Scan Engine & EtherCAT Custom Device Feedback

[StephenB]

yeah we haven't impelemted the buttons or right click menus but you can always go to the scale node and select "map scales" at the top to assign scales to any scalable channel like ours

[Holt]

That worked.  I was unaware of this feature for scaling.  Thanks.

[joshe]

I would like to upgrade from a 3.5 version to the latest release.  I have a system def that I don't want to have to rename and re-map everything if I install the latest update.  Can I install the update without affecting the current system definition?

Thanks!

[StephenB]

If you install the latest update and then open your sysdef(s), save, and close, everything should mutate up automatically and no paths will change.

[joshe]

I shut down VeriStand, copied the Scan Engine directory in to my Custem Devices folder then restarted VeriStand.  I opened up the sys def and then the custme device made a text change, saved it.  Exited to the VS start screen, then reopened the sys def.    When I click on the custom device it shows a version of 3.7 so it looks like it didn't update to the 4.1.2.  Is there something else I need to do?

[StephenB]

You shouldn't even have to make a change. Just opening the system explorer should show a dirty dot because mutation automatically happens up to 4.1.2. You should see the new version there.

Can you close NIVS and make sure that there is no other old scan engine files in your custom devices directory?

[Inertia]

So I've run into a few issues: (using 9144 with NI-9206 module, desktopPC RT system, as far as I know I have installed all the latest drivers for 2013)

1. If I run using FPGA / Scan interface mode=use current (or FPGA) then I can get data but my loop rate tops out around 500hz,  going beyond this rate the EtherCAT chassis appears to fault out.  If I write a LabVIEW program to acquire data, I can get over 10khz before faulting.  Also, CPU usage is <5% on all 4 cores.

2. Updated - EtherCAT Scan Mode not available in VeriStand ( If I select the Scan Mode option, the deploy errors out stating Error -65587: "CompactRIO:  The bitfile required by the RIO Scan Interface is either missing or corrupt.  Reinstall the software on the LabVIEW Real-Time target."  I have reinstalled and repaired a few times with no change.)

[Thør]

1.  Before we investigate too far into the loop rate, I want to make sure you're looking at the correct channel.  The scan engine add on can run asynchronously at a rate different from that of your primary control loop.  The "system channels" section can tell you a lot of info about the primary control loop, but the "Async Loop Duration" channel will tell you about the scan engine directly.  Does this channel have a different value?

2. There's a way to force the redownload of the default bitfile for the RIO scan interface right from the system definition.  I'd like to have you try doing that and let me know if it succeeds or throws an error.  The way to do this is to browse to your EtherCAT master >> Slave >> right click on User Variables.  In this menu, select "Revert Slave to Scan Mode."  This should make the slave attempt to download the scan mode personality again.  What does clicking this option do for you?

[Inertia]

So Async Loop Duration is around 13 usec (acquiring 32 channels on a NI-9206, Slot 1, 500 Hz PCL rate).  Async Loop Late Count = 0.  This is what I would consider "correct" operation, except when I set the PCL to say 1khz. I get the error "This module cannot be configured while the controller engine is in its current mode."

I show the same right click menu as you posted and when I select Revert Slave to Scan Mode, I get an error stating the slave device is already in Scan mode.

[StephenB]

See the known issues list. I think you need the RIO 13.0.1 patch. It fixes an issue like this. Let me know if it resolves the issue.

You should probably also run in sync mode, not async mode, unless you are also using single point data from a R series or PCI/PXI DAQ card. To turn on sync mode click the "sync" checkbox on the scan engine custom device main page.

Its also probably not possible to run at 10kHz with a 9206. So your LV test might have been invalid. Make sure you check the error wire on the read as well as for scan faults in the distributed systems manager. I say it might not be physically possible because even if the module was changed from the default of all RSE to all differential (so there are only 16 lines to read) and you change the default channel-to-channel gap from 8uS to 4uS (the max)... that is a total of 64uS to digitize the module... and then there is a the software scan time and the transfer delay over EtherCAT... that all has to be less than 100uS. If you left the defaults (RSE and 8uS delay), then it certainly isn't possible.