NI USB driver loses device over long period (1 week) - error -50405

[res]

Hello everyone,

It's nice for us (maybe not NI...!) to see others with the same problem - purely for the reassurance that it's not us being awkward or stupid...

It does look like others have this problem for different reasons. Our test, which has now run for the required week, has completed, and is still running, without error.

While we cannot say that our problem is definitely environmental, we are now in a position to assert that it's likely to be simply an environmental problem. We have not yet tracked down what it is that is causing the interference in our environment (NI used a mobile phone next to the cable to demonstrate the problem). The reason we are assuming this is because we have substituted the (NI) supplied USB cable with a gold-plated premium USB cable with two inductors on it. Since we have used this cable we have not had the problem recur.

If you see this problem, then I'd recommend you add a 'premium' USB cable to the (lengthy) list of things to try. But the sheer number of possible causes of this error means this is by no means a guaranteed solution.

RES.

[Tim@UBC]

Hi Andrew,

I tried using a couple other USB cables, including the one supplied by NI and had the same error. I also connected the power supply to a pure sine wave inverter that I use to run the monitor when I am in the field. The inverter is capable of supplying 120V and 150W, so this should have eliminated the potential power supply problem. With the NI usb cable, and the power running through the inverter, the system ran for about 15 minutes (an improvement over the usually two) before crashing. Before I was able to just unplug the USB cable and then plug it back in to get it working again, but with this configuration I actually had to reboot the computer to get it to run again. It also reported an Error 200020 in NI MAX.

I think I may try the premium gold plated cable suggested by Res to see if that helps at all.

Message Edited by Tim@UBC on 02-27-2009 01:47 PM

[Tim@UBC]

Hello,

So I purchased a premium USB cable with gold plated  connectors and sprial shielding in the hopes it would help. It did not seem to have any effect.

Gurshan Sidhu, the NI sales manager for UBC suggested I change the program to ignore errors from the device and contiune logging. (see attached VIs) This seems to work for all the errors except Error 50405. The next option I am considering is trying a different computer.  I am pretty limited by power requirements but I am hoping that I can get a laptop to work, as they generally only use about 35 watts. This is still three times what the current system is using.

[Marcus@Tieto]

I can confim that this problem still persists, i'm using an USB-6216 and my measrument takes 7 weeks. I'm able to reliably repoduce the error by plugging and undplugging a usb memory stick, intrestingly plugging and unplugging the same stick to an connected ubs hub do not prevoke this error.

[Tim@UBC]

I have decided that the USB controller on my field computer is likely the cause of my problems. I  borrowed a laptop and ran the labview program on it in the field with out any errors.  At this point it seems that my best option would be to buy a new computer, and since power is a limiting factor, I am looking at purchasing a laptop. I have been looking at the Acer Aspire One netbook. This laptop uses only 14.3 watts, which is only slightly more than the 10 watts the current computer uses. Before I go and buy one I wanted to check if there were any known issues with running Labview on these netbooks. It seems to meet all the system requirements but I thought I would double check first. If anyone has any experience using a similar setup, I would appreciate hearing how it worked.

Thanks. 

[Nathan]

Hi Tim,

I have tested two netbooks based on the Intel Atom processor and they work quite well.

These netbooks have performance that are roughly 10x less than a conventional laptop, and it's USB bus can only pull at approximately 1/2 to 1/3 of what a normal laptop can do. Otherwise, LabVIEW runs very well. I would say that for a good percentage of DAQ applications out there, Netbooks are a good solution. We will soon publish some content on netbook-based DAQ.

Please let me know if you have any further questions.

Cheers!

Nathan Yang | DAQ Product Marketing | National Instruments

Message Edited by Nathan on 03-30-2009 08:01 PM

[Mache]

Where did you find these gold-plated premium USB cable with two inductors ?

[JÞB]

---

Tim@UBC wrote:

I have decided that the USB controller on my field computer is likely the cause of my problems. I  borrowed a laptop and ran the labview program on it in the field with out any errors.  At this point it seems that my best option would be to buy a new computer, and since power is a limiting factor, I am looking at purchasing a laptop. I have been looking at the Acer Aspire One netbook. This laptop uses only 14.3 watts, which is only slightly more than the 10 watts the current computer uses. Before I go and buy one I wanted to check if there were any known issues with running Labview on these netbooks. It seems to meet all the system requirements but I thought I would double check first. If anyone has any experience using a similar setup, I would appreciate hearing how it worked.

Thanks. 

---

A laptop also has a battery-  this should provide immunity if the original problem was related to a generator surge/ glitch from bad fuel?

[x33]

I am seeing these types of problems to. I am running a CVI based program and am using a Digi Edgeport USB to Serial Converter (16 COM ports). My testing runs for several weeks on multiple systems. I see that after a random amount of time that I will loose one or more of the COM ports and cannot recover unless I reboot the PC. All of my testing is in a noisy test environment with cell phone, GPS and thermal chambers. I would even notice upon walking up to a system that I would here the USB reset (Window default sound) even when no CVI program is running. I suspected that it might be a noise issue but wasn't sure and was looking into higher quality USB cables. I have ferrite beads here and will try this out.

[jboti]

Hi all,

I also have this problem...

My application is written in CVI v.8, I use DAQmx v8.8, measuring six analog input channels of the USB-6008 device. After reviewing this long list of discussion, I've turned towards noise issues, so I've tried each and every thing that is enumareted here, plus an additional filtering of the signals that I measure with the USB-6008 device. Unfortunately nothing's working for me. The longest I could stay "on-line" with the USB-6008 device was about 5 hours, but the occurence of the -50405 error is completely random, and lots of times it appears under the same circumstances.

I may be thinking about changing the PC (Advantech PPC-L106T touch panel PC, with AMD GEODE LX800 CPU, no additional USB hubs), but what other PC should I choose? How can I find one that works? As you can guess, buying new PCs untill I find one that works is just not an option...

Sorry to say, but I really am dissapointed with NI products, not just because of this problem...

Boti