There is some evidence that combining the instrument system controller network and EIKS network interfaces can produce an increased rate of network collisions
Hi John,
That's interesting, how would you give the EIKS a different IP address, in Probewin.ini I only see
WDS_IPAddress="192.168.0.2" ; 8530 instrument address
WDS_IPAddress2="192.168.0.2" ; 8200/8500 only
Thanks
Ben
Hi Ben,
It's a good question because it can be confusing given that JEOL has "evolved" on this issue.
In the past, JEOL told us they wanted to have the (direct) instrument communications separated from the EIKS (JEOL computer) communications. Apparently they were worried about too much traffic on a single network. So they would have their 8230/8530 instruments set up where the (direct) instrument communications were set up on one subnet, e.g., 192.168.0.x and the EIKS communications set up on another subnet, e.g., 192.168.1.x.
So for example, in the Probewin.ini file one specifies the (direct) instrument connection as you showed above, in the [hardware] section of the Probewin.INI file:
[hardware]
WDS_IPAddress="192.168.0.2" ; 8530 instrument address
WDS_IPAddress2="192.168.0.2" ; 8200/8500 only
But for the EIKS interface one specifies the EIKS network connection to the JEOL computer in the jeolEIKS.ini file (which unlike the probewin.ini file, is found in the Probe for EPMA application folder, not the ProgramData folder!) as seen here:
[Communication]
Async Period=0
Server Timeout=40000
Client Timeout=1000
Server Address=192.168.1.3 <- edit for your JEOL PC computer network address
Server Port=22200
Host Address=192.168.1.3 <- edit for your JEOL PC computer network address
ClientSyncPort=22300
ClientAsyncPort=22300
Note that none of this applies to previous JEOL instruments because for the JEOL 8900/8200/8500 instruments we utilize a *single* network connection (and subnet) for all instrument communications. But on the new 8230/8530 instruments, JEOL insisted that we use the EIKS interface for stage and column control, and use the (direct) instrument interface for the WDS and mapping interface.
Since that time JEOL has "evolved" and now we often find that new 8230/8530 installations use a single subnet for both the instrument and EIKS communications. That simplifies everything and we see no problems with network traffic on this single subnet.
Remember, on 8230/8530 instruments, each application folder (e.g., Probe for EPMA, Probe Image, Remote, etc.) all require their own copy of the JEOL EIKS files which are:
jeoleiks.dll
JeolEIKs.ini
mfc100.dll (used only by JEOL eiksJSample.exe test program)
msvcr100.dll (used only by JEOL eiksJSample.exe test program)
eiksJSample.exe
So at installation time our consultants just copy these files from the Probe for EPMA application folder to any other Probe Software application folders, and properly configured jeolEIKs.ini file is simply a duplicate of the jeolEIKS.ini file in the Probe for EPMA folder. This is necessary because unfortunately the JEOL EIKS client DLL is not multi-threaded, therefore one has to have a separate instance of the EIKS DLL loaded in memory for each application folder. This limitation of the JEOL EIKS interface also prevents one from running two apps from the same application folder at the same time, e.g., Probe for EPMA and Stage.
Someday when we have time we'll implement the 8230/8530 stage and column calls into direct TCP/IP commands (we have reverse engineered them!) and then we won't be dependent on the EIKS interface at all. Then we will be able to use a single network connection as we do for the JEOL 8900/8200/8500 (and Cameca) instruments.
We are a community of analysts, that cares about EPMA