Feature #2181
closedK2 Summit linear mode integration
0%
Description
I was able to get 500 ms exposure at ~64 electron/pixel/s (=40 e/A^2 at 1.27 A/pixel) with linear mode on K2 Summit. There are a few things need to take into account for Leginon integration.
In linear mode, Leginon will not use DM's dark/gain references (neither hardware nor software).
- Leginon needs to scale dark reference by exposure time to make proper correction when exposure time is changed.
- There appears to be overflowing or many noisy pixels in the image. These need to be handled.
- The returned array in pyscope.gatanK2.__getImage is a square of (3712,3712) even if I specify as in DM for (3712,3840) before rotation when frame saving is off.
- The returned array is a rectangular (3840,3588) if frame saving is on.
See session 12dec12f for examples. 12dec11f_00009ma_0.mrc is the best example of the noisy pixel or over/underflow effect.
Updated by Anchi Cheng over 11 years ago
Jim changed the type in r17313 and it solved problem 2, and problem 3 and 4 disappears when we tested it together, but the bland corrected image looked a lot worse than the last run. See 12dec14jimtest1_00002ma_0.mrc
Updated by Anchi Cheng over 11 years ago
- Status changed from New to Assigned
- Assignee set to Anchi Cheng
- Priority changed from Normal to High
r17348 uses frame rate and "use frames" to calculate exposure time of the integration so that it works on K2 linear mode both in integration and with dose fractionation on.
Need to test on DE-12 and Tietz F416.
Updated by Anchi Cheng over 11 years ago
- Status changed from Assigned to Closed
- Priority changed from High to Normal
A few people used linear mode for final exposure using ~50 e/pixel/s intensity and 0.5 secoand exposure. It is now reasonably good. There are some lines appears in the image, not correctable with gain/dark correction, but is sufficient if people just want a quick image of their sample.
As a user, caution should be taken if you want to use DM gui directly interchanging with Leginon. DM GUI does not behaves consistently and can be stuck thinking that it is doing linear mode but using hardware dark subtraction and gain correction as well as other non-functional state when the mode is assigned through scripting as in Leginon. Check carefully before starting acquiring in DM after Leginon usage.
DE-12 and Tietz F416 runs fine with the new code.
The remaining problem is transferred to Issue #2250