RE: quadEM FFP plugin update issues

["Ha, Kiman via Tech-talk" <tech-talk@aps.anl.gov>]

Hi Mark,

 

Thank you for the quick update.

Before proceeding with your suggestion, I checked the other waveforms and confirmed that only the FFTAbsValue has stopped updating.

The rest are functioning normally.

 

I’ll follow up with your recommendation and investigate it and will update.

 

Thanks,

Kiman

 

 

From: Mark Rivers <rivers@cars.uchicago.edu>
Sent: Saturday, September 27, 2025 12:22 PM
To: Ha, Kiman <kha@bnl.gov>; tech-talk@aps.anl.gov
Subject: RE: quadEM FFP plugin update issues

 

Hi Kiman, I just tested on my system, and I can’t reproduce the problem. This is not related to the issue of the plugin updates stopping, but I would suggest that you change the number of points from 1000 to 1024. I think the “beats”

ZjQcmQRYFpfptBannerStart

This Message Is From an External Sender

This message came from outside your organization.

ZjQcmQRYFpfptBannerEnd

Hi Kiman,

 

I just tested on my system, and I can’t reproduce the problem.

 

This is not related to the issue of the plugin updates stopping, but I would suggest that you change the number of points from 1000 to 1024.  I think the “beats” you see in frequency are artifacts of the fact that the number of points is not a power of 2.

 

This is what my Y position FFT looks like:

 

 

The first thing to do is open the detailed screens for that FFT plugin.  This is what I see with “More/FFT plugin”.  Note that Array rate=10 Hz, the execution time is 0.474 ms, and the array counter is incrementing at 10 Hz.  Are these values updating on your screen?

 

 

This is what I see when I press the More/NDPluginBase Full”  on the screen above:

 

 

The next thing to determine is whether Channel Access updates of the waveform record are being sent.  The way to test this is with camonitor.  This is what I see with camonitor on the X and Y positions, asking only for the first 2 points in the waveform.

 

(base) [epics@corvette ~]$ camonitor -#2 13IDA:QE1:FFT:PosX:FFTAbsValue

13IDA:QE1:FFT:PosX:FFTAbsValue 2025-09-27 10:56:29.331851 2 0 1.15078

13IDA:QE1:FFT:PosX:FFTAbsValue 2025-09-27 10:56:29.431627 2 0 1.15078

13IDA:QE1:FFT:PosX:FFTAbsValue 2025-09-27 10:56:29.531824 2 0 1.15078

13IDA:QE1:FFT:PosX:FFTAbsValue 2025-09-27 10:56:29.631831 2 0 0.822881

13IDA:QE1:FFT:PosX:FFTAbsValue 2025-09-27 10:56:29.731868 2 0 0.822881

^C

(base) [epics@corvette ~]$ camonitor -#2 13IDA:QE1:FFT:PosY:FFTAbsValue

13IDA:QE1:FFT:PosY:FFTAbsValue 2025-09-27 10:56:34.831667 2 0 0.326255

13IDA:QE1:FFT:PosY:FFTAbsValue 2025-09-27 10:56:34.931805 2 0 0.326255

13IDA:QE1:FFT:PosY:FFTAbsValue 2025-09-27 10:56:35.031571 2 0 0.0809308

13IDA:QE1:FFT:PosY:FFTAbsValue 2025-09-27 10:56:35.131777 2 0 0.0809308

13IDA:QE1:FFT:PosY:FFTAbsValue 2025-09-27 10:56:35.231808 2 0 0.0809308

 

They are both updating at 10 Hz, as expected.

 

What do you see when you run this camonitor command on your X and X FFT positions?

 

The next thing to do is use “ps” to list the threads with names that contain FFT.  This is what I see:

 

(base) [epics@corvette ~]$ ps -eL | grep FFT

1737089 1737178 pts/14   00:00:00 QE1_FFT1_update

1737089 1737181 pts/14   00:00:00 QE1_FFT2_update

1737089 1737184 pts/14   00:00:00 QE1_FFT3_update

1737089 1737187 pts/14   00:00:00 QE1_FFT4_update

1737089 1737190 pts/14   00:00:00 QE1_FFT5_update

1737089 1737193 pts/14   00:00:00 QE1_FFT6_update

1737089 1737196 pts/14   00:00:00 QE1_FFT7_update

1737089 1737199 pts/14   00:00:00 QE1_FFT8_update

1737089 1737202 pts/14   00:00:00 QE1_FFT9_update

1737089 1737205 pts/14   00:00:00 QE1_FFT10_updat

1737089 1737208 pts/14   00:00:00 QE1_FFT11_updat

1737089 1737301 pts/14   00:00:03 QE1_FFT1_Plugin

1737089 1737303 pts/14   00:00:04 QE1_FFT2_Plugin

1737089 1737305 pts/14   00:00:04 QE1_FFT3_Plugin

1737089 1737307 pts/14   00:00:04 QE1_FFT4_Plugin

1737089 1737309 pts/14   00:00:04 QE1_FFT5_Plugin

1737089 1737311 pts/14   00:00:04 QE1_FFT6_Plugin

1737089 1737313 pts/14   00:00:04 QE1_FFT7_Plugin

1737089 1737315 pts/14   00:00:04 QE1_FFT8_Plugin

1737089 1737317 pts/14   00:00:04 QE1_FFT9_Plugin

1737089 1737319 pts/14   00:00:04 QE1_FFT10_Plugi

1737089 1737321 pts/14   00:00:04 QE1_FFT11_Plugi

 

Note that there are 11 FFT threads with _update in the name and 11 with _Plugin in the name.  What do you see?  Are all the threads there?

 

Let me know the results of these tests on your system.

 

Mark

 

 

From: Ha, Kiman <kha@bnl.gov>
Sent: Friday, September 26, 2025 5:59 PM
To: tech-talk@aps.anl.gov
Cc: Mark Rivers <rivers@cars.uchicago.edu>
Subject: quadEM FFP plugin update issues

 

Hi Mark,

 

We are using a new NSLS2EM hardware setup with the quadEM IOC.

 

One of our IOCs has an issue where the FFT plugin stops updating shortly after the IOC is started.

Although the FFT plugin has stopped updating, the circular buffer index continues to update.

 

Time-domain display continues to work without any problems.

Do you have any suggestions or recommendations to help resolve this issue?

Is it possible to restart the FFT plugin without rebooting the IOC?

 

See plot the PosX FFT is still running but the PosY is stopped.

 

 

 

Thanks,

Kiman