Intermittent error 28 with GPIB-USB-B

[trnoe]

I'm using a GPIB-USB-B interface to control my test rack.  The computer is a laptop running XP.  When running my application code, I intermittently get an error 28.  This is 'EPWR", and the error message says that the GPIB-USB-B has lost power.  From the NI 488.2 documentation, "EPWR results when an interface loses power. This often results when the system goes to and returns from a standby state".

Standby and suspend are turned off on the laptop, it never powers down.  I'm at a loss to figure out the source of my error.  Since my application takes several hours to run, this effectively prevents me from running my code.

Here are some experiments I've tried so far:

 - Connect the GPIB-USB-B directly to the PC, or connect it through a powered hub (rated at 500 mA per port).  Results: No difference, error still occurs.

 - Disable all the ACPI functionality that I can in the laptop.  Results: No difference.

 - In the PC's control panel, under System/Hardware/Device Manager, 'USB root hub', when opened, has a 'Power Management' tab.  On this tab, there is a checkbox for 'Allow the computer to turn off this device to save power'.  I have turned this off.  Result: No difference.

 - Monitor USB +5V with a digital oscilloscope, and see if the +5V supply ever goes away.  Result:  The power supply is stable, and doesn't trigger off any transients when the error occurs.

 - Disconnect all GPIB instruments except one, in case there is some problem with cabling or bus loading.  I have tried this several times, using different instruments.  Result: No difference.

 - Try a different PC.  I have tried this some, and have not seen the error, but I don't have enough test time to be certain.

At any rate, monitoring the +5V USB supply with the scope seems to remove any possibility that USB power is really being interrupted, which suggests that the true problem is not what the error message indicates.

If anyone has any ideas, or has seen other problems masquerade as error 28, I would be greatly interested in hearing about it.

Thanks,

Terry N.

trnoe@yahoo.com

[ScottieB]

Terry,

I think you may have e-mailed into support a couple times with this issue, haven't you?  If so, I think I remember being asked these sorts of questions.  And, unfortunately, you've taken the suggestions right out of my mouth.  The powered hub would've been my first, and monitoring the 5V pin on the USB (you're measuring between the hub and the GPIB-USB-B, right?  And not just a different hub port?) seems to eliminate power.

So, couple of questions for you:

1) Can you describe in detail how you're measuring the 5V pin so that I can completely take that out of the picture?
2) Have you tried multiple GPIB-USB-B's (if you have more than one)?  Perhaps we're looking at a faulty power line inside your GPIB-USB-B.
3) What's your driver version?
4) What is the assembly number (ASSY) of your GPIB-USB-B?  For example, 188417N-01.  Should be printed on the case.

Thanks,
Scott B.
GPIB Software

[trnoe]

Scott,

Thanks for your interest in my problem.  But no, I haven't contacted you - or anyone - in the past about this problem.  In answer to your other questions:

1) Can you describe in detail how you're measuring the 5V pin so that I can completely take that out of the picture?

 - The product that I am testing is a USB bus-powered device.  I measured the USB +5V where it enters the product.  It was connected to the same USB hub the GPIB-USB-B was, but of course to a different port.  The supply was measured with a digital scope via a 10:1 probe.  I turned off the scope's auto-triggering, and set it to trigger on a negative edge, at a level 0f 4.5V.

I should mention that I did occasionally see short-duration transients on the +5V, but they were uncorrelated with the occurrence of error 28.

It would be reasonable to suspect that my device-under-test, being connected to USB, might be the source of the problem.  However, error 28 still occurs even if the device-under-test is completely disconnected from the test system.

I have been wondering whether monitoring the +5V off of a different hub port was the right approach.  So, I have since started monitoring the +5V inside the GPIB-USB-B.  I removed the plastic shell, attached wires to the USB +5V and ground, and connected them to the oscilloscope.  Since the scope probe wouldn't reach, I connected the scope's 1 Mohm input to the supply vis alligator clips, through 3m of BNC cable.  I see essentially the same thing as before - occasional transients on the supply, but they are uncorrelated with the occurrence of the error.

2) Have you tried multiple GPIB-USB-B's (if you have more than one)?  Perhaps we're looking at a faulty power line inside your GPIB-USB-B.

Unfortunately, I only have one GPIB-USB-B, and am unable to perform this test.

3) What's your driver version?

I was originally using driver version 2.3.  In an effort to fix the problem, I upgraded to the newest version, 2.4.  This had no effect on the problem.

4) What is the assembly number (ASSY) of your GPIB-USB-B?  For example, 188417N-01.  Should be printed on the case.

On the case I find the following information:

Part number: 188417D-01

Serial Number: 10CE017

I'm interested in any advice or suggestions you might have.  From your post, it sounds as if perhaps this problem has been seen before.  I think I might try placing a large capacitor across the power supply terminals inside the GPIB-USB-B and see what happens.  Since it's over a year old, I would imagine that the warranty is expired anyway, so I have little to lose by experimenting.

Regards,

Terry

trnoe@yahoo.com

[ScottieB]

For anyone following this thread, I'm working with Terry to get his GPIB-USB-B adapter sent back to National Instruments for an RMA that should fix this problem.  After this process is complete, hopefully he'll post and let us know that everything worked out okay.

Scott B.

GPIB Software
National Instruments

[trnoe]

Well, I was finally able to solve my problem with my GPIB-USB-B.  I had tried everything I could think of, including the loaner unit I got from Scott B. at NI, and nothing worked.

For unrelated reasons, I recently purchased a PCMCIA-USB adapter.  Since the computer I use for GPIB control is also a laptop, I decided to try using the card to connect to the GPIB controller.  I connected the GPIB-USB-B to the PCMCIA card, and after extensive testing, I'm certain that the problem is gone.

I don't have any idea why this works, but I'm certainly happy that it does.  The PCMCIA-USB adapter I'm using is an IOGear GPU202.

Thanks to Scott for all the support he provided, especially letting me borrow the loaner unit.

Terry N.

trnoe@yahoo.com

[ScottieB]

Terry,

I'm glad that you finally got it working.  I never would've thought of this solution but it's intriguing.  It makes me want to blame the laptop for not providing enough power to the USB port...inside the USB protocol we request the maximum amount of power from the port (500mA).  Typical usage results in something quite a bit lower than that (somewhere around half), and even under heavy usage we will never come close to 500mA.  I can't imagine a laptop failing to provide 500mA to its USB port--it's required by the USB spec!  But that's what it seems like...the onboard USB ports aren't giving enough power, but your PCMCIA-USB adapter is.  And way back when we tried the powered USB hub apparently it didn't provide sufficient power either?  Very odd.  I still feel that my power theory may not explain all this.

I wonder if you could post your computer brand/model and specifications as well.  Maybe someday someone will stumble across this thread and have the same problem with that configuration.  Also, what is the brand/model of that powered USB hub you used to test it out??  That thing should've provided enough power for the device too.

Thanks for your followup and patience.

Scott B.

GPIB Software

[trnoe]

Scott,

The computer I'm using is a Toshiba Satellite A65.  I understand why the power theory makes sense to you, but I've done enough experimentation to pretty much rule that out.  Here are some of the reasons that I don't think it's the laptop's USB port's power capacity:

1) The problem appears when the GPIB-USB-B is directly connected to the PC, or through a hub.  I tried several different hubs, all rated at 500 mA; the results were the same.

2) Using a different PC, I never saw the problem.  This was true both when the 'good' PC was directly connected to the GPIB-USB-B, and when it was connected through the same hubs mentioned above.

3) I went so far as to monitor the +5V supply to the GPIB-USB-B with a storage scope to see if the voltage level was drooping under load, or if there were transient brownouts.  I saw no problems with the +5V into the GPIB-USB-B.

I know you would rather not hear this, but I feel pretty certain that the error messages from the GPIB-USB-B were incorrect - power was fine.  As far as what caused it, I don't know; I would assume it was some kind of software interaction between the PC's code and the driver.  Thinking that perhaps I had a .dll compatibility problem, I had uninstalled pretty much all the other software on the PC, and had re-installed the NI driver (several times).  So I don't think the problem was related other software I'd installed on the machine, but a more fundamental compatibility issue between the NI driver and, perhaps, low-level USB drivers on the PC.

As I had mentioned to you in a previous email, I had been contacted by one gentleman who was having similar 'error 28' problems who had seen my original post.  He felt that his problem was caused by EMI; but given the intermittent nature of the problem and what he described as the fragility of his solution, I wouldn't be surprised if his problem and my problem are related.

Regards,

Terry

[uyya]

Hi

    i have developed an USB GPIB Converter application which works fine on NI USB HS card.But with the same setup , its not working properly with NI PCI GPIB card. Data is being lost in PCI GPIB card.Data loss is more occuring frequently  for commnad transfers than data transfers.

                 i am Using TNT4882  as GPIB ontroller and cy7c68014a as USB Controller.

Can any one explain the differnce between USB HS card & PCI GPIB card with respect to this problem?

Thnx in advance

R.Sreedhar

[ScottieB]

Hello,

Please start a new discussion thread for new issues.  Your issue does not appear to relate to the thread in which you posted.  This keeps everything orgnized for users searching the forums.

While there are differences in the way PCI-GPIB and GPIB-USB-HS devices operate inside, there should be no difference in lost characters or software functionality.  You should contact the manufacturer of your instrument to see if they make any assumptions about what GPIB hardware you are using this with, or if they happen to have tested with USB-HS.  There are no good reasons explaining something working on PCI and not on the USB.  It is possible that the instrument doesn't fully conform to IEEE 488.1, which might explain this behavior.

Scott B.
GPIB SW

Message Edited by ScottieB on 04-28-2006 09:52 AM

[Martin_Kroner]

I don't know if this thread is already dead, but I have the very same problem. The annoying thing is that the error only occurs after some time. So, my software sometimes runs for more than one day before the error 28 occurs. Then however, I have to shut down the whole LabView in order to reset the GPIB-USB-B interface. I also tried different USB connectors on my computer and USB hubs with separate power supply. Is there an other solution to this problem then to by a PCI-GBIB card?   Best wishes, Martin