mbclient

Client software and on the fly visualisation tools for the moessbauer effect.

The client software

The client software is a command line tool capable of connecting to the Red-Pitaya running the MBFilter program in server mode, providing the server on the Red-Pitaya with all the neccesary information to configure itself properly and to store the data received from the Red-Pitaya in a .csv file for Analysis by the student and also Visualise the data as a pulse-height spektrum and a 2d spektrum of pulse height vs. Digital Function Generator Address. The Visualisation is updated continuously and can be exited without interfering with the data taking procedures.

Configuration options

The client software can be configured in two ways, the main method of configuration is a yaml config file. An example is provided in this repository and printed here for convenience:

basic:
  rise-time: 10
  hold-time: 20
  pulse-decay-time: 390
  pulse-height-threshhold: 1000000

environment:
  server-ip: '192.168.0.2'
  server-port: '8080'

There is an advanced option available to override the internally calculated accumulation-time. To set this add the following section to the config file:

advanced:
  accumulation-time: 22

The configuration options rise-time, hold-time and pulse-decay-time are the parameters of the trapezoidal filter. The pulse-height-threshhold is used to suppress small (and thus low energy) peaks and primarily acts to reduce noise, as the high frequency of the noise can overwhelm the filter. The IP of server should not change through the course of a semester and the port only changes if explicitly set at server startup. The accumulation-time is used to avoid counting one peak multiple times. It should only be set if there are major problems with the setup, as it is automatically calculated from the rise-time and hold-time of the trapezoidal filter.

The second way is via command line arguments, that are similarly named to the configuration options. The command line values, when given, override the configuration options specified in the config file.

Configuring plotting

The command line also has a flag to disable plotting. Disabling plotting should be done when attempting the long duration measurements as plotting the values slows down the process significantly after a while. The plot can be closed at any time during the program execution without interrupting the data taking activities. It however cannot be restarted after being closed.

The --histmin and --histmax options set the lower and upper bound on the pulse-height of the events being plotted. The bin width is adjusted automatically. The plot also automatically rescales so that all the data is visible.

Behaviour of the Program

The server will stop sending events, if the FPGA-Internal hardware buffer overflows. This is an indicator, that the system as a whole is overwhelmed with the number of events. As the buffer is only so large, the overflowing is poisson distributed. The larger the average event rate, the more likely it is for the buffer to overflow due to many events in short succession. If this happens within the first 5 minutes of the experiment, try increasing the pulse-height-threshhold about 50000, this cuts more of the signal that is suspected to be noise and reduces the likelihood of the filter overflowing and the experiment being halted.

To restart the experiment, simply call mb-client again. The Red-pitaya will automatically reconfigure and start the next measurement.

In the (very rare) case where this does not happen, go to the ssh connection to the Red-pitaya and quit the server with <Ctrl>C then call the ./start_server.sh again.

The program structure

The mbclient package consists of the cli application located at cli.py and associated functions in the mbclient.py file. The command line program operates asynchronously with asyncio. There are a total of three tasks.

1. The user facing task waits for a stop input from the user and sends a terminate signal to the other running tasks.
2. The process_data task opens a websocket connection to the Red-Pitaya experiment, reads in the data from there and passes it along to 'consumer' tasks.
3. The write_to_file task writes the decoded data it receives from the task 2 and writes it into an csv file.
4. The plot_data task starts a second process, that is responsible for 'live' plotting of the data and forwards the data to it via a PIPE.

The mbdatatypes.py file contains the Class that represents the result coming from the Red-Pitaya. It contains methods to decode the raw data from the websocket and methods to transform the data read to a csv entry.

The mbplotter.py file contains the Class that is spawned off into the second process to plot the data, as well as a class that acts as an API for sending data to the plotting process via the plot(data) funciton. The characteristics and details of the plot are encoded in the ProcessPlotter class.

Notes on Matplotlib

Matplotlib does quite a lot of things. One of them being the implementation of it's own event loop similar to the one used from asyncio. With the event loop comes the ability to set a timer that fires a callback function that updates the plot. All this is implemented in the ProcessPlotter class. The Pipe between the processes is used as a data buffer.