VME device, invalid address

[sbattazz]

Hi there, I have set up a PCI-MXI-2 card along with a VME-MXI-2 installed in the first slot of our VME crate.
To start, I am trying to access just one VME device in the slot next to the VME-MXI-2.
When I pulled this card out of the crate, the address was set to 0x02900000 and I decided I would leave it there to start out with. In vxiedit, I created a VME device with this base address, but whenever I try to read anything from the device in VIC to test it out, I just get invalid address. Am I missing something here? Do I need to explicitly specify this address window to the VME-MXI-2 or something?

I did find this bit of info on the Resman explanation page:
"Do I need to run Resman if I only have VME devices in my VME Chassis?
No, RESMAN is a VXIbus utility and performs no operations on VME devices. However, RESMAN does perform an important task on the VME-MXI devices. Resman programs the window registers of each VME-MXI to allow bus accesses to go into the mainframe. If you have only a single mainframe system, then you need to write the value 0x4700 to the 0xA register of the VME-MXI, and write the value 0x6000 to the 0xC, 0xE, and 0x10 registers of the VME-MXI, before you can access any other devices in the system. If you have more than one mainframe, consult the VXI-MXI User Manual or VME-MXI User Manual, or contact National Instruments technical support for more information on configuring the VME-MXIs."

I have noticed that the values I get when I use VIC to read these 0xA, 0xC, 0xE, and  0x10 registers are not what this little explanation says needs to be there. Is this the problem? And shouldn't Resman be setting up the correct values when it sees that I only have one frame?


Thanks,

Steve

[sbattazz]

A thought-- perhaps I should just use the VISA VXIO:MEMACC resource instead of VXIO:(pseudo LA):INSTR?

Is there any advantage to one or the other?

[John E]

Hi Steve,
It sounds like you’re having some difficulties setting up a VME Device.  A resource that may be helpful to you is:
http://www.ni.com/support/vxi/vmeinfo.pdf
Particularly the section titled "Why do I need to configure VME devices?"

In addition, what version of the NI-VXI driver are you using?  3.3.1 is recommended
Also what Operating System?  If Windows 2000 or newer, you can check and see if you can add the device in MAX.
You can also try using VISA Interactive Control and see if you are able to communicate with the device (for this you will need the VISA resource name)


Regards,

[sbattazz]

Hi John--
I am using Red Hat Linux 4.2.
When I did some writing/reading to/from the memory allocated to the device using the VXI:MEMACC resource, the device responded as expected, I just can't get it to work in VIC or with the VXI:257:INSTR resource.
This may be ok-- as long as we can write up some VIs to exchange data, I don't know that I need to do anything futher.

In VXIedit -> Non-VXI device configuration editor, I allocated a slot of A24 memory space according to the base address I set on the rotary DIP switches on the device, that's pretty much what I was supposed to do right?

Thanks,

Steve

[sbattazz]

Oh, and I forgot to mention a problem we're having with the GUI versions of various utilities in Linux-- for some reason I can't enter any text or numbers into the entry fields.
For VXI Edit, and VIC, I've alleviated the issue by using the text versions of them.. does a similar thing exist for VISA interactive control?

[Gavin F]

That text entering behavior is a known issue, and has been fixed in NI-VISA 4.0 and later. 
The workaround on previous versions is to turn Num Lock off.
Installing the latest version of other drivers should also fix the issue.

[Gavin F]

To answer your question, there is not a text version of VISA Interactive Control for Linux.

[sbattazz]

Thanks!