Frontside of a picoscope server (dal024):

Mounting the Hardware

• Usually Someone from IN mounts the Pexaria card and connects the ethernet cables / the Pexaria LEMO .. here some details on that:

• The following two figures explain how-to install a PCIe card into a SuperMicro after the top cover has been removed.

Figure: Remove three screws (red) and pull the handle (blue).

Figure: Unplug the frame and mount PCIe modules.

Finally, reinsert the frame, close the handle, insert the three screws and reinstall the top cover.

Here a "Labeled" Pexaria card:

Please note that the Pexaria card must be registered by Timing, BEFORE it gets connected to the Timing Network !!

Management Port (IPMI)

• Name of Managenment Port always is either <FECNAME>i64 or <FECNAME>i
• If management port is not reachable after first boot, unplug + replug of RJ45 should help.
• Reboot/Shuitdown/Console can be used on this website: <FECNAME>i64.acc.gsi.de -- Ask Alexander Schwinn for Password
• Only the left ethernet port seems to work
• This one is the management Port:

Figure: SuperMicro Ethernet ports.

• Uhhh, its not quite as simple; when only one Ethernet port is connected the following will happen
  • LAN Interface 'Failover' : all ports work as 'Management Port'; you can only reach the IPMI via <FECNAME>i64 (or <FECNAME>i)
  • LAN Interface 'Dedicate' : ports behave as indicated in the figure 'Management Port' above
  • thus, check (and configure) the desired behaviour via IPMI -> Configuration -> Network

Figure: SuperMicro Ethernet port config.

• You can modify or even factory reset the management software via IPMICFG-Linux.x86_64.

Booting via PXE (Ramdisk) and Flashing Gateware

• Attention! There is a "proadcast bug" with older Pexaria gateware (v4.0.0 and below), which potentially may lead to broadcasted garbage events on the timing network. Because of that, please emediatly flash a new gateware on it when the system is Powered On !
• execute "pxe-config <FEC> scuxl.doomsday" or "pxe-config <FEC> scuxl.enigma" to set the required symlinks to the correct Ramdisk)
  • ". /common/usr/cscofe/scripts/accdefs"
• Setup required scripts:
  • On asl cluster: mkdir /common/export/nfsinit/<FEC>
  • Here the required init scripts for a Digitizer FEC (v. alpha8):

    02_digitizer_optimization_supermicro -> ../global/digitizer_optimization_supermicro
    20_timing-rte -> ../global/timing-rte-tg-doomsday-v4.0.5
    50_fesa -> ../global/fesa_64bit_pro_environment
    01_digitizer-alpha8 -> ../global/digitizer-alpha8

• reboot the system to load the correct ramdisk (e.g. via the management port)
• Flash the correct gateware to the Pexaria: https://www-acc.gsi.de/wiki/Main/CheckAndFlashTimingReceiverFirmware
• Optionally open bios and to set "resume to latest powerstate after power loss". (Default is to stay off after there was a power loss)

Installing Digitizer Fesa Software

Attention: FesaEclipsePlugin required !

• Checkout and build (validate, codegen, make) DigitizerClass2 and DigitizerDU2
• When DigitizerDU2.deploy is open, press "add FEC", clone an existing FEC (attention: possibly required to first fix instance file)
• Do required changes in new instance file (rename Fesa Device, *.grc filepath, Trigger Events)
• Copy flowgraph (*.grc) file from existing FEC to new folder and modify flowgraph according to what is requested (E.g. naming of channels, etc.)
• Commit changes to git
• Press "Launch FESA explorer" on the instance to get a *.zip which can be used by FEX
• On the instance file, press "export FESA instance to Database" to be able to access the new device via Japc (Check on this website that it is in: https://websvcpro.acc.gsi.de/groups/cscosv/fesaweb/index.html (only works on asl cluster))
• When DigitizerDU2.deploy is open, press "release", pick "inUse" .. actually thats only required to create all folders relevantr for FESA
• Execute "scripts/releaseDigitizer.sh" to do the actual release (Execute it with -h first) ... that will copy the *.grc file besides other stuff

Historic: USB Trouble

• If e.g. two scopes are recognized, lsusb will show something like:
  • Bus 001 Device 013: ID 0ce9:1202
  • Bus 001 Device 012: ID 0ce9:1211
• USB Bug Picoscope:
  • There was a bug in the picotec firmware which led to failing reboot of supermicro systems (endless reboot loop)
  • Picotec fixed the firmware bug and visited GSI to update the firmware on all our scopes (04.12.2018)
  • Due to our ald kernel, it is still required to unbind + bind aller usb ports via script.
• USB Bug PCIE-USB Card
  • We had problems with an older PCIE-USB Card (picoscope connection lost after some time). But now only new PCIE-USB Cards should be used.