Mezurit 2: Virtual instrumentation for electronics experiments

Transcription

1 Mezurit 2: Virtual instrumentation for electronics experiments Dr. Brian Standley FOSDEM 2 Feb 2013

2 Origin of Mezurit 2 Problem #1: Acquire, scale, and record data Problem #2: Sweep a region of parameter space Problem #3: Find and trigger on rare events Common hardware: Semiconductor parameter analyzer (expensive, inflexible) Computer with DAQ and/or GPIB cards Common software: Lab-specific LabVIEW/Matlab/IDL Mezurit "1" by Marc Bockrath and David Cobden measureit by Vera Sazonova

3 Key features Virtual channels: Arbitrary (Python) functions of hardware ports (and GPIB) Invertible functions can be outputs Virtual instruments: Acquisition Scope Sweeps Triggers Command-line terminal Data logging up to ~5 khz Asynchronous acquisition up to hardware limits (~1 MHz) Linear or non-linear output ramps Event detection with predefined responses (also ~5 khz) Continuously-updated plot and data readout

4 Example experiment computer PCI-GPIB Master SR830 SINE OUT Addr :1 divider kω 109Ω Vb junction box Oxford VTI feedthrough PCI-6036e ADC 3 X ADC 4 Y INPUT DL V/A I 130Ω (each) DAC kω V tg DAC 1 10 BGVA 1 MΩ V bg chip carrier

5 Example experiment computer PCI-GPIB Master SR830 SINE OUT Addr :1 divider kω 109Ω Vb junction box Oxford VTI feedthrough PCI-6036e ADC 3 X ADC 4 Y INPUT DL V/A I 130Ω (each) DAC kω V tg DAC 1 10 BGVA 1 MΩ V bg chip carrier

6 A BRIEF DEMO

7 Implementation Code (v0.91): Language files blank comment code C C/C++ Header Python (Linux) libraries: COMEDI Linux-GPIB C Python API Python interpreter GTK+ 2 VTE Platforms: GNU/Linux Windows XP/7 (via MinGW) License: GPL3

8 Architecture Interface loop: Receive input Display status Plot data Terminal: Python interpreter Tools: Widgets Settings MCF nodes RPC nodes Refresh logic Operation logic Widgets Settings MCF nodes RPC nodes Virtual chan. computation Subsystems: Acquisition loop: Binning system GPIB loop Logging Scanning Sweep Trigger MCF system RPC server Wrapper libraries: DAQ GPIB Timing COMEDI NI-DAQ linux-gpib NI GTimer Perf. count. Varset glibc mscvrt Status messaging

9 Architecture GUI GTK events pending? No Yes DAQ RPC received? Yes No Interface loop: Tools: Acquisition loop: Wrapper libraries: Receive input Widgets Settings MCF nodes RPC nodes Binning system Refresh logic Operation No logic Scanning DAQ GPIB Timing COMEDI NI-DAQ Display status GPIB loop linux-gpib NI No Plot Yes data RPC queued? Proc. events Exec. RPC Limit rate (120 Hz) {Xn} readout due? No Yes Ret. buffer status Plot data Copy point Refresh Buffer readout due? Yes Refresh Logging GTimer Perf. count. RPC queue Widgets Acquired point Settings Recorded points MCF nodes RPC nodes Sweep Trigger DAQ Varset glibc mscvrt Type == DAQ? No Yes Terminal: Queue for GUI sadc Acquired point device Sample buffer Exec. RPC Read ADCs Compute {X n} Moving average Binning event? Yes No Record point Exec. cmds. Limit rate (~1 khz) Scanning? No Yes Triggers armed? Yes No T i == 1? Yes No Read samples Transfer due? Yes Store samples Compute {{X n} j} Scanning? Yes No No Scan done? Yes Record points No Python interpreter DAQ device GPIB slots GPIB dev. Virtual chan. computation MCF system Sweeping? No Yes GPIB Slot update due? Yes No Format msg. Send msg. Recv. reply Interp. reply Limit rate (50 Hz) RPC Compute {X'p}, {x'p} server Set GPIB slots Set DACs Status messaging Subsystems: DAQ device

10 Configuration system Text-based config files each line maps to an "MCF node": String identifier Data type Pointer to the setting variable Callback Config is scriptable through the "set_var" terminal function.

11 Configuration system Text-based Terminal: config files rate_widget) each line maps to an "MCF while (running) node": { String identifier 3 Data type 5 Pointer to the setting } variable Callback 7 Config is scriptable through the "set_var" terminal function. lock(mutex) var = new_value unlock(mutex) set_value(widget, Widgets 12 GUI thread: rpc_closure(setvar_rpc, 'set_var') mcf_closure(setvar_mcf, 'sweep_rate', rate_var, } if (rpc_queued()) { cmd, arg = get_msg() func = rpc_lookup(cmd) rpc_call(func, arg) } def setvar_rpc(arg) { key, value = parse(arg) func = mcf_lookup(key)) mcf_call(func, value) } def setvar_mcf(var, widget, new_value) { 9 new_value) Variables 6 delta_var 0.0E+0 rate_var 1.2E+0 hold_var 5.0E MCF Closures 'sweep_delta' setvar_mcf() + var = delta_var + widget = delta_widget 'sweep_rate' 'hold_time' 11 RPC Closures 'get_var' 'set_var' setvar_mcf() + var = rate_var + widget = rate_widget setvar_mcf() + var = hold_var + widget = hold_widget DAQ thread: while (running) { } getvar_rpc() setvar_rpc() 'save_data' savedata_rpc() lock(mutex) step_sweep(delta_var, rate_var, hold_var) unlock(mutex)

12 TODO Ongoing work: GTK+ 3 support Python 3 support Realtime operation (PREEMPT_RT, if possible) Incomplete features here and there Bug fixes Future work: More users >1 developer?

13 Alternative frameworks Current alternatives: LabVIEW Matlab/Simulink Scilab/Scicos/RTAI-Lab A simpler approach?

14 Alternative frameworks Current alternatives: LabVIEW Matlab/Simulink Scilab/Scicos/RTAI-Lab A simpler approach? Model definition (Python): User-supplied script Instrument libraries Interface library Control and display GUI (Python): User-supplied script Widget libraries Interface library Realtime layer (C): General DAQ / control engine Drivers

15 Credits Helpful discussions: Prof. Marc Bockrath (UC Riverside) Prof. Henk Postma (CSU Northridge) Testing: Dr. Hang Zhang, Dr. Wenzhong Bao, Dr. Jairo Velasco Jr., Peng Wang, Tengfei Miao, Oleg Martynov THANKS!