USB to RS232 Interface bug

Over the last several years, my computers were plagued by program stoppages. "Computer found a problem and shutdown..."

The computers use a Parallax USB to RS232 Rev B interface to BS2 micro controllers.

A NASA engineer familiar with USB interfaces analyzed this problem and determined that when the computer reads the driver code from the USB to RS232 chip and loads the driver into the operating system a condition is setup where the computer will detect an interface problem as data is sent to the computer for display using the DEBUG instruction. The computer shutdown usually takes several weeks to occur but it does occur. With the USB to RS232 interface removed the computer runs the program and never fails. Tests have run for over eight months with no computer failure.

Discovery

Comments

  • 12 Comments sorted by Date Added Votes
  • Discovery wrote: »
    Over the last several years, my computers were plagued by program stoppages. "Computer found a problem and shutdown..."

    The computers use a Parallax USB to RS232 Rev B interface to BS2 micro controllers.

    A NASA engineer familiar with USB interfaces analyzed this problem and determined that when the computer reads the driver code from the USB to RS232 chip and loads the driver into the operating system a condition is setup where the computer will detect an interface problem as data is sent to the computer for display using the DEBUG instruction. The computer shutdown usually takes several weeks to occur but it does occur. With the USB to RS232 interface removed the computer runs the program and never fails. Tests have run for over eight months with no computer failure.

    Discovery

    Hello!
    Okay who. Can you cite specific references here? I freely admit I've seen something of a sort happen, but on a computer running Windows XP_SP1 with specific updates applied, and not on Windows 7_SP1 also with matching updates.

    I should also explain that the way Windows does things is positively fraught with implied dangerous activities.

    And you should also tell us what sort of system you yourself run.
  • The current computer that I am using is an EeePc , Intel processor, running Windows 7 Starter. The operating system was configured to run without any automatic functions: no Internet, no updates...no nothing but a basic computer that runs a Basic program for the BS2 micro controllers and a C++ program for the Propeller. The Propeller PropPlug USB was modified by inserting a slide switch on the interrupt signal from the Propeller so that the switch is closed when downloading a modified program to the Propeller then switched to open otherwise. I now run the computer with neither USB connections. The computer does not detect any problem and continues to run the operating system program perfectly well for many months.
    When the USB to RS232 interface is connected to the BS2 micro controller data is sent from the BS2 to the Basic program for display on the screen. After several weeks, the computer encounters a problem and stops.
    So, to avoid the computer stop problem, I only inset the USB interfaces when I make program changes. I can not send data from the Propeller micro controller or the BS2 micro controller 24 hours a day, 7 days a week, month after month without having the computer encounter a problem and stop.

    Sincerely,

    Discovery
  • jmgjmg Posts: 12,304
    Discovery wrote: »
    ... After several weeks, the computer encounters a problem and stops.
    So, to avoid the computer stop problem, I only inset the USB interfaces when I make program changes. I can not send data from the Propeller micro controller or the BS2 micro controller 24 hours a day, 7 days a week, month after month without having the computer encounter a problem and stop.
    If you change the USB bridge to another brand, does the problem go away ? Plenty of choices out there... common are parts like CP2102N/CH340/PL2303HX, and also Microchip/Cypress/Exar have offerings too..

  • jmg,

    Thank you...I will investigate.

    Sincerely,

    Discovery
  • I have a similar problem with he same adapter. I am using a Professional Development Board and an RS-232 adapter(Rev. B) on windows 10. My programs don't upload. When the code attempts to download, there is a loopback but no echo. I loaded the following to a Stamp on my BOE, then transferred it to the PDB:
    ' {$STAMP BS2}
    ' {$PBASIC 2.5}

    DEBUG "hello"
    END
    What appeared in the debug terminal was:
    ($llohello(e(lkHeldo `l,o
    I think the problem may be the RS-232 adapter, but I have been told that it is the latency timer.
  • Revision B adapter, don't know why it did the face.
  • Sometimes, when a device is sending RS232/USB data when first connected, Windows will think it's a mouse. (Think of the havoc that could wreak!) If you go to the control panel and disable the "mouse" on that port, the problem should go away.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • Thank you, I had not thought of that before.
  • I ran a test of the Eee laptop computer that was part of the USB to BS2 interface. The computer had a regular thumb drive installed and the machine was set to operate unattended to see if the operating system would crash like it did with the USB to BS2 interface. After about two months Microsoft Windows detected a problem just as before and cleared the screen.

    With this information, I feel it safe to say that the NASA engineer's analysis of the Parallax interface was wrong. Microsoft crashed without the USB to BS2 interface installed. I give the Parallax interface a clean bill-of-health.

    Discovery
  • And there you go.

    I can't tell you how many times I've seen this computer throw up a BSOD screen on something else, and then restart and work like nothing happened.

    But on the other machine when running Windows XP_SP1 that machine would do much the same thing and cause more problems.

    USB detection and management is still an art form for Windows.
  • Six years ago, I switched from Windows PCs to Android tablets for machine control. The link between the tablet and the motion controller is Bluetooth.
    I braced myself for all kinds of possible glitches......hasn't happened.

    I had an issue, in a couple of installations where I was setting a flag over Bluetooth and almost immediately resetting it on the motion controller. Sometimes it would work and sometimes not, it was a timing thing.

    The customer's in-house "experts" insisted that it was a Bluetooth issue and were pushing me to switch to USB instead. Since I corrected the flag issue, my installations have been set-and-forget.

    However, the very suggestion of wireless can be a deal-breaker so now I intend to provide both BT-RS422 and USB-RS422 on every system so they have a choice.

    Android tablets make ideal HMIs. They are ubiquitous, inexpensive, robust, feature-laden, powerful and easy to code for. Lock them down to Kiosk Mode and they can't be tampered with. I restrict mine to my own app and TeamViewer Quick Support for remote control.
    Failure is not an option...it's bundled with the software.
  • I agree. They do indeed. Except when used to run a space's environment.






    ----
    And this message is being sponsored by the Klingon Demolition Corp, look for us at your favorite watering hole.
Sign In or Register to comment.