NI-RIO 3.6.0 and 3.6.1 Known Issues

Deploying an NI 9870 or 9871 when the module isn't physically present corrupts the CompactRIO RT controller

If a project is configured to contain an NI 9870 or 9871 module but the module isn't physically in the chassis, deploying the chassis will result in software corruption on VxWorks controllers. This and subsequent deployments will result in the following error: LabVIEW: (Hex 0x80DF0019) The deployment connection has timed out. To determine which operating system your controller is running, refer to Knowledgebase 4LRA4IQ0: What Operating System is my Real-Time Controller Running and Why?

Workaround:  

None.

Back to top

Existing User-Defined Variables stop working after adding new ones

If you make a change to the collection of User-Defined Variables (UDV) and then attempt to run your RT VI, you'll get an expected error telling you to recompile your FPGA VI.  Once you do this, your FPGA VI will load but all UDV will return an "invalid refnum" error.  

Workaround:  

To complete the system update and add the new UDV to the system, you will also need to explicitly deploy these variables.  Right-click on the controller, select Undeploy, then right-click the controller again and select deploy.

Back to top

NI 9075/9076 cannot read TEDS data

If you try to use TEDS functionality with the NI 9075/9076, the FPGA VI will compile fine but the host VI will have a broken run arrow when a Read TEDS method node is placed on the block diagram. 

The error is:

An internal software error has occurred. Contact National Instruments technical support at ni.com/support. Error details:
-61499 : Could not find plugin VI path in resource tree.
Method: Read TEDS
Target Class: cRIO-9076

Workaround:  

This has been fixed for NI-RIO 4.0.0.  If you are planning to use TEDS sensors with a NI 9075/9076 you must upgrade to NI-RIO 4.0.0 or later.

Back to top

RIO Device Setup Utility appears to hang when targeting Linux

When using the RIO Device Setup Utility to download a bitfile to the flash of an R Series or FlexRIO board in a Linux system the download will hang and return a timeout.

However, even though the utility says the download has failed, the download is actually taking place; provided the system is not interrupted, the bitfile will download.

A 6MB bitfile will take approximately 40 minutes to download

Workaround:  

This issue has been fixed for NI-RIO 4.0.0 for Linux.  When the driver is updated on the Linux machine a 6MB bitfile will take approximately 2 minutes and will not return any errors.

If the use of the current driver is a requirement, and you do not want to wait,you can move the physical card to a Window system, download the bitfile, and move the card back. Always remember to power down any system before removing the card.

Back to top

NI 9075/9076 do not support RIO Scan Interface or Hybrid modes

The new NI 9075/9076 integrated chassis and controller CompactRIO devices did not initially ship with RIO Scan Interface (RSI) capability.

However, the "Chassis Properties" dialog box used to select FPGA or RSI mode alludes to the fact that the controller may be used in Hybrid Mode.  

Workaround:  

RSI support was added to NI-RIO 4.0.0.  If you need to use RSI or Hybrid mode with the NI 9075/9076 you must upgrade to NI-RIO 4.0.0

Back to top

NI 9075/9076 have modified DMA specifications

These chassis have slightly different DMA specifications than previous CompactRIO chassis.  These differences are listed in the NI-RIO help file at:
NI-RIO>>C Series References and Procedures>>Devices>>CompactRIO Chassis>>NI cRIO-9075

Keep in mind the following guidelines for writing to DMA FIFOs on the NI cRIO-9075:

--Always check the Empty Elements Remaining indicator from any previous DMA write operation before doing the next DMA FIFO write.

--The DMA engine transfers data in 64-byte chunks. If the host FIFO is full, reading data in amounts that are non-multiples of 64 bytes from either the FPGA FIFO or the host FIFO does not necessarily free an equivalent amount of space on the other FIFO.

--The DMA host FIFO size is limited to 2,096,960 bytes.

Workaround:  

This issue has been resolved in NI-RIO 4.1.0. If you need a DMA FIFO larger than 2 MB on the NI 9075/9076 you must upgrade to NI-RIO 4.1.0.

Back to top

NI 9802 Mount and Unmount methods cannot be called remotely from Windows

Normally, these commands would be run from the RT controller host (or remotely on the RT host through VI Server). For the EthernetRIO targets, NI 9146/9148 this is not an option.  This means that the NI 9802 RT API cannot be used with these targets.

Workaround:  

None.

Back to top

NI sbRIO-9632/9642 boards can output incorrect analog output values on built-in AO channels when used in Scan Mode

When used in Scan Interface mode, the built-in analog output channels of NI sbRIO-9632/9642 devices may output a voltage very different from the commanded value. 

Workaround:  

Use the chassis in hybrid mode by compiling and deploying a user-defined FPGA VI. This issue was also corrected by a patch for NI-RIO in December 2011 and the patch was integrated into NI-RIO 4.1.0.

Back to top

Scan Engine VIs fail on a machine with only LabVIEW runtime & NI-RIO

Scan Engine VIs (such as "Set Scan Engine Mode.vi") return an error when use in a built LabVIEW application on a machine which only has the LabVIEW runtime & NI-RIO driver installed (returns error -2132869114: A Module could not be loaded).  The same LabVIEW exe works fine on a machine which also has the LabVIEW development environment installed.

Workaround:  

The Scan Engine VIs need the DLLs located in:

C:\Program Files\National Instruments\Shared\DeploymentFramework

Copy this directory onto the machine with only LV runtime & NI-RIO, (in the same location) and the scan engine VIs will function properly. This DeploymentFramework folder is installed with the LabVIEW development environment, but not with LabVIEW runtime or NI-RIO driver.

Back to top

Possible timing errors when compiling a VI with a 9476/9477 IO node on a Spartan 3 target at 80 MHz

An IO node with 32 channels and error terminals enabled will not successfully compile on a Spartan 3 (sbRIO-96xx, cRIO-907x) when the top level clock is set to 80 MHz

Workaround:

Setting the Design Strategy to Timing performance in the build specification property for Xilinx Options will allow this to compile.  Alternatively, splitting the IO node into two 16-channel IO nodes will be successful.

Back to top

Getting INTERNAL_ERROR:Xst:cmain.c:3464:1.47.6.4 when compiling a 9205/9206 for a Spartan 3 target with an 80 MHz clock

This is a known issue with the Xilinx compiler that only occurs in rare situations.

Workaround:
Changing parts of the block diagram might fix the issue; alternatively, setting the Xilinx Options for Design Strategy to by optimized for Timing Performance should resolve the issue.  

Back to top

Timing violations with certain modules when compiling for a Spartan 3 target at 80 MHz

When using a 9205, 9206, 9263, 9264, or 9269 module with an 80 MHz clock on a Spartan 3 target (NI 96xx or NI 907x) it’s possible to get a timing failure when compiling.

Workaround:

In the Xilinx options, select the Design Strategy to be optimized for Timing Performance.  

Back to top

DMA Host side buffer failing to allocate memory with error -52000 or -50352

Using a 64-bit operating system (either Windows Vista or Windows 7) and configuring  DMA FIFO host side buffers fail to allocate enough memory and give error -52000 or -50352

Workaround:

Depending on how much memory is available to allocate, there are several things that can possibly resolve this issue.  

Back to top

Setting date/time in a tight loop causes the controller to run at an incorrect frequency

If the date/time is being updated continuously in a loop with a period of less than 2 seconds, it can cause the controller to run slightly faster or slightly slower than it should

Workaround:

There is not ever a need to update the time this often.  To avoid the issue described above, do not set the date/time in a loop, or set it much less frequently.

Back to top

Using Windows Server 2003 R2 may require disabling Physical Address Extensions (PAE)

Support for Windows Server 2003 R2 may require disabling
physical address extensions (PAE). To learn how this might affect your use of Windows Server 2003 and what actions you might have to take, visit <a href="http://www.ni.com/info">ni.com/info</a> and enter the info code <i>PAESupport</i>

Workaround:

Disable PAE for Windows Server 2003 R2 as instructed in the knowledgebase referenced above.

Back to top

Configuring a module in scan mode that has been swapped for another without restarting causes error 1070

If a module being used in scan mode is swapped out and replaced, “Refresh Modules” command will pick up the difference; however, trying to programmatically configure the module results in “Error 1070”

The module is implicitly opening a reference that is not being refreshed

Workaround:

(1) To programmatically prevent the issue, explicitly close both references (IO and module) by calling “Close Variable Connection” from the Shared Variables Palette, before running Discover IO.vi  (2) If the VI is stopped and started again the references will be updated properly.

Back to top

Chassis settings don't automatically deploy

Changes to the chassis properties are not automatically deployed.  This is problematic when switching between Scan and FPGA Interface mode because it means the project can easily get out of sync with the target if the user does not manually deploy the chassis.

Workaround:

Manually deploy settings whenever changes are made to chassis properties.

Back to top

Mounting/Unmounting a NI 9802 in Real-Time interrupts Scan Engine

When you take scan mode measurements and then mount and unmount the SD card, on the subsequent reads the scan mode I/O variable throws error -65536 until the FPGA VI is run.

Workaround:

Instead of the blank FPGA VI required to normally make this work, use an empty while loop inside the FPGA VI and leave this VI running while running the RT VI.

Back to top

Changing counter configuration via module refnum always resets counter

When changing properties such as terminal count and counter direction via the RSI module refnum the counter value always resets.

Workaround:

None.

Back to top

Reading Empty Target to Host DMA FIFO with Timeout Set to Zero Gradually Starves CPU in built LabVIEW RT executables on cRIO targets

In built LabVIEW RT applications on cRIO, if a Target to Host DMA FIFO read executes with a timeout of zero and the FIFO is empty a processor leak occurs that increases the CPU usage on the controller.

Workaround:

Read zero elements to find elements remaining, instead of using a zero timeout

Back to top

Performance issues when using more than 16 FPGA sessions

The NI-RIO driver maintains a growable lookup table of sessions that has segments. If the first segment overflows, new segments are created. RIO sessions using those new segments will pay a performance penalty. The RIO driver sets a warning on any session that has this "problem".

Workaround:

Users should be careful to close unused sessions and look for sessions leaks.  If greater than 16 sessions are needed, please contact NI technical support.  

Back to top

DMA is not supported from the host to the FPGA on cRIO-900x

DMA is not supported from the host to the FPGA target on the cRIO-9002/9004.

LabVIEW returns an error if you try to output DMA from the cRIO-9002/9004.

Workaround:

None.

Back to top

Known Issue

Setting date/time in a tight loop causes the controller to run at an incorrect frequency

If the date/time is being updated continuously in a loop with a period of less than 2 seconds, it can cause the controller to run slightly faster or slightly slower than it should.

Workaround:

There is not ever a need to update the time this often.  To avoid the issue described above, do not set the date/time in a loop, or set it much less frequently.

Back to top

NI 9213 return bad first value when the module is inserted and is already running

If an application is executing and the module is removed and readded, or a new module is added to the chassis while the code to read that module is already running, the first value returned is incorrect.

Workaround:

None.

Back to top

Using Windows Server 2003 R2 may require disabling Physical Address Extensions (PAE)

Support for Windows Server 2003 R2 may require disabling
physical address extensions (PAE). To learn how this might affect your use of Windows Server 2003 and what actions you might have to take, visit <a href="http://www.ni.com/info">ni.com/info</a> and enter the info code <i>PAESupport</i>

Workaround:

Disable PAE for Windows Server 2003 R2 according to the knowledgebase referenced above.

Back to top

Configuring a module in scan mode that has been swapped for another without restarting causes error 1070

If a module being used in scan mode is swapped out and replaced, “Refresh Modules” command will pick up the difference; however, trying to programmatically configure the module results in “Error 1070”

The module is implicitly opening a reference that is not being refreshed

Workaround:

(1) To programmatically prevent the issue, explicitly close both references (IO and module) by calling “Close Variable Connection” from the Shared Variables Palette, before running Discover IO.vi  (2) If the VI is stopped and started again the references will be updated properly.

Back to top

RIO Device Setup utility crashes after running a program with User-Defined variables on cRIO

Once a VI is run with user-defined variables, subsequently running the RIO Device Setup will cause the target to become unresponsive.

Workaround:

Before running RIO Device Setup, (1) download a blank FPGA VI  or (2) Reset the device.

Back to top

FPGA Discovery fails silently on sbRIO if FPGA is in a bad state

Anytime the sbRIO is in the following states: safe mode, unconfigured IP address, or the NI-RIO driver is not installed on the controller, FPGA auto-discovery will fail silently.

Note1 -- If the sbRIO controller is in a good state, auto-discovery succeeds.
Note2 -- If there is some condition that prevents module discovery (not FPGA target discovery) from working during sbRIO auto-discovery, it is reported correctly.
Note3 -- cRIO autodiscovery reports errors properly at every stage

Workaround:

Manually add the FPGA target

Back to top

Chassis settings don't automatically deploy

Changes to the chassis properties are not automatically deployed.  This is problematic when switching between Scan and FPGA Interface mode because it means the project can easily get out of sync with the target if the user does not manually deploy the chassis.

Workaround:

Manually deploy settings whenever changes are made to chassis properties.

Back to top

Mount SD Card (9802) Doesn't return error on Pharlap

The NI 9802 is not supported on Pharlap targets (900x). 

If an attempt to mount the NI 9802 in LabVIEW Real-Time is made it will not work; however, no error is returned.  The drive letter returned is “/” which is the only indication this didn’t work.

Workaround:

None.

Back to top

cRIO-9074 set to Ethernet polling still uses interrupt based Ethernet

When Ethernet polling is enabled on this controller it still gets interrupts anyway, causing jitter

Workaround:

None.

Back to top

9401 output glitches high when set to output and sleep is asserted or FPGA VI is downloaded

The 9401 has a pull-up resistor causing this to happen.

Workaround:

Before going into sleep or downloading a new VI, set the outputs to inputs to prevent the glitch. 

Back to top

Mounting/Unmounting a NI 9802 in Real-Time interrupts Scan Engine

When you take scan mode measurements and then mount and unmount the SD card, on the subsequent reads the scan mode I/O variable throws error -65536 until the FPGA VI is run.

Workaround:

Instead of the blank FPGA VI required to normally make this work, use an empty while loop inside the FPGA VI and leave this VI running while running the RT VI.

Back to top

Reading Empty Target to Host DMA FIFO with Timeout Set to Zero Gradually Starves CPU in built LabVIEW RT executables on cRIO targets

In built LabVIEW RT applications on cRIO, if a Target to Host DMA FIFO read executes with a timeout of zero and the FIFO is empty a processor leak occurs that increases the CPU usage on the controller.

Workaround:

Read zero elements to find elements remaining instead of using a zero timeout.

Back to top

DMA from the host to the FPGA target on the cRIO-9002/9004

DMA is not supported from the host to the FPGA target on the cRIO-9002/9004.

LabVIEW returns an error if you try to output DMA from the cRIO-9002/9004.

Workaround:

None.

Back to top

ID

Known Issue

NiRioScanInterface DMA Channel is listed when using Scan Interface along with LabVIEW FPGA Host Interface

When using Scan Inteface along with LabVIEW FPGA Host Interface, DMA FIFO method nodes display internal Scan Interface DMA Channels

Workaround:

None.

Back to top

DMA from the host to the FPGA target on the cRIO-9002/9004

DMA is not supported from the host to the FPGA target on the cRIO-9002/9004.

LabVIEW returns an error if you try to output DMA from the cRIO-9002/9004.

Workaround:

None.

Back to top

Reading Empty Target to Host DMA FIFO with Timeout Set to Zero Gradually Starves CPU in built LabVIEW RT executables on cRIO targets

In built LabVIEW RT applications on cRIO, if a Target to Host DMA FIFO read executes with a timeout of zero and the FIFO is empty a processor leak occurs that increases the CPU usage on the controller.

Workaround:

Read zero elements to find elements remaining instead of using a zero timeout.

Back to top

Cannot Deploy Scan Interface Modules or IO Variables  with Non-ASCII Characters

When deploying Scan Interface Modules or IO Variables with Non-ASCII characters such as Chinese or Japanese Characters it will report that the Deployment Completed with Errors.

Workaround: 

Rename any Scan Interface Modules or IO Variables using Alpha Numeric  Characters.

Back to top

Abort and Reset Methods do not execute on incoming error

If an error is passed to either the Abort or Reset Methods, they will not execute and the error will be passed through

Workaround:

Check, report, and clear errors manually in any error clusters wired to an Abort or Reset; otherwise, do not pass error clusters to these methods

Back to top

DMA from the host to the FPGA target on the cRIO-9002/9004

DMA is not supported from the host to the FPGA target on the cRIO-9002/9004.

LabVIEW returns an error if you try to output DMA from the cRIO-9002/9004.

Workaround:

None.

Back to top

VIs that use the calibration API should be modified to use the RIO Device I/O control

Applications from versions of LabVIEW older than 8.5.x that use 783XR Calibration - Open FPGA VI Reference.vi should be updated to use the RIO Device I/O control instead of the VISA I/O control.

Workaround:

None.

Back to top

FPGA code generation into an existing host VI is very slow in FPGA Wizard

If you use the FPGA Wizard to generate code, open the host VI, then generate code into the same VI, the code generation goes very slowly.

Workaround:

None.

Back to top

Large configuration files load slowly in FPGA Wizard

Opening a large FPGA Wizard configuration file can take several minutes.

Workaround:

None.

Back to top

Maximum Timed I/O rate value is 70 kHz in FPGA Wizard generated code

The maximum rate you can set for the Single Point Timed I/O timing engine on the Single Point Timed configuration page is 70 kHz. The maximum rate achievable varies with different hardware devices.

Workaround:

None.

Back to top

Selecting external clock requires external clock source with FPGA Wizard

If you select External for the Clock Type on the Single Point Timed or Buffered DMA Input configuration page and you do not have an external clock source connected, LabVIEW appears to hang when you run the generated host VI. Press the Global Stop button to stop the VI or add an external clock source to resume running the VI.

Workaround:

None.

Back to top

NI-RIO 3.6.0 does not install support for the NI 987x serial modules for LabVIEW 8.5.x

For serial module support in LabVIEW 8.5 download and install the C-Series Serial Software  installer.  NI-RIO 3.6.0 will install support for the NI 987x serial modules for LabVIEW 8.6.x and later.

Workaround:

Install C-Series Serial Software from the installer linked above.

Back to top

FPGA I/O Property Node returns wrong value for Module Model Code of NI 9217 module

The FPGA I/O Property Node returns the value 9217, which is the Product ID, not the Module Model Code, of the NI 9217. The Module Model Code is 0x712B.

Workaround:

Look for 9217 in applications looking for the Product ID in LabVIEW 8.2.x

Back to top

Read/Write Controls execute more slowly after upgrade to NI-RIO 2.4.1 or later

Read/Write Control now uses a Call Library Function Node to call the RIO driver. The Call Library Function Node executes additional code if debugging is enabled. To restore the performance to the original level, you must disable debugging in any VIs that contain Read/Write Controls.

Workaround:

You can disable debugging in a VI by going to File»VI Properties»Execution and unchecking Allow debugging. 

Back to top

DMA from the host to the FPGA target on the cRIO-9002/9004

DMA is not supported from the host to the FPGA target on the cRIO-9002/9004.

LabVIEW returns an error if you try to output DMA from the cRIO-9002/9004.

Workaround:

None.

Back to top

DMA in parallel While Loops

If you use DMA in parallel While Loops either on the cRIO-9002/9004 or while accessing the FPGA target across a network, one of the While Loops might hang while the other executes.

Workaround:

None.

Back to top