09-12-2017 08:50 AM
Hello,
I have been using PCI-8512 CAN interfaces with no trouble. Now I have to use USB-8502 CAN interfaces for some system and I am having a lot of trouble. I keep getting the following error from XNET Create Session (Single Input Waveform).vi.
Error code: -1074384887
Possible reason: NI-XNET: (Hex 0xBFF63009) The board's firmware did not answer a command.
Solution: Stop your application and execute a self test. Try deactivating/reactivating the driver in the Device Manager. If the problem persists, contact National Instruments.
The error is not very consistent. It happens many times when I begin to run the LabVIEW program but eventually it works after I try doing restart / unplug / plug kind of stuff. But then if I run both ports, data from port 2 doesn't seem to be streamed stably. I never had any problem like that when I was using the PCI interface with the same LabVIEW program. Did anyone have a similar experience? Is the USB version not supposed to be the same as the PCI version except for the connection type?
09-12-2017 04:01 PM
The primary differences lie in the underlying bus of each device. PCI is more deterministic and has more throughput than the non-deterministic USB that has Windows as a gatekeeper. Aside from that, the devices have almost identical firmware, driver stack etc and should be near equivalents.
Are you connecting the USB-8502 to a USB hub with other devices? Or directly into the PC? Does plugging into a different port on the back of the PC improve performance at all?
Can you provide us with a general idea of what your application is doing? Or can you create a small test application that reproduces the problem with a similar architecture to your main app? If I can run a test application I may be able to reproduce the problem and get a better idea of what is going on under the hood.
09-12-2017 04:26 PM
Hello Jeff,
The following shows the CAN devices on MAX.
I created a simple VI to test both XNET CAN devices. First, I tried using the USB-8502 device (CAN1 & CAN2).
I ran the VI many times. I saw three different results as shown below. It was unpredictable which result I would get next. I tried all three USB-8502 devices that I recently bought. They all showed the same problem.
Then, I tried running the PCI-8512 device (CAN3 & CAN4) using the exactly same VI. I never get any errors. In fact, I have used many PCI XNET CAN cards in many projects with no problem. This is my first time to use the USB devices and they are giving me huge trouble. FYI, I am using LabVIEW 2017 and XNET 17.0.1.
Thank you for your help.
09-13-2017 08:08 AM
@JefeL wrote:
Are you connecting the USB-8502 to a USB hub with other devices? Or directly into the PC? Does plugging into a different port on the back of the PC improve performance at all?
I forgot to answer these questions. I am not using any USB hub. I tried using different USB ports on two different computers (a desktop and a laptop) but I didn't see any difference.
09-13-2017 09:04 AM
I have been using USB-8205 and PXI-8512.
When I use USB I also change setting in device manager not to let computer to manage power state. I found out when I run long test I started loosing USB connection.
Most of my code for USB CAN is serial.I read CAN1, then CAN2. Usually I do not have any issues getting the data. Not sure is you hitting bandwidth of USB. I also use them only with USB3.
09-13-2017 12:59 PM - edited 09-13-2017 01:16 PM
I see that SS mark on the USB port so I think it is 3.0. In this simple test, the required data rate is minimal anyway. I tried it with serialized XNET Read's, too, but it didn't make a difference.
I made some changes to the code to allow testing a single CAN port and get some time information. I also disabled some parts to simplify it. In any case, I was still getting the same errors. Sometimes, it takes several seconds for the VI to end after completing the data loop. The behavior is just very inconsistent. I feel I get an error 4 or 5 times out of 10 trials.
Occasionally, I get an error from XNET Read, too. The error code is different. The following screenshot was taken when I was testing with one CAN port. You can see that it passed the initialization part but timed out in the data acquisition loop.
09-13-2017 01:48 PM
Regarding the error -1074384895
09-13-2017 03:26 PM
Thank you for all the detail. This helps a great deal in our efforts to reproduce the problem. So far we haven't been able to reproduce it but we are trying it out on different computers with different USB chip sets etc.
To help narrow it down, can you tell us what OS you are using and by chance the USB chipset?
09-13-2017 04:02 PM
Thank you for looking into it. I tried using a laptop and a desktop. Both have Windows 7 SP1 and have the same problem with the USB-8502. The screenshots below are from my laptop.
What information for the USB would be useful from the Device Manager?
09-14-2017 10:04 AM
In the screenshot above, was the USB-8502 plugged into the computer when the picture was captured?