When starting a data taking run with TOS, a new directory for the run is created in the data/runs subdirectory. The name will be of the form Run_<run number>_YYMMdd-HH-mm where the run number is an increasing number based on the last run present in the directory and the suffix is the date and time of day when starting the run. This directory contains the configuration of all DACs for each chip, fsr<chip number>.txt, the written configuration matrix for all pixels, matrix<chip number>.txt and finally the data files data<event number>.txt. If an FADC is used for the readout additional data<event number>.txt-fadc files are created, one for each file in which the FADC triggered (sec. 17.2.1.1).

The Timepix data files are stored – for historic reasons – in raw ASCII format. Two different readout modes (with different output formats) are supported. For the following explanation it is assumed the Timepix is used in the ToT (Time-over-Threshold) mode.

full matrix readout

reads out the whole Timepix ASIC(s) and writes a single 256x256 pixel matrix as an ASCII file (for each chip). 256 lines, each containing space separated ToT values for each pixel.

zero suppressed readout

reads out only those pixels that have ToT values larger than 0, up to \num{4096} pixels. Stores the data in TSV files (tab separated values) 'X Y ToT' with an additional header. The header contains a global "run" and "event" header, which contains information about the run the event is taken from and a "chip" header, which contains information about the specific Timepix ASIC(s) being read out (up to 8 can be read out at the same time using TOS).

As for our purposes most events are extremely sparse (< 500 pixels active) the zero suppressed readout is the only relevant readout mode. The data files can be split into 3 distinct parts. A global run header, see listing 2, which contains information about the run the event is part of including important settings used as well as the timestamp of the event. Next is an event specific header, which contains specific information about the event in relation to the FADC and the scintillators, see listing 3. The final part of the zero suppressed data files is the chip header and tab separated value part of the 'X Y ToT' pairs of the active pixels for each chip of the detector in that event, see listing 4.

## [General]
## runNumber:        339
## runTime:          7200
## runTimeFrames:    0
## pathName:         data/runs/Run_339_190218-10-36
## dateTime:         2019-02-18.10:36:34
## numChips:         7
## shutterTime:      2
## shutterMode:      verylong
## runMode:          0
## fastClock:        0
## externalTrigger:  0

## [Event]
## eventNumber:      2
## useHvFadc:        1
## fadcReadout:      1
## szint1ClockInt:   0
## szint2ClockInt:   0
## fadcTriggerClock: 647246

# chipNumber: 0
# chipName:   E 6 W69
# numHits:    0
# chipNumber: 1
# chipName:   K 6 W69
# numHits:    0
# chipNumber: 2
# chipName:   H 9 W69
# numHits:    2
106     160     75
211     142     2

Everything related to the HFManager in TOS is controlled by a configuration file, normally located in TOS/config/HFM_settings.ini. The TOS configuration file used during CAST data taking is found in appendix 18.1. We will go through the sections of it one by one and explain them.

Starting with the [General] section, listing 8. This section defines the VME related settings. The VME address of the HV module installed in the crate is used as the base address. The FADC address in the same VME crate is calculated from an offset in units of the VME address spacing of 0x0400. ⁶

[General]
sAddress_fadc = 1
baseAddress_hv = 0x4000

The next section [HvModule], listing 9, are general settings about the used HV module. The settings are related to the KillEnable feature of Wiener HV power supplies, the ramping speed of the HV channels and the time in seconds between sanity checks of the HV during data taking. ⁷

[HvModule]
setKillEnable                   = true
# Voltage and Current RampSped currently set to arbitrary value
# in percent / second
moduleVoltageRampSpeed          = 0.1
moduleCurrentRampSpeed          = 50
# checkModuleTimeInterval       = 60, checks the status of the
# module every 60 seconds during a Run, between two events
checkModuleTimeInterval         = 60

Next up, the [HvGroups] section in listing 10 defines the different groups that combine multiple channels. There are multiple different kinds of groups in Wiener HV power supplies. The important groups are ramping groups and trip groups. Essentially, if one channel in a group starts ramping / trips all others also start ramping / shut off the HV, respectively. The section in the config file mainly exposes the already predefined sets of groups that are relevant for the Septemboard detector in TOS.

# if this flag is set to true, anode and grid
# will be coupled to one group
[HvGroups]
anodeGridGroupFlag              = true
# grid is master channel of set on group
anodeGridGroupMasterChannel     = 4
anodeGridGroupNumber            = 0
monitorTripGroupFlag            = true
monitorTripGroupNumber          = 1
rampingGroupFlag                = true
rampingGroupNumber              = 2                          
gridChannelNumber               = 4
anodeChannelNumber              = 5
cathodeChannelNumber            = 8

After the [HvGroups] section is the definition of the individual HV channels in listing 11. Here the physical channels on the device are mapped to the desired voltages and current bounds as well as to a human readable name. The fields repeat with increasing prefix numbers.

[HvChannels]
# all currents given in A (vmecontrol shows mA)
0_Name                          = grid
0_Number                        = 5
0_VoltageSet                    = 300
0_VoltageNominal                = 500
0_VoltageBound                  = 2.5
0_CurrentSet                    = 0.000050
0_CurrentNominal                = 0.000500 
0_CurrentBound                  = 0

Second to last is the [FADC] section in listing 12. As the name implies it configures all parameters of the FADC. The main parameter to change is the fadcTriggerThresholdRegisterAll parameter, which defines the trigger threshold in effectively \(\si{mV}\). Depending on the amount of noise in the system, adjustments to the threshold may be necessary.

[Fadc] # FADC Settings
fadcTriggerType                 = 3 
fadcFrequency                   = 2
fadcPosttrig                    = 80
fadcPretrig                     = 15000
# was 2033 before, 1966 corresponds to -40 mV
fadcTriggerThresholdRegisterAll = 1966 
# run time of a single pedestal run for the FADC in ms
fadcPedestalRunTime             = 100
# number of acquisition runs done for each pedestal calibration
fadcPedestalNumRuns             = 10
# using channel 0 on FADC as trigger source, thus bit 0 = 1!
fadcChannelSource               = 1
# set FADC mode register (mainly to enable 14-bit readout)
fadcModeRegister                = 0b000

The last section of the configuration file is the [Temperature] section, which deals with the safety ranges of the temperature of the detector. If the temperature leaves the safe range, the detector is to be shut down. ⁸

[Temperature] # temperature related parameters, all temps in °C
safeUpperTempIMB                = 61
safeUpperTempSeptem             = 61
safeLowerTempIMB                = 0
safeLowerTempSeptem             = 0

As is being alluded to in the previous section 17.2.2, the HV control built into TOS can also handle ramping the channels of the detector. This is particularly convenient as it offers a very smooth ramping mode, which keeps the voltage potentials between all channels under a constant ratio. This allows for automatic ramping even for highly sensitive channels (like the GridPix grid).

In order to use the HV control and ramp the channels via TOS, ActivateHFM must be followed by InitHV, which attempts to connect to the HV power supply using the configuration of the config file. If the channels are not ramped up, a call to RampChannels will start the smooth ramping process (see listing 14).

> ActivateHFM
> InitHV
> RampChannels

If the HV is to be ramped down, a call to ShutdownHFM will ask whether the channels should be ramped down. There are a multitude of further commands available to communicate with the module, check the voltages, print status information etc.

Note that TOS can be started without ramping the HV channels and stopped without ramping the channels down. It is capable of connecting to a running HV power supply or leave it running after shut down.

The two MAX31685 micro controllers read out PT1000 sensors, a group of \textbf{R}esistance \textbf{T}emperature \textbf{S}ensors (RTDs), which measure temperatures by its effect on the electrical resistance. They are platinum based and have a resistance of \(\SI{1000}{Ω}\) at \(\SI{0}{\celsius}\). As the expected change in resistance is well understood, the temperature can be precisely measured.

The micro controller communicates with another micro controller, an MCP2210 via the \textbf{S}erial \textbf{P}eripheral \textbf{I}nterface (SPI). SPI allows to address both MAX31685 from the single MCP2210. The MCP2210 is a USB-to-SPI micro controller. The USB connection from the intermediate board with the computer is separate from the rest of the detector communication. TOS communicates with it via the standard \textbf{H}uman \textbf{I}nteface \textbf{D}evice (HID) driver and utilizes an existing open source library for the MCP2210 (Wong 2013), which is slightly adapted.

The ActivateHFM command mentioned in the previous section also attempts to find the USB device of the MCP2210 (the two are intertwined mainly, as the Septemboard detector is the only detector with either of the two features). If it is found, temperature logging starts immediately and the log files are placed in the default log directory of the TOS repository. Once a data taking run starts, the logging location is moved over to the data storage directory of the run. In either case the log file is named temp_log.txt and contains one temperature value for the intermediate board sensor (Temp_IMB) and one for the carrier board sensor (Temp_Septem) computed – based on an average over \(\SI{5}{s}\) – and a timestamp (DateTime). A short snippet of the temperature log is shown in listing 15. ⁹

# Temperature log file
# Temp_IMB  Temp_Septem   DateTime
26.5186     42.1472       2019-02-16.16:11:45
26.5217     42.2798       2019-02-16.16:11:51
26.5202     42.4371       2019-02-16.16:11:57
26.5309     42.5944       2019-02-16.16:12:03
26.5324     42.7347       2019-02-16.16:12:09
26.5355     42.8627       2019-02-16.16:12:15
26.5416     42.9707       2019-02-16.16:12:21
26.5432     43.0771       2019-02-16.16:12:27
26.5616     43.1804       2019-02-16.16:12:33
...

As mentioned in one of the footnotes in the previous section, there are nowadays 2 independent versions of TOS.

The detector used for CAST in 2014/15 (and thus its successor used in this thesis) was based on a readout using the Virtex 6 FPGA. This system was, at the point I started on my master thesis in 2015, already quite diverged from the SRS based system, which was mainly developed for multi chip detectors that were initially planned for a large GridPix based TPC to be used for the ILD (the detector planned for the ILC, the International Linear Collider to be built in Japan).

In addition there was recent a master thesis (by Alexander Deisting (Deisting 2014)), which included work on using an FADC to read out the induced charges on the grid of the InGrid by decoupling the signal using a capacitor. The software library to interact with the used FADC had partially been implemented into the Virtex 6 TOS.

As the FADC was an integral part in the new detector design, it was natural to start with the Virtex 6 version.

At the same time the SRS TOS version at the time was even more ugly than the same code paths in the Virtex 6 TOS, due to its hardcoded extra loops for each FEC of the TOS. At a later time the SRS TOS was required for other detectors and so development effort was spent on both systems unfortunately.