Another example of Remote Server: Cameca electron image navigation

[JonF]

Recently, we've been having problems with the control keyboard on our SX100 resulting in the keyboard being disconnected.
This made driving around samples more difficult than usual, so I decided to have a look at what the Remote server could do.

Using the video capture hardware I bought for monitoring the electron beam image remotely, I coded up a small program to capture the video stream and interface with the SX100 via the Remote Server.

The result: I can now navigate on the Cameca in a similar way to our JEOL EPMA and Quanta SEM: double click on the (live!) image to move the stage to the desired location and scroll the mouse wheel to zoom in or out of the image. 
I've also checked this out on a touchscreen (MS Surface Pro) and double tapping works to move the stage and pinch to zoom in/out also works (!).



The image capture isn't perfect (i.e the vertical stripes, the slightly offset image), but it works pretty well otherwise and makes moving around the sample much, much easier.

[John Donovan]

Wow, that is pretty amazing.  Very cool.

It seems to me that since you're using Remote Server for the instrument communication, this app could be instrument agnostic?  I mean it wouldn't take much work so that it could run on either JEOL or Cameca instruments, correct?

[JonF]

It seems to me that since you're using Remote Server for the instrument communication, this app could be instrument agnostic?  I mean it wouldn't take much work so that it could run on either JEOL or Cameca instruments, correct?

It doesn't interact directly with the EPMA so should be instrument agnostic, yeah.
Not sure why you'd need it on a JEOL though? Do older JEOLs not have the live imaging as part of PC-SEM?
You'd need to find a way of generating a video stream for the electron beam image. With the Cameca, I have a VGA splitter between the instrument and the electron beam monitor, and I feed that in to the video converter attached to the PfE PC.

[John Donovan]

Are we talking about a different Remote Server? I thought you mean this app was based on the Remote Server Active-X object from Probe Software?  The Remote Server interface (which is what this board is about) supports calls made to the instrument for stage, spectrometers, column and imaging control.

[JonF]

Are we talking about a different Remote Server? I thought you mean this app was based on the Remote Server Active-X object from Probe Software?  The Remote Server interface (which is what this board is about) supports calls made to the instrument for stage, spectrometers, column and imaging control.

Yeah, I'm using the Remote Server that you made. When i say "it doesn't interact with the EPMA", I mean it makes the call to the instrument via Probe for EPMA (or Remote Server) rather than direct to the instrument.

For example, the program is calling RemoteGetMagnification() and RemoteGetMotorPosition() to query the instrument, then I've pinched the image-stage calibration from the ProbeWin.ini file to calibrate the pixels from the electron beam image to stage coordinates as a function of magnification, calculate the distance from the centre of the image to where a mouse double click event happens (within the video window) and then use RemoteMoveMotor() to move the stage the appropriate amount.

[John Donovan]

Ah, OK.   So you're using Remote Server for the image mag and stage positions, and the video conversion hardware for the image acquisition.  Note that one can also capture a video image using the Remote Server, but the video capture will be faster I'm sure.

Well I think it's a great idea and I think it could work on both instruments assuming one bought an additional video conversion board.

[JonF]

I've just been doing some googling, for JEOL I guess you're thinking for older probes running on UNIX? They look to have a separate electron beam image monitor like on the SX100 so yes, this should work for those instruments as well.

On my SX100, the refresh rate is set in Peak Sight and the image is displayed real time (i.e. the same as the electron image monitor), so there's no need to acquire an image (e.g as in the imaging window of PfE). Just drive around using the SE or BSE image as you would on a modern SEM.

Do you still create a network stream of your electron beam image?

I was toying with the idea of interfacing with the position.dat file and have this software plot the positions up directly, but I figure it's simple enough to just have the digitize window open next to this.

[John Donovan]

I've just been doing some googling, for JEOL I guess you're thinking for older probes running on UNIX? They look to have a separate electron beam image monitor like on the SX100 so yes, this should work for those instruments as well.

OK, I couldn't remember if the 8530 had a dedicated electron imaging monitor or not.

On my SX100, the refresh rate is set in Peak Sight and the image is displayed real time (i.e. the same as the electron image monitor), so there's no need to acquire an image (e.g as in the imaging window of PfE). Just drive around using the SE or BSE image as you would on a modern SEM.

Yes, that should work great.

Do you still create a network stream of your electron beam image?

A network stream? I don't think we ever did that.  Trying to think how that might be accomplished...

I was toying with the idea of interfacing with the position.dat file and have this software plot the positions up directly, but I figure it's simple enough to just have the digitize window open next to this.

It's an MDB file, so you'd have to read the Access database, so a bit of a pain.  On the other hand, once you're navigated to your area of interest and are ready to digitize points, you can capture an FOV in PFE, and then digitize points either in the PictureSnap window using a right mouse click, or from the Digitize! Digitize Image window using the left mouse click.

[JonF]

Do you still create a network stream of your electron beam image?

A network stream? I don't think we ever did that.  Trying to think how that might be accomplished...

I was thinking of the setup you mentioned here:
https://probesoftware.com/smf/index.php?topic=581.0

The only exception would be screens that aren't part of the windows display system, e.g., the dedicated electron imaging monitor and the optical image monitor. For these monitors we bought video convertors for each monitor and piped those signals to a four channel video server that the PTZ camera is also on. So in quad view seen here:

Regarding plotting up positions, I was using the ebeam navigator yesterday and did as you mentioned, it seems to work pretty well and certainly speeds things along. One thing the navigator does highlight is that I need to work on my beam-stage calibration!

[John Donovan]

Ah, OK. Yeah that was using a PAL to NSTC converter. It's too bad that the new JEOL video imaging system is embedded in the JEOL PC.  The new Cameca tactis system uses a USB port connection for video imaging now. We're actually working on getting that working for a customer (LANL).

On the beam-stage calibration the two point method works pretty good:

https://probesoftware.com/smf/index.php?topic=329.msg10947#msg10947

Funny enough we've recently asked Karsten Goemann to document this procedure a bit, but it's pretty simple now that the image calibrations at each mag can be updated using the Save Image Calibration to Probewin.INI button.

If ones instrument has some scan rotation issues, you can call in your instrument engineer, or one can utilize the three point image calibration method described here:

https://probesoftware.com/smf/index.php?topic=396.0

This is what we had to do on our instrument because the scan rotation is always about 9 degrees off from vertical and varies slightly with changes in magnification.