USB Loopback Plugs

[Dobbs]

I'm using Burn In Test version 3.0 to test industrial PCs (motherboard contains all hardware - no floppy or CD drives).

With one PC, the USB plugs work OK (send/receive & detect OK).
On others, the USB loopback plugs either stop working after a few seconds or are not detected at all.

Also, with all PCs, sometimes when I unplug the plugs, the PC resets (no warning message or anything - just like a hardware reset).
I've tried this a few times even without running BIT, but still the PC resets.

Main concern is the PC reset.

Any ideas?

[David (PassMark)]

Hi,

None of these issues have been reported by other users before.

There are two broad possibilities
1/ That the particular plug(s) that you have are faulty, (maybe an electrical short circuit, etc).
2/ Your industrial PC's are faulty.

I don't think it is a possibility that it is a design problem with the plugs because we have shipped several thousand of them now, without this problem. Also, many of our customers are major PC manufacturers. Dell. HP, etc... They would have told us of such a serious problem a long time ago.

We test all the plugs at the end of the production line so it is unlikely (but not impossible) that the plugs are bad. I note from your previous orders that you have two plugs. Do you see the same behaviour with both plugs?

If you use the plugs on a few different desktop PC(s) do you see the same problem ?

We know that older PC's, that used the very first versions of USB PC chipset, had various problems. But these bugs have been corrected with newer motherboards.

Regards
David
PassMark

[Anonymous]

USB Loopback Plugs

While running the test software, if I unplug one of the USB loopback plugs I see the expected USB data error. But when I replug the USB loopback plug, data transmission is not re-established. This is a problem I believe with the test software not refreshing to see if there is a loopback plug re-attached while the test is running. While running CE tests on the CPU, both USB ports displayed data problems. They never re-established communication until the burnin test was completely shutdown and restarted. This is not an acceptable failure. The software must be able to reconnect to the USB loopback plugs. Please advise on how to fix this problem.

[Anonymous]

USB re-plug handled in the latest version of BurnInTest

The error handling for both a USB loopback plug (what you are using) and a USB 2.0 loopback plug (recently released) have been improved in Version 4.0 (build 1006) of BurnInTest to handle re-plugs. If you unplug/re-plug during a test, the USB loopback plug will be detected and re-initialized such that data transfer will continue. It is now available from our website at:
http://www.passmark.com/ftp/bitpro.exe (for the professional version) or
http://www.passmark.com/ftp/bitstd.exe (for the standard version).

Best regards,
Ian Robinson
PassMark Software

[Anonymous]

Issues with Passmark BurnIn Test V4.0 build 1006

I am having issues with Passmark BurnIn Test V4.0 build 1006:

I am using BurnIn Test V4.0 build 1006 with a Via EPIA Mini ITX SBC with 4 USB 2.0 ports. The operating system is Windows 2000 SP4. I am testing the USB with 4 USB 1.1 Passmark USB Loopback plugs.

If I select to test only 1 USB port (Only testing USB, testing at 50% duty), the test results in 581 KB sent and received (1 cycle), followed by complete loss of transmission. Stopping and restarting the test does not resume / restart transmission (test immediately fails). Restarting the BurnIn Test application does not correct the problem (test immediately fails). Unplugging and replugging the loop back does not resume / restart transmission (test immediately fails.) In all cases after the test initially fails the USB loop back plug is no longer detected by BurnIn Test. Restarting the operating system (Start menu --> restart) without cycling power allows BurnIn Test to detect the plug and test again for 1 cycle before failing.

Selecting 100% duty cycle and only testing 1 port works continuously until the test is stopped. Restarting the test fails with 1KB sent and received.

Testing more that 1 port immediately fails and causes the tested ports to lose detection of the attached loop back plugs.

Occasionally removing the loop back plug causes a board reset.

I have noticed no difference if 1, 2, 3, or all four plugs are attached. It also does not matter which USB port is used when testing. I have tested the plugs on a different machine (Dell system) with no problems. I am using the latest bios and USB drivers from Via for the motherboard.

Do you have any suggestions on what is causing this problem and what I might do to fix it?

[Anonymous]

If the power supply is not providing between approximately 4.6V and 5.25V to the USB bus, the USB Loopback plug will go into a reset mode (by design). It is quite possible that your power supply is not providing a constant Voltage above 4.6V to the USB bus. From a BurnInTest perspective, this would look like the device suddenly disconnecting, with transmission errors. If the Voltage remained below about 4.6V, the device would not come out of reset even on re-plug. If the voltage went above 4.6V after a reboot, the plug would work again. I would try changing the power supply.


In respect of the PC resetting issue, we have undertaken quite a lot of testing with W2000 and have managed to re-create the 'PC reset' under certain conditions. We have just released updated device drivers for the USB Loopback plug (to V2.0.0.0). This resolves the 'PC reset' problems related to power management and surprise removal (particularly relevant on Windows 2000). I would recommend downloading the new drivers from http://www.passmark.com/ftp/USBLoop.zip

You will need to uninstall the old device drivers. Information is included in the distributed readme.txt.

Best regards,
Ian