Advantages/Disadvantages NI, LabView, etc.

[skipjackrc4]

There have been major problems with almost every piece of NI hardware I've bought (and it's their high end stuff, not the lower-cost instruments).  Their highest end digitizer has aliasing problems and cannot reliably trigger at low sample rates (and a Keysight DSOX3000 that was 1/5 the cost of the NI triggers perfectly on the same signal), the vector signal generator disables output power regulation when modulation is enabled (we were getting errors of up to +/-4dB whenever something as simple as AM was turned on, and if you look closely, the manual actually states that this will happen), their RF switches have mediocre isolation specs.  The icing on the cake is that each PXIe card cost more than the equivalent Keysight/R&S/Tek instrument that would have better specs and actually work properly.  For simple things their instruments would probably be fine, but in my mind they're not suited for anything beyond the simplest measurements.

LabVIEW is a pretty good tool for quickly throwing together some simple instruments with a passable GUI.  For anything more involved, it's just very clunky and buggy.  I'm fortunate enough to have an NI certified LabVIEW contractor who develops and maintains our code, but even with his 20 years of experience, we run into stupid bugs and limitations in the language all the time.  I'm often tempted to scrap all of our LabVIEW software and rewrite it in MATLAB or Python.

[joeqsmith]

There have been major problems with almost every piece of NI hardware I've bought (and it's their high end stuff, not the lower-cost instruments).  Their highest end digitizer has aliasing problems and cannot reliably trigger at low sample rates (and a Keysight DSOX3000 that was 1/5 the cost of the NI triggers perfectly on the same signal), the vector signal generator disables output power regulation when modulation is enabled (we were getting errors of up to +/-4dB whenever something as simple as AM was turned on, and if you look closely, the manual actually states that this will happen), their RF switches have mediocre isolation specs.  The icing on the cake is that each PXIe card cost more than the equivalent Keysight/R&S/Tek instrument that would have better specs and actually work properly.  For simple things their instruments would probably be fine, but in my mind they're not suited for anything beyond the simplest measurements.

LabVIEW is a pretty good tool for quickly throwing together some simple instruments with a passable GUI.  For anything more involved, it's just very clunky and buggy.  I'm fortunate enough to have an NI certified LabVIEW contractor who develops and maintains our code, but even with his 20 years of experience, we run into stupid bugs and limitations in the language all the time.  I'm often tempted to scrap all of our LabVIEW software and rewrite it in MATLAB or Python.

I have some horror stories about attempting to use their hardware as well.  Same for working with their "tech" support.   I would never attempt to use their hardware again outside of maybe a one off prototype.   

It would be interesting to know more details about the clunky, buggy Labview problems and what is a simple setup and passable GUI.   I've provided several videos in this thread with examples of how I have used it with success.   I am not NI certified and have no formal training in Labview.  It's been a learn as I go sort of thing.   

It's rare I run into limitations.  One I came up with is I wanted to basically have a spreadsheet and wanted to be able to use the mouse to edit their values.   Basically this was to edit engine tuning maps.    It's something I would have thought would be built into one of the widgets.   In this case, coding what I wanted was indeed what I would consider clunky.  It was also what I consider slow.   The following short video gives you some idea what I was after.   


****  Just to be clear, the video is not showing some full on program.  The GUI in this case is only to demonstrate the table entry.   The plan was to use the table editor as part of an actual program.  Also, this was from a few years ago and it is very possible NI came out with a better way to handle tables since then.   

https://youtu.be/EgLwrH3zrEw

[jeremy]

I use the comedi open source drivers (not associated with NI at all) for the DAQ cards on Linux (also with python). I don’t exactly push the features to the max, but I’m yet to have a problem with them. If your hardware is supported, you should give it a try. These are older cards though.

[Gribo]

I am using both LabView and NI PXI hardware for my day job. If one follows the style guide, their code will not be a mess of spaghetti. It is the same with text based languages, you don't want a 10000 lines monsters, you want small, maintainable files.

Yes, a PXI DMM or scope don't make much sense, but a 32 channel scope in a 4U Chassis makes sense for certain applications. 

[joeqsmith]

I am using both LabView and NI PXI hardware for my day job. If one follows the style guide, their code will not be a mess of spaghetti. It is the same with text based languages, you don't want a 10000 lines monsters, you want small, maintainable files.

Yes, a PXI DMM or scope don't make much sense, but a 32 channel scope in a 4U Chassis makes sense for certain applications.

True, I am sure we have all seen some pretty messy unmaintainable text based designs.  I may have contributed to that! 

For work, we used some of NIs VXI (hardend VME) hardware for a similar reason, lot's of I/O in a fairly compact chassis.  A bit of a side note. This thread seems to segregate the various tools.     For work, we would use a combination of Matlab, LabView and C on a single project. 

[skYfIrE]

Hi Joe,

I'm not sure to follow all you do :
- Did you develop yourself these Labview softwares ?
- Are you a Labview developer or/& architect ?
- Is it your job or do you make Labview Softs for a private passion ?

If yes, how much time is required to get a good skills level to develop a full Labview software ?

[joeqsmith]

Hello,

- Did you develop yourself these Labview softwares ?
All of the programs I show in this thread, I developed.

- Are you a Labview developer or/& architect ?
I use Labview as a tool, like any other tool.  Do I develop programs in Labview, yes.   

- Is it your job or do you make Labview Softs for a private passion ?
Everything I show in this thread is for my personal hobbies.  It has nothing to do with my job.   However, as I mentioned in my previous post, I do use Labview for my job as well.

-If yes, how much time is required to get a good skills level to develop a full Labview software ?
I can't answer that for you or anyone else.  We all have different education levels, skills, and ability to learn.  I am also not sure how to define good.   

On the PC, I started out learning assembler, then mainly used C.   I developed code for the PC for several years before using Labview.   It took me about a week of going through the manuals before I started using it for actual projects.    After several years of using it, I continue to learn better techniques and tricks.  I would say I am much more fluent with it now.   

[frogg]

But the compactRio modules are low priced and would be a good alternative for standardized and repeatable measurements in the private sector (?). The advantage is supposedly the high number of modules and the fast integration into the test environment.

Nevertheless, when i see the many test equipment here on EEVBLOG, which is sometimes very high priced, i wonder why such environments like LabView is not seen more often.

Probably the wrong tree.

First, I wouldn't call compactRIO particularly inexpensive.

Second, NI isn't compactRIO. CompactRIO is just a package. PXI is also good.

Third, NI isn't just Labview. Labview has strengths and it has weaknesses. NI has a lot of other useful software products besides Labview.

I built a startup on scanning light microscopy. We needed several multifunction DAQs (regardless of the number of ADC's, the NI DAQs only have a limited number of control channels), along with a high sample rate reconfigurable I/O module. We used the reconfigurable I/O module to act as a high sample rate digitizer via the on-board FPGA.

NI's Labview provided a good way to perform the FPGA synthesis and build the digitizer quickly. We did not use Labview for anything except the HDL synthesis. We had good C/C++ windows library support and MATLAB interfaces with NI's API's. All stable enough software to build our entire toolchain.

The CompactRIO chassis allowed us to consolidate most of the scanning hardware to a single backplane. Before that, it was a rats nest of wires that only appealed to the disheveled mad scientists.

The NI support is pretty good, as well. They have a big staff of application engineers and the online support is quick. That said, if I was a lone individual trying to get something done on a shoestring budget, I wouldn't consider NI to be a first choice.

Just wanted to add - we moved to PXI chassis for the high resolution / high frame rate imaging...the ADCs were much faster on the PXI's but we didn't have to do hardware rework when moving to PXI from CRIO.

Quite frankly, the alternatives for high-sample rate 14-bit ADC digitizers can be expensive - the NI digitizer solution saved us money. The compactRIO was just a convenient way to package everything in a single chassis. The PXI is really awesome, more capability but larger and more expensive.

[Nerull]

But the compactRio modules are low priced and would be a good alternative for standardized and repeatable measurements in the private sector (?). The advantage is supposedly the high number of modules and the fast integration into the test environment.

Nevertheless, when i see the many test equipment here on EEVBLOG, which is sometimes very high priced, i wonder why such environments like LabView is not seen more often.

Probably the wrong tree. While I'm very new here, from what I have seen, many of the strongest opinions on this forum are not representative of reality, real companies, or real work. Usually, the strongest opinions seem to be tied to personal ideologies or axes to grind taken to obsessive extremes. Take everything with a grain of salt.

I think you're on the right track, not just here, but life in general. The strongest convictions tend to accompany the least knowledge. I've been using linux for 20 years but I don't hang out in linux user communities because of the large concentration of users who can barely use ubuntu but believe beating others over the head with the fact that they "use linux" (or, have a live cd they booted once) makes them morally superior to some other group of users.

Its best not to get caught up in the zealot approved solution to any problem, but rather the real best tool for the job. The hard part is picking that out from the noise.

[skYfIrE]

Hi Joe,

I asked you some questions because I'd like to start a NI project but I'm a bit confused with all possible solution.
I've contacted NI support but they absolutly don't know anything about Cam or Crank sensors topics. Your work exhibits a high level of Labview skills
Maybe could you help me ?

Hello,

- Did you develop yourself these Labview softwares ?
All of the programs I show in this thread, I developed.

- Are you a Labview developer or/& architect ?
I use Labview as a tool, like any other tool.  Do I develop programs in Labview, yes.   

- Is it your job or do you make Labview Softs for a private passion ?
Everything I show in this thread is for my personal hobbies.  It has nothing to do with my job.   However, as I mentioned in my previous post, I do use Labview for my job as well.

-If yes, how much time is required to get a good skills level to develop a full Labview software ?
I can't answer that for you or anyone else.  We all have different education levels, skills, and ability to learn.  I am also not sure how to define good.   

On the PC, I started out learning assembler, then mainly used C.   I developed code for the PC for several years before using Labview.   It took me about a week of going through the manuals before I started using it for actual projects.    After several years of using it, I continue to learn better techniques and tricks.  I would say I am much more fluent with it now.

[joeqsmith]

Hi Joe,

I asked you some questions because I'd like to start a NI project but I'm a bit confused with all possible solution.
I've contacted NI support but they absolutly don't know anything about Cam or Crank sensors topics. Your work exhibits a high level of Labview skills
Maybe could you help me ?

NI support "absolutly don't know anything about Cam or Crank sensors topics" is your opinion based on the few people you spoke with, but it's not a fact.   You are not talking to the right people is all.  They have a fairly large forum.  I suggest you do a little homework before calling them or posting generic questions on their forum.   Google  "labview HIL engine" and start reading.     

You could also consider D-Space.  Applied Dynamics may still be around.   Depending on your requirements (schedule, budget) D-Space used to offer turn key solutions.   If you know exactly what your requirements are, you could get a quote from them.

You could post about your project here as well but it seems there are not many people using the tools.  Being an electronic forum, you may not find many people interested in engine controls and such. 

*****

https://www.dspace.com/en/inc/home/products/systems/ecutest.cfm
https://www.adi.com/

A few NI links from a Google search:
http://www.ni.com/example/31187/en/
http://www.ni.com/en-us/innovations/white-papers/17/what-is-hardware-in-the-loop-.html
http://www.ni.com/en-us/innovations/white-papers/18/automotive-hil-testbed-reference-architecture.html




 

[skYfIrE]

Hi Dr Joe,

I'll be more accurate : People from NI France I've spoken with were not aware about what is a cam or crank sensor.
I'm not a NI forum user but I've seen example and white papers you've mentionned. I think so partners developers from NI know exactly how this kind of sensor work.

I've contacted dSPACE and they'll make an offer for my topic

Thanks a lot for your info !

[joeqsmith]

Hi Dr Joe,

I'll be more accurate : People from NI France I've spoken with were not aware about what is a cam or crank sensor.
I'm not a NI forum user but I've seen example and white papers you've mentionned. I think so partners developers from NI know exactly how this kind of sensor work.

I've contacted dSPACE and they'll make an offer for my topic

Thanks a lot for your info !

I would be interested in understanding more about what you were after and how it goes with dSPACE.  They were doing some very impressive things 20 years ago.  I can only imagine what they have to offer today.  If you decide to made a blog about it, please post a link here.   

Good luck.

[skYfIrE]

Hi Dr Joe,

My topic is a bit complex because acquisition systems have to use timing to synchronize signals together. For my setup, I need to synchro an angular position delivered by an encoder (not sine and cosine signals but pulses) with an analog voltage applied on an input of the acquisition system. The final is to get an Excel sheet with an angle assciated to a voltage value and for example draw a graph with the angles on X axis and voltage on Y axis. It seems maybe easy but not so that due to this synchro system.

Hi Dr Joe,

I'll be more accurate : People from NI France I've spoken with were not aware about what is a cam or crank sensor.
I'm not a NI forum user but I've seen example and white papers you've mentionned. I think so partners developers from NI know exactly how this kind of sensor work.

I've contacted dSPACE and they'll make an offer for my topic

Thanks a lot for your info !

I would be interested in understanding more about what you were after and how it goes with dSPACE.  They were doing some very impressive things 20 years ago.  I can only imagine what they have to offer today.  If you decide to made a blog about it, please post a link here.   

Good luck.

[joeqsmith]

Hi Dr Joe,

My topic is a bit complex because acquisition systems have to use timing to synchronize signals together. For my setup, I need to synchro an angular position delivered by an encoder (not sine and cosine signals but pulses) with an analog voltage applied on an input of the acquisition system. The final is to get an Excel sheet with an angle assciated to a voltage value and for example draw a graph with the angles on X axis and voltage on Y axis. It seems maybe easy but not so that due to this synchro system.

That's pretty common practice really.   For experimental work when we needed finer resolution than we could get with the cam and crank sensors, we would mount an encoder to the camshaft.   The encoder will provide an index.  Then basically do everything in the digital domain.   Maybe if you provide a little more detail, I could offer you some suggestions. 

[jwells777]

Another vote for Python.

It has the following benefits:
1. No license to purchase, maintain, or renew
2. Can do everything from simple text scripts to full-blown web driven GUI
3. Can interface with many other programs
4. Easy to find solutions to programming problems
5. Python skills transferable to other tasks besides instrument control
6. Many more people know Python than Labview
7. Python doesn't crap all over your Windows installations (although VISA still does ).

Number 1 has proven to be a really big benefit for my company, since we have several sites with common test setups. Purchasing multiple licenses, dealing with license servers, renewals, etc., has not only direct costs, but all the indirect costs as well. We already do this with Altium, and you can sink a surprising amount of time into dealing with them. And, every time you need another license, their salepeople use it as leverage against you. Don't want to repeat that with another company.

John

Out of curiosity, how do you create LABVIEW-like GUIs in Python?  We would love to ditch LABVIEW at my office but I haven't found a great substitute for creating front panels in Python or C/C++.

If anyone here has any suggestions, I would be greatly appreciate hearing them.

Thanks,

Jason

[frogg]

Out of curiosity, how do you create LABVIEW-like GUIs in Python?

MATLAB - GUIDE
Python - Tkinter

it's pretty easy to do procedural UI's with Tkinter...GUIDE isn't as much fun, but it can be done.

[joeqsmith]

Out of curiosity, how do you create LABVIEW-like GUIs in Python?

MATLAB - GUIDE
Python - Tkinter

it's pretty easy to do procedural UI's with Tkinter...GUIDE isn't as much fun, but it can be done.

Interesting.  I did a search for Tkinter and only found examples similar to what I was doing with Visual C back in the Windows 3.1 days.  Can you provide some decent examples?   

We would love to ditch LABVIEW at my office but I haven't found a great substitute for creating front panels in Python or C/C++.

It seems the GUI for you is the motivator for Labview.   Personally, I like the whole package.  The extensive libraries have really saved me a lot of time. 

 

[jwells777]

The GUI is the only motivator for us.  Most everything else is much more cumbersome for us in LABVIEW compared to other languages; however, we don't do too much interfacing with other people's lab equipment through it.  If you are interfacing lots of such equipment, I can appreciate that the existing drivers for many pieces of equipment can be a real time saver.

[joeqsmith]

Most of the time, I am using it with Ethernet, GPIB or RS232.   So talking with the various equipment hasn't been much of a problem.  Actually, the worst one I came across was when they dropped support for my Ethernet GPIB controller that they make.   It would have been simple to enough to control it if they would have released the documents for it but instead, it took me some some time with Wireshark to sort it out.

The motorcycle simulator is custom hardware, sitting on the PCI bus.  I had been doing all the low level interface in C  but have since used their driver wizard.  Basically the interrupt that runs the the control loop was all in C and now even that is in Labview.   It's a lot easier to make changes to the model.   If you look at some of my programs, I certainly take advantage of some of their advanced math functions.   

Now days it's the only tool I use to develop software for the PC.   Even if I just want some simple program to parse a file, it's in labview.   It's just faster for me to get things done with it.     

That said, a few years ago I looked at their GPU toolkit and didn't see the point of it.  In that case, I wrote some code in Labview and C.   Even my single threaded C code would run circles around my best attempt as Labview.  The CUDA code would runs circles around my best C attempts.    Still, it was a rare case where I was really just trying to do tun a benchmark.

I've never looked at their FPGA tools as again, I didn't see the point.  The motorcycle simulator is using an Altera and most of what I work on is Xilinx based.  I just use their tools and talk to the parts directly.  No big deal.   

The GUI is nice but for me, that's normally about 1% of the job.   I wonder what you are doing that Labview is more cumbersome. 

[thermistor-guy]

...
Nevertheless, when i see the many test equipment here on EEVBLOG, which is sometimes very high priced, i wonder why such environments like LabView is not seen more often.
...

Years ago ,when investigating how to automate my test gear, I decided:

* NI and Agilent s/w software were too expensive for me; spend my money on test h/w, and spend my time writing my own s/w;

* connect all test gear to ethernet; if necessary, use serial-, gpib-, and usb-to-ethernet adapters;

* use SCPI;

* use GNU tools like gawk and octave to run experiments and process the data; the tools run on a variety of setups: Linux, Cygwin, rpi.

I'm happy with the results, so far.

[JohnG]

Out of curiosity, how do you create LABVIEW-like GUIs in Python?

MATLAB - GUIDE
Python - Tkinter

it's pretty easy to do procedural UI's with Tkinter...GUIDE isn't as much fun, but it can be done.

The person who writes our GUIs uses wxPython, although there are many ways to do it. Labview may be easier for GUIs, but you can definitely do them in Python as well, from very simple to very complex. We went with Python because we felt that the combination of benefits listed was beneficial. Note that a few people in our company do use Labview, but they have a couple people who have done years of work with it. Unfortunately, they have little extra time to help others out.

It just seems to get really expensive very quickly if you have multiple sites, and to do it right, you need a lot of expertise with Labview. With Python, we also found it easy to get support and help.

YMMV, as always...

John

[joeqsmith]

Looks like they release the free versions of LabView 2020 and NXG yesterday.   

[Kosmic]

Time flies so fast. Bought the home edition of LabVIEW 2014 last year I think. Didn't had the time to play with it yet.

Ho well, I guess I will try to do my LabView rampup on LabView 2020.

[DRL]

At that price point (FREE), it might be worth a look.  From the download page:

LabVIEW Community Edition
Use the same powerful graphical programming language that engineers and scientists use for your non-commercial personal projects.
 
The LabVIEW Community edition includes:
Access to LabVIEW 2020 Community edition and LabVIEW NXG Community edition
All of the capabilities found in the LabVIEW Professional editions
The LINX toolkit for use with Raspberry Pi, BeagleBoard, and Arduino
Access to the LabVIEW NXG Web Module for creating web-based applications