Timescales for Imaging with PFE

[Dan R]

Hello,
I am wondering if anyone out there with a JEOL 8530 or similar instrument (new EIKS interface) has noticed whether it takes quite a while to collect an image using PFE... (512x512 pixel image is taking me ~2-3 minutes to collect, but I have not timed it exactly). Compared to our JEOL 8200, this is considerably slower. Does this have to do with the new EIKS interface? Could it be related to my network card? Any thoughts would be helpful, thanks!
-Dan

[John Donovan]

Hi Dan,
I can answer this question for you because I know exactly what the problem is.

The problem is that JEOL has provided us with the mapping call for x-ray mapping, but they have *not* provided the video call for fast analog signal imaging.  Therefore, because the minimum pixel dwell time for the mapping call is 100 us or 0.1 msec, a 1024 x 1024 map will take 100 seconds just for the mapping, not counting any overhead.

Here is the list of JEOL instrument calls for the 8230/8530 that we are still waiting for:

1. Get/set image shift.
2. Read gun vacuum (existing EIKS function works) and sample vacuum (existing EIKS function does not work).
3. Get video image (fast X/Y image for analog signals, BSE, SE, etc)
4. Open/close the gun valve (e.g, JEOL FEG V1), to protect gun vacuum
5. Acquire SE and BSE signals simultaneously (JEOL "comparison mode")
6. Command to set picoammeter auto-range on/off?  May not exist, but why does the new JEOL picoammeter take 1 to 2 seconds to "settle"?
7. Command to get/set JEOL 8530 EPMA/SEM condenser lens mode.
8. Command for the PHA multi-channel analyzer (MCA) call, to improve the PHA scan speed.

Please feel free to bug Charlie about this. It helps when his customers call him!

[Dan R]

Thanks for the clarification John! I was just trying to determine whether or not it was my network card or a known problem with the EIKS interface.

Yes, it's a known problem with JEOL!   
john

[Dan R]

John-
I timed how long it took for my system to acquire an image and there maybe another underlying problem. A 128X128 pixel image (16384 pixels at 100 microsec dwell) should take 1.64 seconds (unless my calculations are wrong)... my image acquisition took ~30 seconds. I can't image that there would be that much overhead. Can you think of another way to check why this is taking so long?
Thanks,
Dan

[John Donovan]

Hi Dan,
As I said, until we get the fast video call from JEOL it's going to be like that. Also you are comparing apples to oranges. Did you try a 128 x 128 mapping call for the BSE image in the JEOL software?

I'm sure they have optimized the mapping call more than us, but I'll bet the times are closer.  If I remember right from the PNNL testing I did a few years ago, the JEOL mapping call takes 20 secs just to start once one clicks the "start" button.
john

[Dan R]

John-
I must be comparing apples to oranges. I have not yet tried to 'Map' a BSE image, but that makes sense. Thanks for the clarification.
-Dan

[Probeman]

The problem with the JEOL 8230/8530 is that JEOL has removed the video scan generator from the instrument and placed it in an external box which is only accessed by the JEOL PC.  The mapping scan generator is still in the instrument which is why PFE can still access it using the mapping commands.

Therefore the BSE/SE etc imaging speed that PFE can utilize is limited by the JEOL mapping scan generator which is 100 usec or 0.1 msec pixel. So not counting any overhead that means a 1024 x 1024 BSE image will take 104 seconds or almost 2 minutes using the mapping scan generator.

We've asked JEOL to provide an API for the video scan generator so we can also acquire fast video images in PFE, but they have ignored us so far. You know that it will take someone buying a new instrument from JEOL, who insists on this addition to their EIKS API, to get them to respond.

The only other solution is to utilize the video imaging scan generator in a Thermo or Bruker EDS system since PFE can interface to both for imaging. This is done simply by specifying that the video acquisition interface defined in the Probewin.ini file [image] section is set to:

[Image]
ImageInterfacePresent=1
ImageInterfaceType=9   ; 9=Bruker

or

[Image]
ImageInterfacePresent=1
ImageInterfaceType=10   ; 10=Thermo

This solution works fine if you have a newer Thermo NSS or Bruker Esprit SDD EDS system on your probe. The only other solution is to buy a Cameca instrument, which offers an integrated fast video acquisition for PFE in both the SX100 and SXFive instruments... or as I said, someone in the community buying a new JEOL instrument who adds a fast video imaging acquisition API specification in their instrument purchase specs.

[Dan R]

Hi-
I found one other potential option... apparently Thermo can interface with a JEOL EDS system and route all signals through their computer and software. This way we would only have to buy the computer/scan generator package without having to get another EDS detector.
-Dan