Jump to content

Rolf Kalbermatter

Members
  • Posts

    3,245
  • Joined

  • Last visited

  • Days Won

    180

Rolf Kalbermatter last won the day on November 26

Rolf Kalbermatter had the most liked content!

8 Followers

About Rolf Kalbermatter

  • Birthday 06/28/1966

Profile Information

  • Gender
    Male
  • Location
    Netherlands

LabVIEW Information

  • Version
    LabVIEW 2011
  • Since
    1992

Recent Profile Visitors

12,170 profile views

Rolf Kalbermatter's Achievements

  1. It's still the same. You can not have multiple tasks accessing the same DAQ card for Analog input. You need to combine the channels into one task and one scan rate and then pick out the data as needed and distribute it to the different subsystems as needed.
  2. Most likely that compile worker is a 32-bit application and you only have 64-bit libgdiplus installed? Or another possibility, the /usr/local/natinst/mono/lib64 directory was not properly added to the /etc/ld.so.conf file and/or ldconfig was not executed afterwards
  3. Couldn't view your VI as I don't have LabVIEW 2020 installed on this machine but you want to basically detect that the boolean you use to find out if you want to write to the file has switched from false to true. An edge detector is very simple to do with a feedback register like this. Change the boolean inputs inversion to detect the falling edge. When the rising edge boolean is true, write your header and then the data, otherwise only write the data.
  4. It's not entirely without its own pitfalls. Different network controller chips support different support for jumbo frame, some older do not support it at all, and all your routers, hubs and whatever in between needs to support it too. It's definitely helpful for full HD and higher resolution cameras and/or when you need to use multiple GigE cameras in parallel, but it makes your setup more sensible to quickly switching to a different ethernet port or replacing intermediate network infrastructure suddenly causing issues.
  5. As mentioned, the cvirte.dll is not version specific. The CVI developers apparently tried to avoid that version incompatibility shenigan. So a newer CVI runtime engine install should work. There is a small chance of a version compatibility bug in a much never version, but as long as it works, it works. 😀
  6. The private method is also available in earlier LabVIEW versions, although not as a VI in vi.lib. I checked for it in 2018 and 8.6 but suspect that it may have been present as early as 8.2 0r even 8.0.
  7. Well I started in April 1992, went to the US for 4 months in May and heard there that there was this big news about LabVIEW not being for Macintosh only anymore, but telling anyone outside of the company would be equivalent to asking to be terminated 😀. They were VERY secretive about this and nobody outside the company was supposed to know until the big release event. In fall of 1992 LabVIEW for Windows 3.1 was announced and the first version shipped was 2.5. It was quickly followed by 2.5.1 which ironed out some nasty bugs and then there was a 2.5.2 release later on that made everything more stable, before they went to go to release the 3.0 version which was the first one to be really multiplatform. 2.2.1 was the last Mac version before that and 2.5.2 was the Windows version. They could not read each others files. This was Windows 3.1 which was 16-bit and still just a graphical shell on top of DOS. LabVIEW used the DOS/4GW DOS Extender from Tenberry Software, that was part of the Watcom C development environment used to compile LabVIEW for Windows to provide a flat 32-bit memory model to the LabVIEW process, without nasty segment:offset memory addressing. It was also the reason that interfacing to Windows DLLs was quite a chore because of the difference in memory model between the LabVIEW environment and the underlying OS and DLLs. Only when LabVIEW was available for true 32-bit environments like Windows 95 and NT, did that barrier go away. NI was at that time still predominantly a GPIB hardware company. A significant part of support was for customers trying to get the various GPIB boards installed on their computers and you had at that time more very different computers architectures than you could count on both hands and feet. There was of course the Macintosh and the IBM compatible computers, with all of them running DOS which Windows computers still were. Then you had the "real" IBM computers who had abandoned the ISA bus in favor of their own, more closed down Microchannel bus and also were starting to run OS/2 rather than Windows and about a dozen different Unix based workstations all with their totally incompatible Unix variant. And then even more exotic beasts like DEC VAX computers with their own expansion slots. Supporting those things was often a nightmare as there was literally nobody knowing how these beasts worked except the software driver developer in Austin and the customers IT administrator. NI had just entered the data acquisition marked and was battling against more established manufacturers like Keithley, Data Translation, and some other small scale speciality market providers. The turning point was likely when NI started to create their own ASICS which allowed them to produce much smaller, cheaper and more performant hardware at the fraction of the cost their competitors had to pay to build their own products and still selling them at a premium as they also provided the full software support with drivers and everything for their own popular software solutions. With other manufacturers you usually had to buy the various drivers, especially for NI software, as an extra and some of them just had taken the blueprints of the hardware and copied them and blatantly told their customers to request the software drivers from their competitor as the hardware was register for register compatible with theirs. The NI ASICS made copying of hardware by others pretty much impossible so NI was never concerned about making their drivers available for free.
  8. When I started at NI Switzerland in 1992, things were indeed very different. For 4 months I went to Austin and I would get technical support calls that I of course had no idea how to solve. But we could walk up one floor or two and talk directly with the software or hardware engineers that were responsible for the product in question. As NI grew, this support model wasn't quite supportable anymore. Engineers still usually started out in support and often moved rather sooner than later to another position and walking up to the developers wasn't as simple as they weren't always just one floor higher but in a different building. Support still was handled by inside engineers though, usually with a tiered level, first support was handled by first line supporters who would respond to the standard problems, and if it got to complicated it moved up to 2nd or 3rd line support. Then it was outsourced to some extend to telephone support somewhere in Costa Rica or wherever and from then on it was often pretty much impossible to get any meaningful response. The answers we would get were sometimes so abysmal that I could have asked my son and would have gotten more meaningful information and it was often impossible to explain to them the problem, as they understood nothing but what their onscreen step for step problem solving tutorials told them. Then a few years back, like 2 or 3, NI recognized that this was not really working and changed to a more professional support infrastructure with a dedicated Technical Support Engineering model that actually deserved that name. If someone has a support contract or maintenance software contract, then this support works again very well, although in comparison to 25 years ago, it is practically impossible to get non official solutions that are just gathered by the support person from walking up to the actual developer who would throw together a quick (and sometimes dirty) solution to achieve the goal. Things are much more formalized, and unless someone is from a huge multi-million $ account, it's impossible to get a bugfix release of software or something like that, before the official release.
  9. LabVIEW realtime and LabVIEW FPGA are pure Windows (32-bit only before 2021) extensions. The Linux downloads you found are for the Xilinx toolchain. You can install them under Linux, then setup your Windows FPGA compile interface to interface to them just as it does when you use the NI cloud compile solution.
  10. ISO8601 just as it's semi-sibling RFC3339 also supports an optional timezone offset identifier. Mr. Powels library deals with that and you should probably use his library. Basically if it is not present you need to treat the timestamp as being in local timezone (which can be tricky if interpreting data that was acquired somewhere else). Local Timezone offset means using the -1 indicator for the DST value in the cluster, and yes LabVIEW still needs to learn to listen to the OS message WM_SETTINGCHANGE that indicates among many other things a date/time attribute change as requesting that information every time from the OS when it deals with time conversions would be to costly.
  11. They also support a shared library interface!
  12. Typically, writing such a beast isn't exactly rocket science. But validating it is, and whoever could send you some code doing this, it would be: - Just implementing as much as they needed at that point and everything else might be not there or wrong - Almost certainly not validated: "Look mam it works, without hands!" "Great son, now don't touch it!" So the major part of the work would still be on you unless someone had a real product that they did some good validation on, and will cost considerable money. Writing your own will be unsexy: Who the heck is still writing assembly binary files anymore? 😀, but it is most likely unavoidable in this case and the real work is in the validation that you mention and that will have to be done anyhow if your product is in such a regulated industry.
  13. It's pretty unclear what you really try to do. You mention CLNF and all that and have a DLL but you talk like you would like to implement everything in LabVIEW. My suspicion is that you still want to call UA_server_run(), but from LabVIEW and the bad news in that case is that you can't do what you try to do. LabVIEW is dataflow driven. While it allows to call C functions, it doesn't and can't implement the full C semantics. Once it passes control to the UA_server_run() function in the Call Library Node, this code executes on its own until it decides to return. But since LabVIEW is dataflow driven and doesn't know pointers like in C, you do not have access to the memory location the running boolean is stored at that you pass to the function. If you branch the wire, LabVIEW will very "helpfully" create a seperate copy of that variable, not just pass a pointer around, since passing pointers around would completely go against the entire dataflow paradigm. You will have to create a wrapper DLL that implements and exports the stopHandler() function as well as a function that calls UA_server_run(), passing it the DLL global "running".
  14. That's quite a blanket statement. I'm sure it can be done, not directly but the myRIO does have a Wifi interface, and all cRIO have wired Ethernet. So if you have a Wifi router that you can let the myRIO connect to, you can certainly connect it to the cRIO which all have wired Ethernet in one way or the other. It also support Ethernet over USB but that is generally only easily usable to connect from a computer to the device. In order to connect this to another device like a cRIO connected to the network, you would have to somehow bridge the virtual USB network adapter with the normal network interface. It's possible but not straightforward.
×
×
  • Create New...

Important Information

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