simulation of the behaviour of a PCB Spark Gap

Discussion in 'General Electronics Chat' started by marc1996, Aug 1, 2018.

  1. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    Hello,

    I'm doing an University project fabout PCB Spark gaps.

    My question is:
    Someone know if there is any software to simulate the Electrostatic discharges?
    I tried Quickfield, a FEM software that allows you to simulate a transient electric in the time domain but this is not still what I'm searching for.

    I attach here a youtube video in case someone wants to know more about PCB Spark Gaps:



    Thanks.
     
    Last edited: Nov 6, 2018
  2. Bordodynov

    Well-Known Member

    May 20, 2015
    1,979
    597
    See
    2018-08-01_10-26-00.png
     
  3. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    This allows you to simulate with different spark gap shapes? Could you explain how it works? because I'm a bit lost when looking to this photo.

    I can see that you placed some different ISO tests or atleast looks like that. The one that I'm more interested in is the ISO 10605 ( for automotive ).

    Thank you!
     
  4. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    .
     
    Last edited: Nov 6, 2018
  5. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,836
    2,071
    The moisture content of the air would be an important factor. In parallel with that air gap will be the pcb surface itself, possibly with contaminants on it. I don't know how you would get realistic calculations of breakdown voltage. Perhaps you would need to run some experiments?
     
  6. Bordodynov

    Well-Known Member

    May 20, 2015
    1,979
    597
    I made the models of dischargers by datasheet. I realized that you want the program to calculate the breakdown by geometric parameters. In my models this is not so. About the standards. Any source can be reproduced. you only need to know the parameters of the pulses. You simplify the process of discharge. The discharge has an inertia. If you quickly apply voltage to the arrester, it will break through at a higher voltage than the voltage you calculated for the static mode. That is, the breakdown voltage of the arrester depends on the rate of increase in the input voltage, which reflects my models.
     
  7. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    I know that the humidity and pcb dirt are very importants factors to consider but, there is a formula that gives you an idea about where this value will be approximately: Vbd=3000*P*d+1350, being P pressure in atmospheres ( we can consider 1) and d the distance in milimetters.

    I'm just trying to get a simulation of this, maybe the results won't be that realistics but it's just to have an idea of how the protector will react against the ESD.

    After that, of course I will do experimental simulations in order to get more empiric results but, I want to do computer simulations if its possible before going to the experimental area.
     
  8. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    .
     
    Last edited: Nov 6, 2018
  9. Bordodynov

    Well-Known Member

    May 20, 2015
    1,979
    597
    As already mentioned above, self-made arresters will not give reproducible work. The breakdown is affected by the air composition, humidity and surface cleanliness of the board. They can be placed in non-critical cases.
     
  10. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    So you think that there are too much factors that can vary the results for the simulation process, right? maybe is better to focus on the experimental section and try different spark gaps shapes with different distances in order to achieve the results that I'm looking for...
     
  11. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,836
    2,071
    You can certainly simulate it, but the results would be far removed from reality; so is it worth the bother?
     
  12. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    Yes, because it is a research project too, so it would be nice to comment some differences in the conclusions about how different is the supposed results in front of the experimentals.

    If you know a method to simulate this, could you help me please?

    Marc
     
  13. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,836
    2,071
    Here's one way, but I doubt it's much like the real thing :-
    SparkGap.PNG
    The bits inside the dotted rectangle are the simulated gap.
     
  14. RichardO

    Late Member

    May 4, 2013
    2,274
    890
    And don't forget the effects of altitude. Arcs happen a lower voltages at higher altitudes.
     
  15. Bordodynov

    Well-Known Member

    May 20, 2015
    1,979
    597
    See Change in the breakdown voltage as a function of the rate of change in voltage. Changing the parameters of the model, you can reproduce the actual behavior of the arrester.
    2018-08-02_09-02-52.png
     
  16. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    But the simulation that you did doesn't specify the spark gap shape, right? I mean it can be triangular, semicircular, or any invented shape.
    Btw, where did you get the equivalent circuit of a spark gap?

    Thanks for the info. We will consider it in the experimental phase.

    I don't understeand at all the image, can you explain it a bit please? I'm a bit newbie with these softwares. I have used PSIM before but I don't think I can do this with it
     
    Last edited: Aug 2, 2018
  17. Bordodynov

    Well-Known Member

    May 20, 2015
    1,979
    597
    I change the linear voltage input voltage.
    In order to show the dependence of the breakdown on the rate of voltage buildup, I do the calculation several times, changing the rise time to 1000 volts during the "T" time. This is done by a directive (several calculations):
    .step param T list 100N 1u 100u 1m 10m 100m
    The first calculation (the result is green) slew rate: 1000V / 100ns = 10000V / us = 10kV / us. The last calculation (slow buildup of voltage), slew rate: 1000V / 100ms = 0.01V / us
    I also used postprocessing processing of results (measured breakdown voltage) "MEASURE".
    Measurement: max
    step MAX(v(y)) FROM TO
    1 467.632 0 9.99e-008
    2 421.139 0 9.99e-007
    3 231.864 0 9.99e-005
    4 144.359 0 0.000999
    5 111.016 0 0.00999
    6 101.941 0 0.0999
    2018-08-02_10-46-38.png
     
  18. Bordodynov

    Well-Known Member

    May 20, 2015
    1,979
    597
    I'm trying to convey to you the idea that, assuming you calculated the breakdown voltage of your self-made 100-volt discharger, this does not mean that the limit will be 100 volts (there will be more). It is necessary to give real voltage and take into account the inertia of the arrester!
     
    ebeowulf17 likes this.
  19. marc1996

    Thread Starter New Member

    Aug 1, 2018
    18
    0
    Yes, I knew that, but still, I'm not able to understeand the pictures at all, maybe because I don't know what those parameters mean (for example Ron, R0..) Also, I don't know where you modify the gap distance between the electrodes.
     
  20. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,836
    2,071
    I devised the model using the formula you posted to give the breakdown voltage of a parametric diac and added in some track inductance, some leakage resistance and gap capacitance. The model assumes the breakdown would occur only at the point of minimum separation (1mm in the example) of two track portions, hence shape is irrelevant to that model. I have no idea how you would model any shape effects, but I suspect they would mainly modify local inductance and capacitance around the breakdown point. I did warn you the model is probably unreal :) .
     
Loading...