|
..........
 |
Stand alone/network diagnostic instrument
The Beam Current Monitor is in a development stage with a Design Review scheduled for June 2001. This interface control document (ICD) represents the current design concepts and present plans for capabilities.
The Beam Current Monitor is a PC based instrument, which acquires data from Beam Current Transformers placed at various locations in the accelerator. Its purpose is to measure the longitudinal current profile of the beam. It supports both local and remote operation via a local console and EPICS (Experimental Physics and Industrial Control System) channel access.
A beam current transformer sends an analog representation of the current to the electronics located remotely from the pick-up element. The beam current passing through the transformer is adjusted for signal level, digitized, conditioned to reduce droop and restore the base-line, averaged, integrated, and stored locally in a circular buffer capable of holding a minimum of 10 macro-pulses. These data are available for up-loading to the control system. A minimal set of data is provided on a regular basis for normal operating mode, while more detailed data are available by specific request commands. Current integration and measurement is internal to the instrument. In addition, a time stamp is generated to define the macro-pulse data. Each instrument will provide current monitor electronics for two current transformers.
The instrument is a stand-alone device with a few exceptions. External timing is required to synchronize the data acquisition with the accelerator macro-pulse and mini-pulse structure. A separate calibration subsystem provides a known standard current pulse that is routed to one of many current monitor instruments for calibration purposes. We need to talk about this.
The beam current monitor will provide the following capabilities
1. Detailed time history within the macro-pulse
2. Droop compensation and base-line restoration
3. Average current for each mini-pulse/macro-pulse
4. Total charge for each mini-pulse/macro-pulse
5. Time stamping capable of defining a single macro-pulse
6. Provision for external high speed acquisition rate digitization
7. Manual local/remote gain adjustment
8. Manual local/remote calibration using a separate calibrate subsystem discussion
9. Redundant? Local/remote system testing using an external calibration system discussion
The time history within a macro-pulse will be provided from a digitized version of the analog signal. The digitizer is planned to operate at near 65MHz and will yield 14 bit resolution including sign. To present this data (2 bytes/sample x 65000 samples/pulse x 60 pulses/sec = 7.8MBytes/s) a great deal of communication bandwidth is required. Therefore, the detailed data will be stored in a circular buffer for read-out on demand only. The circular buffer is planned to store only about 1 second worth of data.
The current transformer is designed to provide a rise time less than 1ns and a droop of 0.1%/us. With a mini-pulse of 1 ms duration the transformer will have drooped to near e-1 or 0.37. This droop must be compensated, and the baseline restored to zero to properly average and integrate the waveform to compute charge. The instrument will provide this compensation.
The compensated digitized data will be averaged over a single mini-pulse/macro-pulse to yield a single number for each mini-pulse and the macro-pulse. This data will be available for up-load to the control system at a one second refresh data rate.
The compensated digitized data will be integrated over each mini-pulse and the macro-pulse to yield a single number for each mini-pulse and the macro-pulse. This data will be available for up-load to the control system at a one second refresh data rate.
The instrument will provide a time stamping capability. The Beam sync clock will provide timing information to accurately define time to 1/16 of a revolution. This reference (an external signal input to the instrument), together with an event marker, will provide the information to place a unique time stamp with resolution sufficient to define the single macro-pulse. In addition, by counting mini-pulses after the event marker, one can define the individual mini-pulses.needs clarification
An output will be made available for a signal to be externally processed by a high-speed data acquisition system.
For accurate measurement of the current the amplifier gain must be adjusted to maximize signal to noise ratio and achieve the dynamic range required by the particular sensor and location. Due to the large dynamic range of the input signal in the Ring, the gain must be changed as the input signal increases. The instrument will store a default table of gain settings. This table describes the gain required on a turn by turn basis, and includes information to change gain (at which turn the gain must change). Gain settings can be changed locally or remotely. The MEBT, Linac, and HEBT do not require gain changes. The Ring requires as many as 7 gain changes. The RTBT, although not requiring gain changes, will operate with two gain selections to accommodate a single turn signal level for testing and the normal 1ms accumulated turn case.
Again, we need to talk about this For accurate measurement of the current the system must be calibrated. A separate calibration source will provide a known current pulse to the system and will stimulate the transformer. Presently, the calibration system is conceived as a device that would be switched to individual transformers. This reduces the number of calibrators required for the entire accelerator. The calibration pulse will be timed to occur during the intra-pulse period. The calibration system will be programmable to achieve signal levels that can exercise the system over its dynamic range thereby checking all amplifier gain paths. Data collected during this calibration pulse will be stored within the instrument and used to adjust the data before it is uploaded to the control system.
The same type of current transformer will be used in all areas of the SNS. Approximately 27 current transformers will be installed. The present distribution is shown in Table 1.
Table 1
|
Location |
Size (ID) |
Quantity |
|
MEBT |
5.5cm |
2 |
|
DTL |
* |
6 |
|
CCL |
* |
2 |
|
SRF |
* |
6 |
|
HEBT |
13cm |
5 |
|
Ring |
22cm |
1 |
|
RTBT |
22cm |
5 |
|
* |
2.5cm, 3.0cm, 8cm Not finalized |
All Current transformers are Bergozβ Fast Current Transformers (FCT). They are designed to provide less than a 1ns rise-time with a droop of less than 0.1% per microsecond. The transformer has a 50 turn winding, and is equipped with an auxiliary winding of 10 turns that must always be terminated in 50 ohms for proper operation. This will provide the capability to follow the chopper edge, if necessary, for chopper diagnostics. The auxiliary winding can be used for calibration purposes. The droop will be digitally compensated in the instrument.
Channels will be grouped according to common region. Typically, each instrument provides electronics for two transformers. Racks will be located so that cable runs are not longer than 100 meters.
The intelligent current monitor instrument is comprised of a personal computer (PC), that houses the instrument PCI card. On the card is appropriate logic to interface with the PC, and utilize direct memory access (DMA) to store data in real time. The system architecture is shown in Figure 1.
 |
Shown in Figure 1 are two beam current monitor instruments on a communication line that permits data transfer to a VME collector for correlation with other detectors, and a workstation for data manipulation and presentation.
Generally, arrays of data will be acquired which are sufficiently long to encompass the start of the beam macro-pulse, the end of the beam macro-pulse, the calibration pulse and a sufficiently long section of pulse tail to calculate the transformer droop time constant The actual array length will depend on the sampling rate, the beam pulse width, calibration pulse timing and required sample of tail signal. It is estimated that a total of 3 to 4 ms of data will be taken for each macro-pulse.
Data are acquired and stored in the PC using DMA transfer from the system PCI circuit board. A circular buffer is generated in the PC to store 1 second of discontinuous data. Data are processed to provide the capabilities mentioned earlier. In addition, data taken during the tail of the macro-pulse, after the 1ms macro-pulse has been acquired, will provide information associated with any calibration pulse used and an ability to calculate the transformer droop constant. Data recorded just prior to the macro-pulse will be averaged to determine any DC offset adjustments required to restore the baseline to zero. This data is analyzed within the instrument. The calibration pulse provides self test capability. The tail is analyzed by a least square fit routine to fit it to an exponential. Appropriate adjustments are made to the droop compensation algorithm to correct for observed changes. An appropriate time stamp will be generated by the instrument to correlate the data set to the macro-pulse.
Provide the:
Average Current for the last 1 to 60 macro-pulses, 1Hz update rate.
Total Charge for the last 1 to 60 macro-pulses, 1Hz update rate
Current vs. time (10usec, 100 values) for last macro-pulse. One point for every 10 mini-pulses.
Current vs. time (1usec, 1000 values) for the last macro-pulse. One point per mini-pulse.
Current vs. time (16ns, 60 values) for a single mini-pulse. 1Hz update rate
Each of the above is for one detector. If data is correlated in a VME then the data for many detectors will be reported together.
Install NT, Labview and Channel Access Server on a PC
Assist in debugging hardware.
Assist in definition of Data format
Assist in testing RTDL, Event Link Interface. Need prototype timing system setup.
Setup VME to read BCM. ( Global Controls Interface.)
Determine VME data format.
Write application software to display data from BCM.
Write application software to display data from VME.
To properly load and query the instrument a number of variables are required to be transmitted to the instrument and expected data received from the instrument. A list of presently established variables is shown in the Control Channel Listing below.
NAME
DESCRIPTION
Gain Array
Array containing gain settings
Calibration table
Table coefficients defining gain multipliers
Operating Mode
Mode numbers describing the desired data to up-load
Data type Request
Defines avg current and charge format
Data Pulse number
Requested macro-pulse time stamp
Turn Number
Requested mini-pulse time stamp
Gain Array: This channel defines gain setting array to be down loaded to the instrument. It contains an 8 bit word for DAC control, an 8 bit word for a DAC register address, and amplifier switch coding. It will define the turn vs. gain relationship.
Calibration Table: This channel defines the calibration array. Normally this array will be stored in the instrument in a local calibration file. This file is loaded into the AFE electronics upon start-up. After a calibration is executed, data will be compared to the calibration data and an appropriate new table down loaded to the instrument.
Operating mode: This channel defines type of pulse expected. Special single bunch pulses, low power pulses, or full power pulses will be definable. This will permit alternate gain tables to be loaded in preparation for the pulse expected.
Data type: This channel defines the desired data requested for upload. This is of particular importance for data requests for time histories.
Turn number: This channel defines the turn of interest for requested data, and can include a single turn, a number of turns, or a full dump of the circular buffer.
This section reflects a target schedule to satisfy a requirement for MEBT commissioning. Ring electronics require gain changing that is not required for the MEBT, Linac, or HEBT. Our experience with MEBT operation will be reflected in things learned and will influence the final design effort for Linac, HEBT, Ring and RTBT.
For MEBT:
Prototype system is needed on or about OCT 2001
There will be only 2 BCMs so the VME system is not needed.
VME will acquire data from several systems and combine data according to time stamp.
It is estimated that the timing systems will be ready in the May, June 01 time frame. Testing must wait until then.
BCM Displays will be available.
VME Displays will not be available.
Need mebt schedule here, too
General Schedule:
BCM PDR? 01 JUN 01
Deliver BCM to Site 21 JUL 04
BCM Install LINAC 15 APR 04 to 17 JUN 04
BCM Install HEBT 11 JUN 04 to 16 AUG 04
BCM Install RING 21 JUL 04 to 22 SEP 04
BCM Install RTBT 21 JUL 04 to 22 SEP 04
BCM Installation Compl. 22 SEP 04
BCM Comp. Test @ Site - LINAC 15 JUL 04 to 29 SEP 04
BCM Comp. Test @ Site - HEBT 17 AUG 04 to 01 NOV 04
BCM Comp. Test @ Site RING 23 SEP 04 to 10 DEC 04
BCM Comp. Test @ Site RTBT 23 SEP 04 to 10 DEC 04