Speaker
Description
At Diamond Light Source, several Macromolecular Crystallography (MX) beamlines focus on, or include, completely automated data collection. This is used primarily for high throughput collection, which has historically meant several hundred samples per day. Diamond is building its next generation, service-based, data acquisition platform Athena using NSLS-II’s Bluesky experiment orchestration library. Using this platform, the MX data acquisition team at Diamond is developing an application named Hyperion to increase the possible rate of automated MX data collection, both for immediate use and in preparation to take advantage of the upgraded Diamond-II synchrotron, due in several years. The automated data collection routines are currently built on legacy experiment orchestration software which includes a lot of redundancy originally implemented for safety when human users are controlling the beamline, but which is inefficient for automated data collection processes.
Using an agile development approach, we have systematically replaced sections of the original data collection routines with calls to the new Hyperion service. This has allowed continuous uptime for delivery of automated data collection to users, without interruption by lengthy testing periods, while immediately seeing the incremental throughput gains as the new sections are implemented.
Hyperion currently runs on beamline I03 and, although it has not yet been thoroughly optimised, is responsible for collecting around 1000 samples per day. Although this is already double the throughput of the legacy implementation there is a requirement to increase this to 5000 samples a day, which Hyperion is expected to achieve by the end of the year. Currently available routines include triggering robot exchanges of samples, sample orientation and centring, and finally collection of rotational data, which represent everything needed for standard data collection. Detailed timing metrics are collected about all of the above to facilitate further speed-up. To come are fluorescence scanning, collection strategy determination, and rotation data collection at multiple positions on a single sample pin – which will enable us to reach the high end of the target throughput – as well as expansion to other MX beamlines.
| Abstract publication | I agree that the abstract will be published on the web site |
|---|
