what trigger fired and other operating conditions. It would be costly and take a lot of time to setup another production run of ADF-2 cards. Quantity: 15 ADF-2 cards are required for the 480 channels Cost: $0 unofficial loan from D-Zero Schedule: 9 ADF-2 cards are available now - SCLD Serial Command Link Distributor This is a card from Saclay with all MSU firmware on it. It controls the synchronous operation of up to 4 crates of ADF-2 cards. There are a total of 3 of these cards in existence. One is the running D-Zero system. One spare at D-Zero. The one "extra" SCLD card is in the PAB Bo LArTPC system. I do not even have one to run stuff at MSU any more. To get around this problem I have built a SCLD_Sub card. This card has all of the functions that are necessary to run a crate of ADF-2 cards in the LArTPC application. It can not run the ADF-2 cards in the D-Zero application (nor is it practical to ever make it do so). Building another one of the SCLD_Sub cards will require a few hundred dollars of components and takes a full week of uninterrupted work. Quantity: 1 SCLD_Sub card Cost: $200 for the components Schedule: requires one very full week of work to hand build a second SCLD_Sub card Boxes, Crates, and Power Supplies: ---------------------------------- - Outer Shield Box For this estimate I will assume that there are two preamp boxes. I have very little understanding of the T962 cryostat and what mounting space is available for the preamp boxes. Thus the managers of this project should use very wide error bars on my "estimate" for the outer shield box. The outer shield box can be built from a commercial steel NEMA box. It mounts right over the readout port. It provides the first layer of shielding and it provides mechanical support for parts inside of it. The design work for the outer shield box consists of designing: how it will mount to the cryostat and cover the readout port the DC power entry the cooling air entry and exit the readout cable exit the Wire Bias Voltage entry the supports for the DC Power Distribution Cable Assemblies the supports for the Wire Bias Voltage Input Filters prepare the mechanical drawings of the work to be done on the box itself and of the various brackets The mechanical work to build the outer shield box and the various brackets that are associated with it consists of: cutting openings in the NEMA box making removable covers for some of the holes making cable clamps and support brackets making the structure to mount the box to the cryostat painting the box Time and cost estimate for the Outer Shield Box Quantity: 2 Outer Shield Boxes Cost: $300 for the basic NEMA box Schedule: Design time 1 week Mechanical work time to modify/build the box and its brackets, 1 week per box. I do not know how this should be split between mechanical people at Fermi and the MSU Physics Machine Shop. The MSU Machine Shop rate is $30 per hour. - Inner Shield Box As above I will assume that there are 2 inner shield boxes. Once again the managers should use wide error bars on the "estimate" for the inner shield box because it is an estimate for an item that does not yet have a design. I believe that the inner shield box is a fully custom made box. It mounts right onto the readout port and provides shielding for the PFC-16 cards and for the signal cables from the feedthrough to these cards. It provides the mechanical support and the major signal ground connection to the PFC-16 cards. Design challenges include how to mount the PFC-16 cards so that you can pull one out without unstacking the others, and how to provide DC power cables, readout cables, and cooling air while still providing the required level of shielding. The actual mounting of the PFC-16 cards (up to 20 of them) can be taken care of with a commercial 6U x 160 mm card file. The rest of the inner shield box can be built around this card file. In the direction that the card can slide in the card file a removable panel on one side gives access to the input cables and ground connection to the PFC-16 cards. Cooling air enters from this side. A removable panel on the opposite side gives access to the pleated foil output cables and to the DC power cables. The cooling air exits on this side. The cooling air entrance is through an air duct that includes 90 degree bends for shielding. The design work for the inner shield box consists of designing: the mounting structure to the signal readout port the internal ground structure for connection to each PFC-16 card the cooling air input duct the DC power cable entry the pleated foil readout cable exit the cooling air exit the 6 sides of the box with two of them removable prepare the mechanical drawings of the work to be done on the box itself and of the various brackets and ducts The mechanical work to build the outer shield box and the various brackets that are associated with it consists of: make the 6 sided box with 2 sides removable make the air ducts make the mounting brackets to the readout port make support brackets for mounting to the outer shield box make the internal ground structure paint the various parts Time and cost estimate for the Inner Shield Box Quantity: 2 Inner Shield Boxes Cost: $200 for the basic VME card file components Schedule: Design time 2 weeks Mechanical work time to build the inner shield box and its brackets and air ducts, 2 weeks per box. I do not know how this should be split between mechanical people at Fermi and the MSU Physics Machine Shop. The MSU Machine Shop rate is $30 per hour. - Preamp Cooling For the estimate of the this part I will assume that there are two remotely located air blowers, one for each preamp box. These are commercial air blowers and use flexible air duct to carry the cooling air to the preamp boxes. In total there is about 250 Watts of heat to be removed from the preamp boxes. The noise of the input FETs comes up rapidly with increasing temperature so there is a significant advantage in providing lots of well distributed cooling air. Quantity: 2 cooling air blowers Cost: $150 for the basic air blower each Schedule: Design time 2 days Mechanical work time to mount the air blower to a panel, connect a line cord and power switch and fuse and panel light, make the adapter from the blower to the flexible air duct, paint and label total of 1 week for both air blowers. - Preamplifier Power Supply for PFC-16 cards The PFC-16 cards will require: -6V and +8V for the Preamps and -5V and +5V for the Filters. They need about 1/2 Watt per channel. These power supplies can be made from commercial linear power supplies mounted in a commercial 19" chassis. An isolation transformer should be included to eliminate the possible need to add it later. Each power supply will actually consist of 2 chassis, one for the preamp power and one for the filter power. This will make the assembly easier and faster. Quantity: 2 Preamplifier Power Supplies Cost: $3,932 total cost for the components to build all of the preamp power supplies bricks for preamp +8V $110 each bricks for preamp -6V $58 each bricks for filter +5V $185 each bricks for filter -5V $185 each isolation transformer 500VA $140 chassis $205 line cord, fuse blocks, fuses, power switch, test points $100 Schedule: Design time 1 week Mechanical work to prepare the chassis 1 week total Technician time to wire all the power supplies 2 weeks - VME Crate A commercial VME crate can be used. For this estimate I will specify two crates in case we need to split the ADF-2 cards across two VME backplanes in order to achieve the required readout bandwidth. Note also that the ADF-2 cards require +-5V supplies on the normal +-12V VME backplane bus. This is accomplished by installation of the proper modules in the Wiener VME crate power supply. The cost and delivery time shown below are the current numbers from Wiener. Quantity: 2 Wiener VME-64x 6U with Wiener +-5 Volt analog supplies instead of +-12 Volt VME supplies Cost: $8,632 each Schedule: 8-12 weeks Computer and VME Interface: --------------------------- - DAQ Computer to VME Interface For this estimate I will specify 2 computer pci bus to VME bus interfaces to match the two VME crates above. The interface is from a company that was called Bit-3 and is now part of SBS or GE or something. See: sbs.com/products/371 This is the standard computer to VME interface that is used in a number of setups at Fermi. This has an optical connection to the VME crate which nicely gets rid of all the grounding issues. Quantity: 2 "Bit 3 Model 618" Cost: $3,800 each plus $120 for the optical cable These numbers are from our last purchase of this equipment about 2 years ago. - DAQ Computer It would be best for those who are doing the overall DAQ software to specify the computer that they want. In any case I assume that this will be purchased through Fermi and setup by the Fermi Computer Department so that it meets all of their security requirements. Software and Firmware: ---------------------- - Software If T962 chooses not to use the existing LArTPC DAQ-96 software then it is my understanding that MSU will only need to provide the low level software routines that directly interact with the hardware. Specifically the routines that MSU would provide are: configure the FPGAs on the ADF-2 cards (i.e. load firmware) initialize the ADF card prepare the ADF-2 card to capture the next event test to see if a new event is in hardware read the data from ADF-2 hardware into a software buffer Obviously the functionality of these routines already exists but work will need to be done to remove the D-Zero environment dependencies from these routines and provide them as generic c++ library. The other known area that will require work is making the event readout run as fast as possible. Philippe and I have discussed a number of ways to improve the readout speed. The best way forward and the amount of Philippe's time that is required will depend on how much faster the readout needs to run. For now my best estimate is that we will need 2 months of Philippe's time to prepare generic versions of the hardware routines along with a framework to test and demonstrate them and to work with me on speeding up the readout. - Firmware As far as I know, the current LArTPC firmware for the ADF-2 card will work at T962. Simple modifications would include things like changing the sampling period from 198 nsec to 250 nsec or something like that. Major modification work would be required for faster VME readout time.