This is an old revision of the document!
Supplied 1PPS and 10 MHz reference, using the cables previously connected to the SRTP2 system. BNC→SMA adaptors were used to make the connection.
Booted into Windows from the onboard PC, configured the network settings to IP 131.217.63.164, NM 255.255.255.128, GW 131.217.63.129. Used “System” in control panel to change computer name (to dbbchb) & DNS lookup to phys.utas.edu.au.
Ran “DBBC control”, answered “y” to reconfiguration. Sampling modules were re-initialised and resumed processing with the 1PPS sync.
1 dbbc2.bit 612.99 4
1 dbbc2.bit 652.99 4
1 dbbc2.bit 752.99 4
1 dbbc2.bit 912.99 4
1 dbbc2.bit 632.89 4
1 dbbc2.bit 752.89 4
1 dbbc2.bit 812.89 4
1 dbbc2.bit 832.89 4
1 dbbc2.bit 325.99 4
1 dbbc2.bit 345.99 4
1 dbbc2.bit 365.99 4
1 dbbc2.bit 395.99 4
1 dbbc2.bit 445.99 4
1 dbbc2.bit 465.99 4
1 dbbc2.bit 319.99 4
1 dbbc2.bit 319.99 4
The 1 at the start says to use the core at the relevant address (1-16), the second line is the Xilinx config file to use (again, there is only one of these available and it is not user-configurable).
The third number is the frequency of the band edge for that module. If the upper band is selected for recording, then the number is the lower band edge. The planned connection for IVS observations is to use RCP only, with X-band into modules A and B, with S-band in C & D. This replicates the Ho setup with VCs 1-8 at X-band and 9-14 at S-band. The file shown is an adaptation of an R1 experiment (R1415 I think). The frequencies for X-band have been increased by 480 MHz (relative to Hobart's VC frequencies) which S-band has been increased by 120 MHz. Cores 15 and 16 have “dummy” values as they are not used at Hobart.
The last number is the bandwidth of the recorded channel.
dbbcifa
Which reports the current setup. Note that the commands into DBBC Control are very laggy - it will take up to 1 second for the commnd to be recognised. When you change settings and then check that it has worked, allow at least 2 seconds between the commands.
The output of the dbbcifa
command should be of the form dbbcifa/1,36,agc,1,35347
where the first number is the filter (where 1 = 512-1024 MHz, 2 = 0-512 MHz, 3=1024-1536 MHz (Currently N/A) and 4 = 1536-2048 MHz (also N/A)), the second number (36 in this example) is the current attenuator setting (which ranges between 0 and 63 where 0 is +16 dB and 63 is -16 dB. If the reading is > 50 or < 10, consider changing the settings on the IF unit first.), agc
indicates that the active gain control is enabled, the fourth number is also the filter (Maybe…), and the fifth is the current 1s averaged total power for the filter. This last value should be . You need to turn off the AGC and replace it with a single value. If the reported value was acceptable, then use in the following command
dbbcifa=1,36,1
Wait 2 seconds and then check the levels with dbbcifa
again. The 16 bit power level should be ~32000 ±8000. If this is ok, then repeat the procedure for dbbcifb
, replacing the attenuator setting with the appropriate level but still using filter 1.
For dbbcifc
and dbbcifd
, these are the S-band channels and should use filter two. The initial reading will use filter 1 and report gains set to 0 and power levels at zero. Set the band correctly using dbbcifc=2,32,2
and dbbcifd=2,32,2
. Wait 2 seconds and then check that the power levels are acceptable. If not, adjust the settings and/or the IF attenuator levels.
dbbcform=geo
This should conclude the set up for the DBBC. Next is the Mark5B which is mercifully simpler and less prone to mysterious faults.
dimino -m0 &
In the second, run tstdimino
OR use the field system to send the following commands. Into the tstdimino
/fs, send the commands
clock_set=32:ext
dot_set=:force
Then check the synchronisation has worked with
dot?
This last command will return a string which should have the correct date and time, with syncerr_eq_0 and FHG_off. Next, set the mode on the recorder and then you can start the recording.
mode=ext:0xffffffff:1
The :1
at the end of the line indicates how the data should be reduced. The default recording mode is 16 channels, at 16 MHz (32 Ms/s), and 2 bits per sample for a 1024 Mbps data rate. For IVS experiments recording 16 channels, the reduction value should be set to 16/bw (i.e, for an experiment using a 4 MHz bandwidth, the mode should be set to mode=ext:0xffffffff:4
.
At this point, you are set to record with the tstdimino
command record=on
. The fs equivalents of these tstdimino
commands should be something like mk5=clock_set=32:ext
going from Hobart.
For IVS schedules using the 12m, there is no support for the Mark5B recorder in fs 9.9.2. When this is upgraded to 9.10, less kludging should be necessary. For now, the skd file is drudg'd with the rack set to none
and the recorder as Mark5A
. The procedure file is then loaded into the field system and edited (using pfmed
from another terminal). The procedures to edit are setupsx
(or its equivalent), preob
and systemp
.
In setupsx
, edit the section to read
mk5=play_rate=data:4
mk5=mode=ext:0xffffffff:2
mk5=clock_set=32
mk5=1pps_source=vsi
mk5=dot_set=:force
bank_check
tpicd
mk5=mode?
mk5=clock_set?
mk5=1pps_source?
mk5=dot?
Some of these entries are probably superfluous or incorrect (especially @@tpicd@@).
preob
should be edited to read
mk=dot_set=:force
systemp
systemp
should read
sy=/home/oper/tsys12m.sh
This last procedure is a link to a kludged bash script to make an estimate of the system temperature using the noise cal and 0.5 dB attenuators of the RF box together with some of the unused samplers in rack 1. The measured Tsys values are currently injected into the log as comments but a better solution would be to write an ANTAB file during the experiment.
Dave Graham provided a current version of the Linux implementation on 2010 May 11. The files were downloaded and placed in a directory on a Linux PC. An empty USB-mounted HDD was used and partitions created to match the first two as described in the READ.ME file provided by Dave. This was done using fdisk
. The first partition was tagged as bootable.
Then ext3 filesystems were installed on each partition using mkfs.ext3
.
The image of the first partition was then copied to the disk:
dd if=dd_sda1 /dev/sdc1
The file boot_sector
was copied to the master boot record:
sudo dd if=boot_sector of=/dev/sdc bs=512 count=1
Then the first partition was mounted at /mnt/tmp
sudo mkdir /mnt/tmp
sudo mount /dev/sdc1 /mnt/tmp
Then find out the disk ID:
ls -l /dev/disk/by-id | grep sdc
Then edit the fstab file in the mounted partition so that the / partition has the ID of /dev/sdc1 and /usr2 is at /dev/sdc2, /data at /dev/sdc3. /etc/fstab then looks like this:
/dev/disk/by-id/usb-WDC_WD75_00AACS-00D6B0_DCA425712418-0:0-part1 / ext3 acl,user_xattr 1 1 proc /proc proc defaults 0 0 sysfs /sys sysfs noauto 0 0 debugfs /sys/kernel/debug debugfs noauto 0 0 usbfs /proc/bus/usb usbfs noauto 0 0 devpts /dev/pts devpts mode=0620,gid=5 0 0 /dev/disk/by-id/usb-WDC_WD75_00AACS-00D6B0_DCA425712418-0:0-part2 /usr2 ext3 acl,user_xattr 1 2 /dev/disk/by-id/usb-WDC_WD75_00AACS-00D6B0_DCA425712418-0:0-part3 /data ext3 acl,user_xattr 1 2
also edit /mnt/tmp/boot/grub/menu.lst and change the disk ID.
The disk can now be booted from.
Use a fast (133x or better) card with 4 GB or more. We're using a SanDisk Ultra 200x 8 GB.
Mounted the CF on a Linux PC via a USB card reader. Card appears as /dev/sde1
As the card is larger than 4 GB I partitioned it to match Dave's card:
disk /dev/sdb: 4009 MB, 4009549824 bytes 255 heads, 63 sectors/track, 487 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk identifier: 0xe7cf73dd Device Boot Start End Blocks Id System /dev/sdb1 * 2 487 3903795 83 Linux
sudo fdisk /dev/sde p d (deleted FAT partition) n p 1 2 487 (added Linux partition) a 1 (made partition bootable) p (should now match Dave's card) w
Then installed an ext3 filesystem:
sudo mkfs.ext3 /dev/sde1
Copy the image onto the CF:
gunzip dbbc_if.gz sudo dd if=dbbc_cf of=/dev/sde
Mount the disk
sudo mount /dev/sde1 /mnt/tmp
Edit /mnt/tmp/fstab
and change first line to read
/dev/sda1 / ext3 noatime,acl,user_xattr 1 1
Edit /mnt/tmp/boot/grub/menu.lst
and replace references to the disk partition to /dev/sda1
:
# Modified by YaST2. Last modification on Fri May 14 13:24:28 UTC 2010 default 0 timeout 8 gfxmenu (hd0,0)/boot/message ##YaST - activate ###Don't change this comment - YaST2 identifier: Original name: linux### title openSUSE 10.3 root (hd0,0) kernel /boot/vmlinuz-2.6.22.5-31-default root=/dev/sda1 vga=0x31a splash=silent showopts initrd /boot/initrd-2.6.22.5-31-default ###Don't change this comment - YaST2 identifier: Original name: failsafe### title Failsafe -- openSUSE 10.3 root (hd0,0) kernel /boot/vmlinuz-2.6.22.5-31-default root=/dev/sda1 vga=normal showopts ide=nodma apm=off acpi=off noresume nosmp noapic maxcpus=0 edd=off 3 initrd /boot/initrd-2.6.22.5-31-default
Unmount, eject and boot on DBBC!
Tim has been having trouble getting the 8 GB card to boot so I got a 4 GB SanDisk Ultra 200x card.
Partition table initially showed
4011 MB, 4011614208 cylinders 128 heads, 63 cylinders/track, 971 cylinders units = cylinders of 8064 * 512 = 4128768 bytes
so in the “extra functionality” mode of fdisk I changed number of cylinders to 487 and number of heads to 255
partition table then could be made to match Dave's.