Shop OBEX P1 Docs P2 Docs Learn Events
High RF Noise enviroment resets Propeller — Parallax Forums

High RF Noise enviroment resets Propeller

PhilldapillPhilldapill Posts: 1,283
edited 2008-12-21 17:25 in Propeller 1
I'm working on a project that deals with large currents(~50A). Sometimes, while testing, I disconnect the power from a low-inductance load, and get a bright spark. This wouldn't the problem itself, except that the propeller resets nearly everytime this happens, or, my adc freezes up until I reset the whole system.

The ADC aside, any idea how to shield the Propeller so it doesn't reset? In addition, any idea what is causing this? I have fairly long traces to all the I/O pins that are floating. Could these be, collectively, picking up alot of noise and disrupting the inner workings of the propeller, causing the reset?

Comments

  • sosaraujososaraujo Posts: 24
    edited 2008-12-21 04:20
    ...maybe if you first build a faraday cage....

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Adrian Sosa Araujo.- San Luis, ARG - Time Zone (GMT -3)
    El planteo de un problema es mas importante que su soluci
  • PhilldapillPhilldapill Posts: 1,283
    edited 2008-12-21 04:34
    Yes, this will be in a plastic box, but I could always spray the inside with conductive paint, eh? I've heard of this method being done before...

    The question is, what is causing the reset? I'm assuming the noise, but why?
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2008-12-21 04:37
    Are you saying that there is no electrical connection whatsoever? Not even a common ground?
  • sosaraujososaraujo Posts: 24
    edited 2008-12-21 04:56
    I think the "spark" is the problem.
    You should read this link.
    en.wikipedia.org/wiki/Faraday_cage
    Dont forget to connect (the cage) to earth ground to dissipate any electric currents generated from the external electromagnetic field (the spark).

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Adrian Sosa Araujo.- San Luis, ARG - Time Zone (GMT -3)
    El planteo de un problema es mas importante que su soluci
  • PhilldapillPhilldapill Posts: 1,283
    edited 2008-12-21 05:03
    Erik, that is what I'm saying. There is NO electrical connection from the spark circuit, to the propeller. I am using an ISO721 high speed isolator that has 4KV of protection. I gave up on my "mystery mosfet driver failure" problem in a prior post, and now I'm working on WHY the chip resets. I'm an EE and have a good understanding of farady cages and all that stuff... I'm just mystified that there is actually enough energy in that spark to put off that kind of RF noise to disrupt a totally isolated propeller. If this is in fact a case of high-energy sparks putting off alot of EM noise, then I guess the conductive paint/farady cage could be a fine and dandy solution, however, I'm suspicious that this really is what's going on.
  • Mike GreenMike Green Posts: 23,101
    edited 2008-12-21 05:34
    Anytime you have wires or PCB traces that are more than an inch or so, you have a perfectly good antenna. If you've got enough voltage and current to make a spark, you have enough to generate a pretty potent electromagnetic field. It wouldn't take much to couple to a low voltage, possibly high impedance circuit (the Propeller) nearby. You need some shielding between the power circuitry and the Propeller and some bypassing, maybe some ferrite chokes on the power supply for the Propeller, the reset circuitry, and maybe some of the other leads as well.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2008-12-21 05:35
    Pull your reset line high through a 1K resistor and maybe even add a cap to ground. The reset line is fairly sensitive and it only needs a small spike to trip it.

    For a quick elimination test you can just tie the reset to VDD to make absolutely sure it is or it isn't this.

    Other things to check are the oscillator connections, they should be very short and hopefully have a ground plane. An elimination test to to completely bypass the oscillator and just run the circuit using the 12MHz internal and if it still resets then you know it isn't the oscillator.

    *Peter*
  • PhilldapillPhilldapill Posts: 1,283
    edited 2008-12-21 05:46
    Mike, do I still need all this power supply bypassing when the propeller circuit and power(spark) circuit are totally isolated? For the most part, we can say that these have NO electrical connections, outside of the isolator chip. They share no common ground or any wire, other than internally in the ISO721 isolator.
  • PhilldapillPhilldapill Posts: 1,283
    edited 2008-12-21 05:47
    Thanks, Peter. I'll try those suggestions.
  • ChuckTChuckT Posts: 12
    edited 2008-12-21 06:19
    I had similar problems, adding cap on prop chip and the reset pin, snubbers on motors. But the killer was eliminating a ground loop.
    good luck. I'm an EE too but EMI is magic.
  • LeonLeon Posts: 7,620
    edited 2008-12-21 08:07
    Judicious use of TVS devices can get round problems like that. I like the AVX TransGuards.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2008-12-21 13:43
    I agree with Leon. I have a hunch that having some sort of "surge protection" would help solve some of this. I also assume that you have incorporated a ground plane into your pcb.
  • Mike GreenMike Green Posts: 23,101
    edited 2008-12-21 14:20
    Philldapill,
    People have communicated around the world with 10mW and a good antenna. At a range of a few inches and with the amount of current you're discussing, all you need is a couple of inches of PCB trace or loose wire running in the right physical relationship to each other and a sensitive (high impedance) input and, bingo!, you've got communication.
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-21 14:54
    I'm not a EE, nor do I know much about EMI, but I can vouch for Mike Green's comments about tiny "antennae" messing up the Propeller. When I hook up everything using solderless protoboards, I am always on the lookout for exactly that kind of interference. In the past, when I've had long (6" to 12") lines wigging out across my protoboards, I have often experienced one subsystem interfering with another, or with stray radio signals, etc. giving me bogus pulse counts when none were present in my actual system. Just for grins, take various pieces of wire and hook them up to an oscilloscope and watch the radio traffic pass through your workshop/body. At a wide variety of frequencies: it's a wild drunken orgy of EMI. Use that same Oscope setup to test the insanity of signals generated by your nasty spark gap. You'll be shocked at the injustice and immorality of it all.

    To dispell those demons, you must gird them in iron and touch them to earth.
  • Timothy D. SwieterTimothy D. Swieter Posts: 1,613
    edited 2008-12-21 16:00
    What kinds of bypass caps are you using now Phill?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Timothy D. Swieter, E.I.
    www.brilldea.com - Prop Blade, LED Painter, RGB LEDs, uOLED-IOC, eProto for SunSPOT, BitScope
    www.tdswieter.com
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2008-12-21 16:58
    Never underestimate EMI, the first radio transmiter was a spark gap and the reciever was a loop of wire.
    Physics, not just a good idea, it's the law.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent, only $1.
    Send cash and signature to CannibalRobotics.
  • LeonLeon Posts: 7,620
    edited 2008-12-21 17:25
    It's the reason why we have such stringent certification requirements in the EU. People grumble about the cost of it, but it does mean that equipment doesn't usually fail because of poor design.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
Sign In or Register to comment.