GPIB-usb does not work. Endless NI Misery

[vincent1]

Every time i have to install the Ni488.2 software there is nothing but trouble to get it working.

Windows device manager finds the driver and finds the GPIB-USB-HS fine. yet Nationals provided MAX never finds the hardware. And then the endless reboot ,uninstall reinstall reboot cycle begins until it by wonder all of a sudden works.

Now i have yet another installation that i cannot get to recognise the gpib-usb at all.

Windows device manager correctly lists the GPIb-USB as a device with NI logo and reports it as working properly.

Control through GPIB0 device works correctly ( open a dos promt and do a copy con GPIB0 with some commands in it).

When i launch MAX ( Measurement Automation Ecplorer) this thing stubbornly keeps on claiming that there is no hardware to be found. Also acces using GPIB32.dll does not work.

Can someone at NI for once please make something that works ?

[vincent1]

And here is another pet peeve with this thing : it is NOT USB compliant. If you plug it in non-powered USB hubs (hubs that do not have their own power supply but use power from the USB line) it will often not work correctly because it cannot request the correct amount of current it requires.

On enumeration the device should communicate correcly what power class it belongs to so that the operating system can flag that this should not be connected to unpowered hubs.

[Dennis_Knutson]

Strange. I installed the latest drivers just last week onto my laptop and had zero problems, One of my clients has dozens working with no issues either. What type of computer and OS? What version of NI-488.2?

[vincent1]

Ok, i spent a considerable amount of time getting to the root cause of this problem and these are the 2 cases i have found and their workarounds.

The symptoms are always the same. Plug in the USB-GPIB , windows detects new hardware correctly installs driver. The driver shows up in system manger and is flagged as 'working properly'. Ni MAX cannot find any interface. Adding GPIB interface fails. Unplug / replug / reboot reinstall ; nothing changes. Windows keeps telling the device is there, the driver is loaded and it is working properly. As soon as you launch NI MAX or add hardware or try actually talking to the device : it's not there.

- Plugged in to a bus powered hub. If you have a USB hub between the computer and the USB-GPIB 'dongle' this restricts downstream current to 100mA max. the NI USB-GPIB dongle apparently does not correctly negotiate USB power requirements. It will enumerate correctly and windows will find it ( shows up correctly in windows device manager) . However, as soon as NI software wants to talk to the GPIB-USB this changes power mode and will disrupt communication since the hubs cannot deliver the operating current. The additional cable from comuter to hub may drop Vusb too low for correct operation of the NI-GPIB dongle.

Solution : ALWAYS use POWERED USB hubs. These can supply the full 500mA per port and will maintain Vusb at 5 volt ( although per USB spec a device should be able to run as low as 4.75 volt, the NI GPIB dongle has trouble with this).

- plugged in to a computer motherboard that does active power control on the usb ports ( motherboards that actually measure the current flowing into the usb port and will slam the brakes if this exceeds the required power setting as negotiated during enumeration. )  <- this was my root problem. The motherboards used in the lab all have active power control and they all have trouble with the USB-GPIB.

Solution : put a powered usb hub inbetween the computer and the usb-gpib dongle. The donlge can now suck all the power it wants.

[JasonSmith]

The GPIB-USB-HS request 500 mA of current in its USB descriptors, which should prevent it from being loaded in a port which can only supply 100mA. I have a couple questions for you that will help me understand what you are seeing on your hardware:

1. When plugged into the non-powered hub, can you view the properties of the hub in device manager? You should be able to see there the requested power for each device attached to the USB hub. If it shows 500 mA for the GPIB-USB-HS (as it does on my machine), then it may be that the hub is falsely claiming that it can supply 500mA on each port, even when the power cord is not connected.

2. With the motherboards that do not work with the GPIB-USB-HS direct connected in your lab, can you provide me with the make and model of the motherboards and/or computers you are using?

3. When you are plugging in the GPIB-USB-HS for this testing, does it have a cable or instruments connected, or is it the donle by itself? If a cable and/or instruments are connected, how many meters of cable and how many instruments are connected?

Thanks,

Jason S.

National Instruments

[vincent1]

Well the funny thing is that windows does load the driver and enumerates the gpib-usb even if the port cannot supply 100mA. I guess the usb-gpib is essentially dormant unless someone opens the devices and starts talking to it. ( hence the device showing up in windows devices explorer. ) I guess that is a flaw in windows then ( XP sp3 ). the OS apparently does not check that the chain (usb controller, root hub , hub, endpoint) is capable of providing the current requested by the newly attached device.

The hubs are standard off the shelf usb hubs from various brands (no-name brand ,ching-chang-chong tech and the likes. Nobody makes bulletproof usb hubs anymore. I used to have some that were made using Texas Instruments USB hub controller as and TI usb power switches. they were 7 port and came with a 4 ampere power adapter so each port could source half an amp without problems. But those seem to have gone the way of the wind . They were made by d-link i believe.)

As for the motherboards : We have a range of motherboards , all made by intel ( original motherboards sold by intel ) . Especially the motherboards that have the ich4 and ich10 exhibit this problem as these motherboards actually sense the current draw and slam the port if it exceeds. these motherboards also have problems with usb powered harddisks during spinup. If a usb devices steps out of line the motherboard will terminate power to the port.

I have seen the same problem with IBM Thinkpads T40's i believe as well as some newer HP business laptops. If you run them off battery they close the port.

Anyway , i now supply my engineers with a standard powered hub and the problems are gone.