[3dem] K3 CDS mode issue

[Tomohiro Nishizawa]

━━━━■横浜市立大学からのお知らせ■━━━━
本メールの添付ファイルは、以下のURLからダウンロードをお願いいたします。ダウンロード可能な期間は、最大で90日間ですのでご注意ください。
Please download the attachments from the following URL. Please be aware that the attachments will only be stored for 90 days.
[ダウンロードURL]
https://urldefense.proofpoint.com/v2/url?u=https-3A__f-2Dshare.yokohama-2Dcu.ac.jp_public_B6hEwAPGh8iA-2DdwB-5FtB9nb7VcqqcqE2bNC6gVagYfHHl&d=DwIDaQ&c=-35OiAkTchMrZOngvJPOeA&r=L7-zyQ-04fFCMRqzLIOnx7H0exGZHwIQe_wMPuY600I&m=AxT2VCxHxqyhblTGMKjs9o9xyERUh90iuUtZ9E-79D4-HYsPEvHIxNcc1N8LVJCX&s=rPPCt4MDmX-JelXupylDRHNtCKgk33iIQ3dHt2P9P-U&e= 
公開期限 / Download deadline： 2022/03/19
━━━━━━━━━━━━━━━━━━━━
CDS movies contain heavy noises on the first frame. This problem has not 
been solved even after very recent updates in some Krios in Japan. But 
the attached images imply wrong gain-correction.

Have you checked that the path to the gain reference file is correct? 
After recent update on GMS, the file name of the gain-reference was 
slightly changed. If you are using SerialEM, you have to update the path 
to the gain-reference, or you can simply make a symbolic link to the 
correct gain file.

Did you ever try making gain-file from the data (by using 
relion_estimate_gain)?
If it works, internal setting for the gain reference is somewhat wrong.

Best,

Tomohiro


On 2021/12/18 4:07, Matthijn VOS wrote:
>
> Exactly the same experience at Institut Pasteur. Took more than a 
> year, then GMS upgrade, gatan engineer playing with the settings and 
> gain recorded not at the dose rate requested by the procedure, but a 
> bit higher.
>
> No explanations from gatan as were the origin of this all comes from. 
> To this day not very satisfied and working on replacing it with a TF 
> filter and camera.
>
> Happy holidays!
>
>> On Dec 17, 2021, at 13:06, Wim Hagen <hagen at embl.de> wrote:
>>
>> 
>>
>> Hi Dan,
>>
>> Our K3 has/had this exact problem from the very start when it was 
>> installed. For SPA one can somewhat work around the problem, but it’s 
>> a killer for tomography. Gatan played a bit with the detection 
>> settings for CDS mode (which required a GMS upgrade), that somewhat 
>> fixed the gain ref symptoms, but a subsequent ApoF SPA test of normal 
>> vs CDS mode showed zero difference in performance. We kinda left it 
>> there and only use normal mode on that system.
>>
>> Hearsay:
>> - Some systems have this in normal mode, or CDS mode or both modes.
>> -Some systems were fine after swapping the chip.
>> - One system was fine after swapping the K3 power supply.
>>
>> Best,
>>
>> Wim Hagen
>> EMBL Heidelberg
>>
>>
>> > On 17. Dec 2021, at 02:44, Daniel Toso <dtoso at berkeley.edu> wrote:
>> > 
>> > Hello Everyone,
>> >
>> > We have been having an issue with one of our older K3 cameras since 
>> last year. Namely, in CDS mode we are getting stripes in our images 
>> that do not get removed by gain correction and become worse as the 
>> dose goes further from 7.5 e-/pixel/sec., which will occur quite 
>> often in thicker ice. See attached image for an example.
>> >
>> > We tried having Gatan (Ametek) fix the issue but they do not have a 
>> solution beyond buying a new sensor.
>> >
>> > I was wondering if this is a widespread issue. Also, if there is 
>> any solution besides no longer imaging in CDS mode. We have heard 
>> that it is also seen at UCSF. Any advice would be appreciated.
>> >
>> > Thanks very much!
>> >
>> > Best regards,
>> >
>> > Dan
>> > _______________________________________________
>> > 3dem mailing list
>> > 3dem at ncmir.ucsd.edu
>> > 
>> https://urldefense.com/v3/__https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem__;!!JFdNOqOXpB6UZW0!9vF5z_1W-eVZmSon_oIp055wuboPJvquDo4-2pyT6Z469lEpFAflpvmMk9HLyCLvZ7hNNg$ 
>>
>> _______________________________________________
>> 3dem mailing list
>> 3dem at ncmir.ucsd.edu
>> https://urldefense.com/v3/__https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem__;!!JFdNOqOXpB6UZW0!9vF5z_1W-eVZmSon_oIp055wuboPJvquDo4-2pyT6Z469lEpFAflpvmMk9HLyCLvZ7hNNg$ 
>>
>
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20211220/dde0612b/attachment.html>