Jump to content

Phillip Brooks

Members
  • Posts

    875
  • Joined

  • Last visited

  • Days Won

    42

Everything posted by Phillip Brooks

  1. Do you understand the protocol? You need to know the basics of sending vs receiving before coding your solution. Start with this... http://wiki.synchro.net/ref:xmodem#initial_handshake_rules
  2. Sounds like a perfect use for a joystick. LabVIEW supports a joystick.
  3. Haven't used winzip for ages. What a piece of garbage it is now :(

  4. Even at 2400 baud, you should be able to easily send 10 characters in 10ms. Your requirement indicates 4 characters if we include the conditional decimal point. Add in a terminator and maybe a U16 as a message index and you should still be fine. The question is, what does the other end accept? Is it another LabVIEW app, a SCPI instrument or something like an Arduino with limited resources? LabVIEW to LabVIEW could use SI notation (%_3p) which would always be exactly 4 characters for the data. SCPI devices normally include support for scientific notation (%_3e) which would always be 7 characters in length. If you have a custom system, you will need to understand the parsing routine or capabilities in order to make the best choice. If you are writing the remote device code, simply multiply your value by 100 in LabVIEW and send the value as 00000 to 99999 and divide by 100 on the remote side.
  5. SI notation? http://zone.ni.com/reference/en-XX/help/371361J-01/lvconcepts/format_specifier_syntax/ %_3p
  6. A few years ago the forums experienced a database corruption and many of the attached files were lost or their links to specific posts were lost. An admin may still have some ability to search for this file, but its been a long time. I checked the OPs status and they haven't logged into LAVA in several years.
  7. Wishing that the diags I use were written for testing the system, not debugging application code.

  8. If you use the "Clean up Block Diagram Feature", the default spacing can be excessive. Look under Tools -> Options... Block Diagram and scroll to the bottom of the pane and you can experiment with various types of spacing. There is an option to adjust spacing to reduce bends in wires; I forget if it defaults to OFF or ON. If you're scrolling more than one screen up/down or left/right then you may want to consider a state machine or sub-vis.
  9. Added to The Idea Exchange... http://forums.ni.com/t5/LabVIEW-Idea-Exchange/Set-quot-Context-Help-quot-using-VI-Reference/idi-p/3094027
  10. Been there, done that... https://lavag.org/topic/7328-programatic-control-of-the-context-help-window/ https://lavag.org/topic/9167-context-help-for-a-vi-without-loading/
  11. Maybe the customer VI contains multiple event structures and becomes somehow deadlocked?
  12. I remember using a VI Analyzer plug-in years ago to validate my driver. I'm not sure if that is now included with VI Analyzer or maybe its in the Driver Design Studio. The first thing an NI employee would likely do after receiving a driver for validation would be to test it against the requirements above using VI Analyzer. No sense submitting a driver for certification if it can't pass the automated test...
  13. Inspiration? http://en.wikipedia.org/wiki/Category:Messenger_gods I sort of like Isimud
  14. LV2011 is not listed as supported with Windows 8. I know that doesn't mean the code won't run, but the behavior sounds similar to a problem I saw when trying to run LV 8.6 on Windows 7. Specifically, I had UI code that responded to various events that worked fine under XP, but locked up my app under Windows 7. I think one of the events was "Application Instance Close?" This specific event might be used as an exploit by some viruses to force the close of antivirus software. Windows changes the interface and your 'unsupported' version of LabVIEW fails the OS call.
  15. You might also want to consider a license server. It seems that you have more licenses and stations than programmers. With a name based licensing setup, you could load LabVIEW on all potential test stations and program anywhere. If you have test equipment that you need to interface to, access to a locally installed LV instance is much easier than using a remote desktop to your dev machine and Remote VISA back to the machine you are working from. The only limitation is that you can't be logged into multiple LabVIEW instances as the same user name simultaneously. In an R&D or continuous manufacturing environment, you might need to make an update to a test in situ. That may be why you have local licenses installed. With network licensing, you could log into the network, load LabVIEW, open your source code from a network location (source control assumed), make the change, compile and build. Install the new build, test and leave the station running the newly updated EXE. If your test stations are "off the network" then you've got other problems... Just checked the NI site and it looks like LV2011 does not support Windows 8. Windows 7 for business will be supported for a while, but if your company upgrades to Windows 8.x or Windows 10, you might not be able to run your LV programs reliably....
  16. The OP and James mentioned running a LabVIEW server app without an X server to host it. If you had a minimal platform without a GPU, you might be able to use Xvfb to host the GUI and run a linux based LV app on embedded modules like Intel Galileo.
  17. Could a dummy X server be used? http://en.wikipedia.org/wiki/Xvfb
  18. Sophos doesn't like it either... High Risk Website Blocked Location: www.siphos.net/Products Access has been blocked as the threat Mal/ExpJS-N has been found on this website. Return to the page you were previously viewing.
  19. Locals are WiFi, Globals are Cellular Data. If you want performance, wired is always the fastest and safest
  20. I guess I was stuck in 8.6 for so long that I never learned this. Thanks!
  21. I'm rewriting a test for a family of products that include multiple 10/100/1000-T Ethernet interfaces (2 or 5) built into our system's main board. The existing test code uses CLI commands to query the operational state, speed and duplex state of the interfaces. It turns out that the various standard linux style commands don't really return the state of the interfaces; they always appear as UP and 1000 and full. No one performed negative test cases or combinational testing to see if the auto-negotiation was working. There is a second 'ping' test where the Ethernet ports are assigned IP addresses and a second NIC in the test station is connected to a hub that is in turn powered on and off to see if the ports pass traffic. A bit cludgy, but that seems to work. I found I can use a REST interface on the UUT to query the current state of the interface, but I need programmatically configurable source ports to 'wiggle' the inputs. I found this TP-LINK TL-SG3216 managed switch that includes a CLI that can be accessed via telnet, but I am a bit concerned about the long term reliability and repeatability of a consumer grade device being used as test equipment. I'll be testing a worst case of 50 UUTs a day, and would expect this to be in place for a few years. I'll have between 4 and 8 stations in various locations. Anyone have any experience with Ethernet interface testing that they would like to share?
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.