The Unified Board (Version 5)

1 Introduction

In production phase, this mother board is replaced by a denser single board containing :

• a Power Supply system
• the CPU core (CPU)
• the Time Tagging (TG)
• the Slow Control (SC)
• an interface for the GPS, FE and LED Flasher modules
• Front Panel connectors
• The conception should allow easy production testing and mounting, including possibility for future extensions. It must give high reliability for 20 years of life with minimum maintenance in difficult environment ( -20 to +70 degrees Celsius with high day variation, humidity due to water staying on salt ground and water tank proximity)
• The mean consumption must be closed to 10 W

• Power PC 403GCX 40/80MHz - 32bit data and address - cache running Microware OS9000 (Unix-like Real Time)
• 32 Mo EDO DRAM
• 4 Mo in system programmable Flash memory
• An in system programmable PLD for flexibility
• DMA channels are used for fast FE data transfer
• A 4 serial ports peripheral is used for communication with other devices (GPS, Telecom, TPCB, external test device).
• A connector is provided to connect an Ethernet interface for software development and board tests.

2 Mechanical Constraints

3 Electronic Hardware

4 The Proposal

Finally a fourth single (34 pins) HE10 connector will provide the CPU bus signals for the Ethernet card (ETH). A bus buffering interface will assure the reliability, the EM safety handling and the plug-in compatibility, between the ETH card and the CPU. This may be done only for pre-production. Then Ethernet will be deported with this new design, as consequence stability with Ethernet will be lower. We must keep in mind, that Ethernet was provided principaly for laboratory software development, and test. For a simplest way of communication and programmation for new UB, and for operations in the field, we will use in the new UB a serial port with a PPP protocol communication. Note that Ethernet connector will be not provided on the front panel connector, but only on the UB.
 

Figure 1 :Unified board electronic hardware

5 Power supply

*After several cost estimations, turns out that we will need to make a compromise between the homemade and the industrial solutions. In that sense, for the whole digital circuitry we are planing to use the same homemade concept, nevertheless we must minimize component dimensions (capacitors, shelf etc). For the analog section, as its power supply quality requirements are lot more restrictive, we are planing to use a of the shelf industrial-crafted switched power supply.

*Using shelf industrial-crafted switched power supply, for all the supply. This solution is better for reliability, but cost is a little higher.

*Using shelf industrial-crafted switched power supply for 12, +3V3 digital and +5Volts sections and switched regulator low power (Linear) for +/-3V3 analog section.
 

Power system protection :
 

In the Unified Board, we plan to have a electronic basic circuit, to shutdown the power supply on the UB during a specified time (closed to 3 hours) after had sent an alert to the CPU (during one minute), if an under or voltage occur, or if temperature on the UB is too high.
 

For a complete information on power supply, see reference [1].
 

6 FRONT END INTERFACE PIN-OUT AND DEFINITION

Local bus signals connector C1 - 2.54mm - 3x32 - type DIN 41612 - male on UB, female on FE

Table 1. FE connection list.
 

 Figure 2 : specifications for Front End board dimensions 
 
7 Connectors 

7.1 The front panel connectors are as follows:

2) Console Serial - DB9 male

3) TPCB serial - DB9 male

4) ``Spare'' serial - DB9 male

5) PMT Power (x3) - DB15HD female

6) PMT anode and dynode inputs (SMA female, on front end, 6 total)

7) LED driver outputs (SMA female, 2 total)

8) Sensor cable - DB15HD female (with water level, air temperature inside or outside, external trigger input)

9) Power connector - two pins per B. Courty

10) Grounding plug on box exterior controller software
 
 
 

7.2 UB to PMT Connections

For the EA the SDE group decided to use a DB15 HD female for each PMT. This connector will transport the following signal:

• +3.3V analog section voltage ; GND
• -3.3V analog section voltage ; GND
• +12V high voltage supply ; GND
• DAC (high voltage) control signal ; GND
• High Voltage (voltage) read-back ; GND
• High Voltage (current) read-back ; GND
• Temperature Sensor ; GND

Concerning the weather protection treatment, the PMT will be definitively attached to its base by soldering and finally the whole set will be potted. Once the PMT, the base and the cables are tested, the set will be potted together for good.
 
 

8 LED FLASHER :

• For the pre-production we plan to have a daughter board on the UB for Led Flasher command unit.
• For production, we need to discuss about a possible integration of the Led Flasher command unit in the UB.
• During the September SDE workshop, we have decided to two led at one location offset from the center toward a hatch cover, so that the LED can be repaired after liner installation if necessary.
• Connection between UB and Led Flasher is a HE14 type connector female
• 2 SMA connectors are used the connection between command unit and the 2 LED's.

 

 

We are studying two solutions in parallel, the ASIC solution, that seem working well, and a backup solution consisting of the use of PLD ACEX. For pre-production, we plan to have the two footprint (one for each solution) on the unified board. The final choice, then, should be made for the production after a full test of each solution.
 

10 GPS receiver :

• For production, GPS module will be soldered on the UB.
• A buffer is used between GPS and electronic of the UB.
• HE10 connector on the UB will provide to access to GPS signals.

For the Unified Board, the list is reduced to essential sensors to achieve 16 channels of ADC, to simplify the unified board slow control section. The TPCB does all solar power monitoring.

• PMT temperature (× 3)
• PMT HV monitor (× 3)
• PMT current monitor (× 3)
• Unified board temperature
• Unified board voltages (× 4)
• We will probably add measurement for 3 other voltage on the UB and the current input of UB
• 2 spare (water temperature + water level?)

 

 

12 Slow Control

• All sensors and monitoring for Slow Control are described above
• J.Beatty purpose to reduce the multiplexer capacitance
• 16 ADC inputs (14 + 2 spares)
• 4 DAC outputs
• 4 digital outputs (3 for addressing LEDs + 1 spare)
• Impedance input for each way is set to 10 Kilo-Ohms
• We use a buffer for DAC outputs
• ESD protection are provided on ADC inputs (protection against electrostatics discharges)

13 Test equipment and procedures:

L.A.L. laboratory (Orsay - France) is designing a test bench for all parts of the unified board. The Front End board, and the LED flasher will be tested separately. The status of the TPCB is still pending.
This will be done by connecting a PC with Linux system and ADC/DAC interfaces and using a specific electronic board to test interface between UB and FE. We must test all bits and signal (DMA, chip select...) on this interface with a simple electronic design. The software test will be running in both controller and PC with exchange information via Ethernet and/or the terminal serial port.
The test bench, is in a first step designed to test in a same time only one UB. A multiple test in the same time, will be study after this first step (maximum 5 UB tested in the same time).
We plan to have a complete test in temperature only for prototypes and samples of pre-production and production, to study the performance of the UB ( Tests with the test bench connected to the UB, with temperature cycling between -20 and +70 Deg / C, during 48 hours at least).

For the rest of the production, we plan to do a burning, with UB powered on, and a simply test mode (without a complete test bench in the oven).
We plan to have test bench in the factory, in the PCC laboratory, LAL laboratory and in Malargue in a first time.
Temperature range should be set between -20 Deg/C and 70 Deg/C
For Burn in, we plan to do 3 cycles with temperature cycling between -20 and +70 Deg/C during 6 hours, fallow with 16 hours at 70 Deg/C (This is the purpose of the Front End team). Then, the total time for this burn-in is 22 hours.
The first test bench should be ready for April or May 2001.

The principal components to test are :

• Micro-controller must bootstrap correctly
• Power supply : 4 voltages are monitored directly by Slow Control (battery input , 12V, 5V and 3V3 digital. We will probably add the measurement of the + and -3V3 analogical section and of the GND). We can use the power supply connector for Front End to allow ADC measurement for the other supply (+ and - 3V3 analog)
• Ethernet or serial port : at bootstrap, the CPU tries to find Ethernet board (if we use it), or serial port, load the OS9000 software test from the PC. In the first step, our decision is to work for test with a serial port (perhaps with PPP protocol).
• Serial IOs : just by connecting RX-TX and RTS-CTS on each RS232 port
• Serial port GPS : Rx/Tx, 1PPs; We use an HE10 connector for access to GPS signals
• Slow control DAC and ADC : we need to measure the DAC outputs (3 to control HV PMT bases, 1 internal output trough an HE14 connector for LED baseline). Our purpose, is to have an electronic board with 3 continuous amplifier (*2) with 2 outputs. The amplifier output, then is connected to the input ADC on the Slow Control used for PMT high voltage monitoring. Then, with the same scale, if we write data on the DAC, the read on the ADC must be the same (if note, we probably have a problem of bit, or linearity...) We need to use a free input ADC to test LED baseline DAC output. We need to have 3 temperature sensors on the electronic test bench board to test ADC inputs for PMT temperature measurement (we only want to test that teh measurment chain is good). Connectors used on the test bench must be compatible with Unified Board connectors. We need to test also the 3 PMT current monitoring with tested the related ADC inputs. This can be done by applying a continuous voltage on this 3 input.
• For the LED Flasher module, we need to measure the DAC output and the 3 logical outputs from Slow Control towards the LED Flasher.
• For the test bench we need an external power supply controlled by PC with a sufficient power, to provide between 18 and 30 Volts continuous.
• To provide the interface Front End test, we plan to use one PLD board (Given by Zbigniew) with is config file. Then, we should build an interface board for this PLD board (we don`t need the analog Front End board). We plan to have, on this interface some ADC to measure the +/- 3V3 voltages for analog section. We should use this interface board, to put on all the electronic needed for the test bench.
• To test the FE interface, we need to use the calibration routine, write and read data in PLD, we need to test the DMA...
• To test the Time Tagging, we can generate an external trigger, and 15 micro-seconds latter go to read the Time Tagging.
• To provide this tests, GPS (UT+ by MOTOROLA), should be mounted (deported) on the test bench bench. We don`t need the antenna GPS to provide this test.
• Having a connector for the FE board seem preferable than a soldered board, for tests, cost, and reliability (If not, we need to have a special connector for test the interface with FE). This is an open question!
• All the access routines for the hard and communications functionnality should be present in the flash (or at least in the Ram Disk) before the test. All OS9 usefull files, will be given by Laurent GUGLIELMI, to enable the development of the software for the test bench by the LAL team.
• We must have a look on the electromagnetic compatibility. At least, we need to use a simple metallic box to provide a good EMC. We should fallow specifications given by the industrials...

14 EMC considerations

The unified board will be located on top of the large hatch cover (hatch 1) mounted in an inner metal enclosure. The whole set will be placed into a plastic weather proof box. The two main issues, concerning this enclosure design, are:

1. Permits the signal and control cabling to reach the inside of the tank, but keeping light and environmental sealed.
2. Avoid any RF and/or LF noise interference on the UB.

To minimize the virtual capacitance between the box bottom plane and the UB the UB ground plane must be tied to box at least every 100 mm. The same scheme must be adopted for the FE - UB fixation with brace brace, so, the longer side would have 3 contacts points and the shorter side 2 contacts.

Note: We must pay attention to the mechanical assembling quality of the contact points between ground plans, otherwise the LS could suffer of a capacitive coupling with the environment!!!!!

To avoid any RF frequency on the UB, with passing cables, all connectors will have their metallic part grounded and electrically connected to the front panel.
 

To protect electronic against electrostatic discharges (ESD), we will use some protections components against ESD (the SurgX family by COOPER). We want to protect in&outputs connected to external signals source.
 

Some safety considerations :

A floating electric powered equipment is a safety issue. The tank accumulates electrical charges, and the electrostatic potential increases, becoming an electrical chock hazard for humans and electronic circuits. Moreover, the tank is not lightning protected. As a full electrical protection of the tank, i. e., a high quality ground electrode (capable to be the Earth or Safety ground), a lightning rod and its appropriate cabling, a lightning protection for the radio unit, a metal grid or a conductive paint all over the tank, would be an expansive and complex issue for installation and maintenance, we are not considering this possibility, unless as an ultimate solution. What we are doing by now, is bet that a floating tank can remains out in the pampa for good, working fine. In that sense we would to point out some preventive measures, once working in tanks electronics:

1. Use an anti-electrostatic carpet, connected to the zero volt reference (solar panel brackets).
2. Use an anti-electrostatic bracelet, connected to the zero volt reference (solar panel brackets).
3. Be careful during the cabling, avoiding large cable loops in order to prevent large inductive areas and a pick-up coupling. providing an indirect-impact lightning protection.

Quality insurance :

• In our public tender we have asked for the ISO 9000 quality insurance (or equivalent), technical and man power resources, tests procedures, production logging system and boards backtracking.
•  The boards reliability is defined by the MIL-HDBK-217F standard (ground fixed conditions)
• The MTBF index is determined and must be equal to the Auger life time.

16 References

[2] http://cdfinfo.in2p3.fr/experiences/Auger/work.html: ``Local Station Controller User Manual''.

Retour