List of assumed responsibles who need to provide information: Tracker POWER Blasko/Menicelli ToF+ACC POWER, HV Castellini/Cindolo TRD POWER Schwering TRD GAS,HV - Koutsenko RICH POWER, HV - Marin ECAL POWER, HV - Cervelli SRD POWER, HV - Viertel DAQ POWER - Chang other - Chang (not USCM) CAB - Mike (not USCM) CCOOLER POWER +++ - Viertel, Foster PDB POWER - Viertel MCB T,B,... - Viertel LaserAlign - Wallraff Other people receiving this mail are welcome to contribute, but it if more "for your information". Please reply by 18 Sept 00 to Mike Summary of Discussion on Monitor and Control Electronics 8 Sept 00 at ETH-Zurich M.Capell/MIT * Source: URL=http://ams.cern.ch/AMS/Electronics/Docs/000908_Capell_MC.txt ------- * Present: C.Camps, M.Capell, V.Commichau, V.Koutsenko, -------- A.Lebedev, U.Roeser, G.Viertel, Hp.vonGunten * CAN BUS STANDARD: ----------------- All CAN nodes in AMS 2 should conform to the CAN Standard 2.0B. If you are NOT planning on using a USCM module as the interface to the AMS CAN busses for monitoring & slow control, please forward the part number of your CAN controllers to Mike. ALL CONTROLLERS MUST BE 1) APPROVED BY MIKE "NOW" 2) TESTED BY LEBEDEV "LATER" OR THEY WILL NOT BE ATTACHED TO ANY AMS CAN BUS. * USCM & GSI Beam Test: --------------------- After discussion, it is considered critical to place several elements which are central to the USCM design in the beam test. Volker et al can provide the semi-operational board(s) plus software which runs on a PC. However they cannot support the test in person. Volunteer(s) are solicited. Possible items to test are (in rough order of priority): the dallas microcontroller (DS80C390, which includes the CAN controller), the MAXIM latch up protector, the dallas temperature sensors (DS1820?), the ADC & DAC. * Dallas Temp Sensors: -------------------- The three wire (Power, data, return) connection will be used. Under certain failure modes the 1(2) wire protocol will be used -- but all planning should be for the 3-wire mode. * Power & the USCM: ----------------- The boards will be relatively low-power (c. 1.5W?) and they >> WILL NOT DRIVE LARGE CURRENTS <<. Applications which require significant currents (eg PT100's, Lasers, B-field measurements) should plan on an additional "front end" signal conditioning board which is specialized to their particular needs and which draws power from, for example, the usual crate power supply voltages. Mike's proposal to slip a dedicated 28V to 5V DC-DC converter on each USCMs that needs it (for example, that controls the crate power supplies) was not warmly received but he considers that this will be reviewed later. However a dedicated converter located off in the crate power supply would be fine with Volker. * Redundancy & the USCM : ----------------------- The core elements of the USCM take about ~1/4 of the available AMScard (double eurocard) space. A key concern is to match the physical space for I/O - that is the edge length available for connectors - to the capabilites of the proposed microcontroller (which are much more than those of the one used in the RWTH MCUs in AMS 01). Instead of having "redundant" (1 hot, 1 cold: "1H1C") circuits on one PCB we adopt the working assumption that we will have "redundant" (1H1C) boards at each location and that each such board will have larger than Mike-at-any-rate anticipated functionality. Mike will update his <

> to reflect this working assumption. USE OF BACKPLANE CONNECTIONS - WITH VME CONNECTORS AS IN AMS 01 WILL GREATLY ENHANCE THE CAPABILITY OF THESE BOARDS. CABLE ONLY BOARDS WILL HAVE REDUCED FUNCTIONALITY JUST BECASE OF THE NUMBER OF CONNECTOR PINS THAT CAN BE ATTACHED. ALSO THEY ARE CHEAP. Care must be taken to allow both parts of a USCM pair to independently monitor and control whatever it is they are supposed to control. "Cross-strapping" at the "front-end" is higly recommended, if not practical then duplicate, colocated, for example, sensors can be used - we just have to review each case. For example: to measure the current on a power line, which is unique, we need to fanout to each USCM the corresponding voltage drop - we can either use two resistors or one resisitor feeding two amps (each of which then feeds an ADC on a USCM board). In contrast a dallas temp sensor chain can only be plugged into one USCM - so we will have to pair up these sensors. * Another working assumption: --------------------------- Practically every box with electronics (see list below) will have a USCM pair controlling the power distribution within the box (as well as a host of other functions -- we hope). The usual tortuous discussion about applying (eg 28VDC) power when the box is "too cold" ensued. Here is the working resolution: 0) We assume that survivial limits will be maintained per box. The USCM will use extended industrial grade temp range parts: Operational -40 to +85C/Storage -?? to +??C. 1) Each box will be equiped with a (redundant pair) of PT100-like thermosensor(s) which cabled to some mysterious ultrareliable box. Based on the PT100 readout the 28VDC can be applied to the box. [Actually, I'm not sure this is needed in this scheme but lets leave it in for now and see how big a headache it becomes]. This array of thermosensors can be operated with very low power and high reliability - for example when we are required to be in absolute power minimization mode. [Well, I guess we need something like this.] 2) The 28VDC is connected via simple, reliable (expensive and long delivery time, but tested in mag field and with which we have experience) thermostats to (eg kapton foil) heaters and the USCM (and it's DC-DC cvrtrs) - such that in the temp band between the lower survivial limit, T0, and the lower operational limit of the USCM (and it's DC-DC cvrtrs and its local temp monitors), T1, only the heaters are powered. Above T1 the thermostats allow power to flow to the USCM, which then takes care of further thermal control (ie powering other stuff in the box only when they are within their own operational ranges). * USCM Modifications from V1.2 (28 June 00, attached or see ---------------------------- URL=http:/ams.cern.ch/Electronics/Docs/000628_Commichau_uscm2.ps,.pdf) Concerns were expressed about the implementation of the service module ("mezzanine") concept - how do we get the heat off, does it really save PCB real estate, is it needed, how will the software manage the different combinations, etc. Looking at the functions required we feel there is some hope that a "universal/unique" USCM board can fit most all the required functions - albeit with a high degree of over kill in any given location. As this won't cost much power or weight (as long as the total number of boards doesn't mushroom) this should be workable. So: Starting from a board with one microcontroller plus required peripherals (memories, transceivers, etc) Volker will estimate how many channels of each USCM function can be provided per board - given the constraints of the processor, connectors cm**2 on the PCB, etc, etc. At the same time: All potential users need to at least estimate how many of which type of function they need in each location. I have included my first guess at functions below. Would the responsibles listed please provide at least a first guess at the TYPES & QUANTITIES by 18 Sept or respond when they will be able to provide any better guess (or respond asap who is the responsible).... Here is the summary of the known CAN slaves with location, function, and power supply. Possible "brothering" indicated by "(*2)". "#" indicates USCM candidiate - this is extracted from URL=http://ams.cern.ch/AMS/Electronics/Summary.txt however, given the extended functionality of each USCM, the function of each module should be extended... # 8(*2) TPC 1/T-Crate : Tracker electronics power supply: 28VDC. # 2(*2) UHVC 1/U-Crate : TRD HV: Crate LV. # 2(*2) UPC 1/U-Crate : TRD electronics power supply: 28VDC. # 1(*2) UGBS 1/UGBS : TRD Gas Supply: 28 VDC or UPS # 1(*2) UGBC 1/UGBC : TRD Gas Control: 28 VDC or UPS # 4(*2) SHVC 1/S-Crate : Scint HV: Crate LV. # 4(*2) SPC 1/S-Crate : Scint electronics power: 28 VDC # 4(*2) RHVC 1/Rx-crate: RICH HV: Crate LV. # 2(*2) RPC 1/R1-crate: RICH electronics power: 28 VDC. # 2(*2) EHVC 2/E-Crate : ECAL HV: Crate LV. # 1(*2) EPC 1/E-Crate : ECAL electronics power: 28 VDC 4(*1) JIM 1/JMDC : Main DAQ Control: Crate LV. 2(*1) JFOM 1/JFOM : ?? 1(*1) J422 1/J422 : ?? # 2(*2) JPC 2/J-Crate : DAQ, T/T electronics power: 28 VDC 1(*2) CAB 1/CAB : Cryomagnet: TBD. # 2(*2) CCDB 1/CCDB : Cryocoolers: UPS. # 1(*2) PDB 1/PDB : 28VDC distribution via SSPCs: UPS # 4(*2) MCB 1/MCB : Global Sensors (T,B,...): UPS # 1(*2) LAB 1/LAB : Alignment Laser: Crate LV. # N(*2) # 1(*2) For a total of 50+N slaves, or 93+2N with brothers. Note that the Dallas temp sensors are easy to implement so feel free to use them. * USCM Functions: --------------- Note: If you have a prefered voltage level (TTL, ECL, etc) please mention. However you'll have to live with whatever Volker implements. a) Binary Output (a binary level from the USCM) b) Binary Input (a binary level to the USCM) c) Dallas Temp Sensors - number of chains * sensors/chain. d) Analog Output (a variable level from the USCM, specify number of bits required, Volker will tell you the voltage range later). e) Analog Input (as d), 10bits no problem) f) Pulsed Output (?) g) Pulsed Input (?) h) Other (eg RS-xxx) - specify (one RS-232 is likely "free" per USCM). Remember: the USCM is NOT going to be driving a lot of current on its outputs. You'll have to do that yourself on a seperate front- end board. However, if the user base (you) requests it it can possibly be configured to do pretty unusual things. So tell us what you need - not what you think we want to hear or what you think we (Volker et al) can deliver.