Losing Connection usb to NI Daq 6008 in high noise environment.

Discussion in 'The Projects Forum' started by Domophone, Oct 5, 2015.

  1. Domophone

    Thread Starter Member

    Sep 10, 2013
    42
    0
    Hi,

    I've got an intermittent problem using a National Instruments DAQ (usb-6008). The DAQ is controlled using labview. Partial Schematic is included.

    annoying.jpg

    When P0.0 (as per the schematic) is driven high, the SSR switches power to a Xenon Arc Lamp ballast. The ballast first charges, then sparks the lamp, then runs the lamp in a steady state. Upon the sparking event, seemingly randomly (about 25% of the time), the USB will lose power temporarily, and the connection will be lost, and DAQ will need re-initializing. The usb power is provided independently via the computer (no usb hub is used etc).

    If I use a scope to monitor the the PWR and GND of another USB port on the same computer, when the lamp sparks i see a spike on the USB VCC line, where it drops down to 2Volts for 40ns or so. I don't know if this is a valid way of monitoring USB power for transient conditions, but if it isn't i'm open to any advice. I'm wondering if this could be the cause, and if so, how could i go about solving it.

    The ballast has a lot of EMF issues, but I am stuck using it in my design. The noise issues appear largely as conducted noise on J7, with 30-40Vp-p spikes riding on the 120Vac line once the lamp is running. During the initial lamp spark, there is a 40-100ns enormous spike that conducts and radiates throughout the system. I have tried to reduce the conducted and radiated noise by using a Corcom 6EMC3 between J7 and the ballast, keeping it as close as possible to the ballast to reduce radiated noise. There is a ferrite bead on the power cord where it enters the equipment, to try and stifle any noise cross talk to the usb cable. There are low voltage MOVs throughout the system to stop radiated noise from causing overvoltage spikes on MOSFETS during the spark.

    I'm sure I have missed important information here, so fire away and i'll answer everything i've missed to the best of my ability. Any help solving this would be greatly appreciated.

    Thanks,
    Dom
     
  2. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    2,907
    2,165
    Grounding, shielding and isolation. I would think the root problem is the vast amount of RF energy from the wiring during the initial lamp spark. For problems like this you have to design it for EMI protection in the beginning as band-aids later are usually ineffective.

    You might try an EMI wrap like this for all lamp wiring and on the USB connection with the foil well grounded to the ballast thats well grounded and shielded. Good luck.
    [​IMG]
    [​IMG]
     
  3. Domophone

    Thread Starter Member

    Sep 10, 2013
    42
    0
    I have attempted to take measures from the start to ground and shield, though not sure how effective they've been since the problem persists. I do actually have some shielded heat shrink lying around that i was thinking about using on the High voltage lamp wires. These HV wires come from the ballast and run about 2 feet to the lamp. They are unshielded, but run inside a well grounded chassis where no other wires are run.

    I am fairly new to designing electronic circuits in high noise environments (~1.5 years). Could you give me some quick general pointers that I may have missed? Specifically, should i be isolating the USB entirely from the on board power (supplied by PS1 & PS2)? Maybe something else i can do to the USB cable to help condition the power?

    I've tried contacting the manufacturer of the ballast for extra info and the engineers who designed it have long since left the company, so they essentially churn out copies of a previous design, without knowing much of the technical details.

    Thanks
     
  4. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    2,907
    2,165
    I would try USB isolation to break the ground loop between the DAQ with a locally powered hub and computer first. Most of our designs use light-links to isolate the computer from high energy circuits.
    http://www.bb-elec.com/Products/USB-Connectivity/USB-Isolators/USB-Isolators.aspx

    I tend to use overkill on even simple devices that I know will be near machines capable of generating large transients. For even a simple RS-232 touch converter full shielding and filtering of I/O and power is typically used.
    [​IMG]
    [​IMG]
     
  5. Domophone

    Thread Starter Member

    Sep 10, 2013
    42
    0
Loading...