Author Topic: Pathfinder crashing during combined WDS-EDS acquisition with Probe For EPMA  (Read 1686 times)

Karsten Goemann

  • Global Moderator
  • Professor
  • *****
  • Posts: 179
All,

I'm doing a lot of combined WDS-EDS point measurements in Probe For EPMA now (EDS for majors, WDS for traces...). We're getting very good and consistent major element data for the combined WD-ED acquisitions, but every now and then Pathfinder crashes during an overnight run (yesterday when acquiring point no. 90 at 1am in the morning). When I come in the next morning the Pathfinder window has closed and PfE sits there waiting with the beam still on the last position where it crashed. At least this was olivine and not a more beam sensitive material...

This is using PfE 11.9.5 and Pathfinder 1.2, Workstation Build 72, Analyzer Build 190, both running under Windows 10 on the same PC.
In addition to the ED spectra I also acquire automated BSE images for each point through Thermo. But I think this happens during spectrum acquisition, as the beam conditions I encounter the morning after are the ones for the x-ray intensities measurement, not the the BSE imaging. This is a JEOL 8530F Plus instrument.

Any suggestions how to troubleshoot this? So far it only has happened twice during long runs over night.

Many thanks,
Karsten

Probeman

  • Emeritus
  • *****
  • Posts: 1883
  • Never sleeps...
    • John Donovan
I'll let someone at Thermo respond, as I assume they'll want you to capture a log of the issue...  but this sounds more serious than a socket timeout, as I assume there was no error return from the Pathfinder software in PFE?  If PFE receives an error from Pathfinder it automatically inserts the Faraday cup and displays the error. By the way, I ran 500+ points with spectrum acquisition using Thermo Pathfinder last week:

http://probesoftware.com/smf/index.php?topic=888.msg6123#msg6123

with no issues, and I'm running spectrum acquisitions tonight, so we'll see what the morning brings, but I have to say I haven't seen this issue on my instrument, with the same Pathfinder version.

john
« Last Edit: August 01, 2017, 10:07:31 pm by Probeman »
The only stupid question is the one not asked!

Probeman

  • Emeritus
  • *****
  • Posts: 1883
  • Never sleeps...
    • John Donovan
I'll let someone at Thermo respond, as I assume they'll want you to capture a log of the issue...  but this sounds more serious than a socket timeout, as I assume there was no error return from the Pathfinder software in PFE?  If PFE receives an error from Pathfinder it automatically inserts the Faraday cup and displays the error. By the way, I ran 500+ points with spectrum acquisition using Thermo Pathfinder last week:

http://probesoftware.com/smf/index.php?topic=888.msg6123#msg6123

with no issues, and I'm running spectrum acquisitions tonight, so we'll see what the morning brings, but I have to say I haven't seen this issue on my instrument, with the same Pathfinder version.

john

Hi Karsten,
Ran two overnight runs with EDS spectrum acquisition using Thermo Pathfinder 1.2 and no problems.  I was even uploading some big files during the run.  I hope you figure out your issue soon!
john
The only stupid question is the one not asked!

emma_fisi

  • Post Doc
  • ***
  • Posts: 23
Hi Karsten, John,

Are either of you still noticing issues with running many points using Pathfinder and PfE? I've tried to do a couple of overnight runs this weekend, and Pathfinder is crashing partway through. It got to #70 on the first night, and #88 on the next night. There was no error message in PfE - Pathfinder was just frozen, and PfE looked like it was waiting for a response. (On both occasions, the PCD was in, so I'm not sure if the crash occurred at the end of an analysis).

Cheers,
Em

PS I'm on Pathfinder 1.2, v22.

Probeman

  • Emeritus
  • *****
  • Posts: 1883
  • Never sleeps...
    • John Donovan
Hi Emma,
Are you sure you don't mean Pathfinder v. 1.3.22?

I recently updated from 1.3.22 to 1.3.31 and did get an "EDSAcqRunning timeout error" earlier this week, but I haven't reported it to Steve Seddio as I was seeing if it occurred again.

If you are running v. 1.2.x you should definitely update as the 1.3 version is vastly improved over 1.2.
john
The only stupid question is the one not asked!

emma_fisi

  • Post Doc
  • ***
  • Posts: 23
Hi John,

Sorry, I did mean 1.3.22! Ok, I'll see if I can upgrade to 1.3.31 and if that improves stability. Going to try one more long run over the weekend and see how that goes.... (The PFE side of things is great btw!)

Cheers,
Em

John Donovan

  • Administrator
  • Emeritus
  • *****
  • Posts: 2443
  • Other duties as assigned...
    • Probe Software
...The PFE side of things is great btw!...

Cheers,
Em

Hi Emma,
Thank-you so much for saying that.  It was a fair amount of work to get everything seamlessly integrated and glad to hear it's working well for you.

It's really pretty cool to be able to automatically acquire EDS spectra on standards and unknowns and perform high quality quantitative analyses on them, using full standard k-ratios with beam drift and standard drift corrections.

One can utilize EDS quant in PFE with standards on most major elements, and even most minor elements, and get results with similar accuracy to WDS analyses.  The exceptions being trace elements, light elements and elements where there are pathological spectral overlaps. It really comes down to the quality of the peak stripping. 

Though of course PFE does nicely support the use of integrated spectral interference corrections between EDS and EDS elements, EDS and WDS elements, and of course WDS and WDS elements. Someone really should write up a short paper on how well the EDS/WDS spectral overlap corrections work. We've tested the EDS/WDS spectral interference corrections and they work, but we haven't performed any systematic studies of it.

https://probesoftware.com/smf/index.php?topic=482.msg2826#msg2826

But the nice thing is that now we can dedicate our WDS spectrometers for trace elements and do the matrix elements using EDS, which is really the best of both worlds!    8)
« Last Edit: November 10, 2018, 09:35:30 am by John Donovan »
John J. Donovan, Pres. 
(541) 343-3400

"Not Absolutely Certain, Yet Reliable"

Karsten Goemann

  • Global Moderator
  • Professor
  • *****
  • Posts: 179
Hi Emma,
 
We're still running Pathfinder 1.3.22 and doing a lot of combined WD-ED point measuremens with PFE (also automated BSE image collection as part of the runs through Pathfinder) and I haven't had a crash in a while.

Cheers,
Karsten