Wasted day at the track thanks to LV 2011, One more bug...

[JoeQ]

For starts, I would have expected that they would have had us in contact with a person/s knowledgable about the design of the board. Knowing there was a deisgn problem and we needed a fix, I would have expected them to work with us under an NDA and provide us full details about their design so we could address it on our end. I would have expected them to offer to return all existing stock.

While I am sure they believe the customer needs to verify each of their designs. The way I see it, fool me once shame on you. Fool me twice....

In the end, it would be VERY hard to sell us on Labview hardware and support.

[ShaunR]

Well. Here's my experience.........

I was working on an FPGA and we wanted to transfer huge amounts of data from a 3rd party FPGA aquisition board, accross the PXI backplane, to an NI board for crunching. We couldn't use the NI streaming VIs since the technology is proprietary and NI wouldn't liaise with the 3rd party so they could implement it in their FPGA (which is fair enough). However. NI said that they could DMA at about 700MB/Sec in each direction (1.5 GB/sec) across the back-plane which was "good enough for our team".

The only problem was that all examples never addressed this sort of throughput apart from mentioning that, under the right conditions, it was possible.

So long story short. The local NI rep hooked me up with the UK FPGA guru. I sent through an example of what we wanted to do (with which I was getting about 70MB/Sec) and he sent through a modified version with comments about where and what was important in my example for getting the throughput. It could do it at 735MB/Sec (each direction). He also sent me through an internal (not for distribution) benchmark document of all the NI PXI controllers. what their capabilities where, what measured throughput's could be obtained, with what back-planes and which board positions within the rack (which is important).

Saying all that, It did take me two weeks to get through to him. I had to go through the "correct channels" first before the NI rep had a good excuse to "escalate" the issue through the system. The key is really building up a contacts list of direct dial numbers to the right people. If you know what you are talking about, they will be happy to take your call as they know it's not a silly problem. NIs problem is that there are too many inexperience people calling support for trivial things and, unfortunately for us, their system has been setup so that the engineers are well buffered from this.

Edited October 2, 2012 by ShaunR

[hooovahh]

Well. Here's my experience.........

...

Saying all that, It did take me two weeks to get through to him. I had to go through the "correct channels" first before the NI rep had a good excuse to "escalate" the issue through the system. The key is really building up a contacts list of direct dial numbers to the right people. If you know what you are talking about, they will be happy to take your call as they know it's not a silly problem. NIs problem is that there are too many inexperience people calling support for trivial things and, unfortunately for us, their system has been setup so that the engineers are well buffered from this.

I agree. NI has some of the best product support I've seen. But getting to talk to the right people, or convincing the right people you know what you're doing, can make the support you get very different.

I had a memory leak issue when using user events on real time. The issue turned out to be my fault, but I called NI looking for assistance before I knew that. After I explained what I was doing to the first guy I talked to, he told me user events were not allowed on real time. It was at that point I knew I was not talking to the right person.

One more (not from me but a co-worker). We were ordering some parts from NI, I forget the actual part but it was a common enough part that we thought lead time would be a week or so. When we ordered them they said it would be 8 weeks. We called NI to see if there was anything that could be done to speed it up. After talking to the right person at NI we found out that there was some mounting bracket that was on back order from another supplier in another country which wouldn't be available for 8 weeks. We told NI that we didn't need the mounting brackets and would take them off anyway, so they took an part from the assembly line without the mounting hardware and sent us one. Note that I was told this story and parts of I may have remembered incorrectly, and your results may vary when needing help to get your hardware sooner.

[JoeQ]

I would guess that deal created over a half million in sales for NI. It was more a trial than anything but for that amount, I don't expect to have to beg for real support. We had an FAE at the time and he was given every opportunity to help us locate someone. If this would have been the case of a one off prototype, my expectations would have been different. All water under the bridge now, we learned our lesson the hard way.

[ShaunR]

I would guess that deal created over a half million in sales for NI. It was more a trial than anything but for that amount, I don't expect to have to beg for real support. We had an FAE at the time and he was given every opportunity to help us locate someone. If this would have been the case of a one off prototype, my expectations would have been different. All water under the bridge now, we learned our lesson the hard way.

In my case it was one FPGA card (~$4K). However. That reminds me of the other trick. Ask for a loaner to test before you buy if you have never used the product before (they've always got a couple for demos and conferences kicking around). It concentrates a sales reps mind They tend to be focused on "potential" sales rather than "previous" sales, but they can pull the strings.

Edited October 3, 2012 by ShaunR

[JoeQ]

The Tripp-Lite adapter appears to use it's own chipset. It has a VID 2478, PID 2008, REV 0300. This adapter appears to work with all flavors of Labview and Windows.

Strange enough, using C# with nothing more than

com1.Open();

com1.Close();

prior to running Labview fixes the problem with all the OS and adapter combos. Labview can be closed and reopened and will continue to work correctly until the adapter is pulled and pulled back in or the PC is power cycled.

NI has ".... exhausted all of our options. " I now doubt very much that they will fix it and I have asked that they go ahead and close out the case. Where I work, we have purchase several known working adapters to get around the problem for now. Had I known the PCI bus and serial ports were going to be problems, I think we would have stayed on 2009. The only push to "upgrade" to 2011 was NI dropping support for 2009.

For my home projects, I have gone back to LV6. The speed of the UI is just too fast for me to ignore and it does everything I need it to.

Till the next great bug hunt.....

[Phillip Brooks]

I agree. NI has some of the best product support I've seen. But getting to talk to the right people, or convincing the right people you know what you're doing, can make the support you get very different.

I had a memory leak issue when using user events on real time. The issue turned out to be my fault, but I called NI looking for assistance before I knew that. After I explained what I was doing to the first guy I talked to, he told me user events were not allowed on real time. It was at that point I knew I was not talking to the right person.

You didn't use the correct secret code word...

Edited October 11, 2012 by Phillip Brooks

[Jordan Kuehn]

You didn't use the correct secret code word...

Reminded me of this:

http://www.youtube.com/watch?v=lhwLj2qIT4k

[JoeQ]

https://www.youtube.com/watch?v=6PpAJeAurr4&feature=youtu.be

 

Video showing a home made simulator I am working on.  

[JoeQ]

Playing with an open source C code generator to build the model for my simulator.