[Home][2017 Index] Re: [TCML] MMC [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] MMC

Thank you all for your response, I made a short video to show the Caps I was referring to. Had I known more about what I was doing when I built them I would have added 3 to each group as Bert suggested, isn't hindsight wonderful? 
https://youtu.be/P13sFpms2tI
Doug
-----Original Message----- From: Jay Howson LAST_NAME 
Sent: Tuesday, February 21, 2017 9:39 PM
To: tesla list
Subject: Re: [TCML] MMC

Let me give this a shot,
You are essentially asking what cap value would give the best performance out of you 12/30 NST using what seems to be 3 series lengths of 15 caps arranged is either parallel 15x3 or series 45x1. 

Doug,
Bart Anderson ( I am not sure if Bart was actually the originator, but ok) made a nice looking chart which proposes some efficient cap designs for MMC's, Specifically for the 942C20P15K (.150uF, 2kV) caps. Not sure if that is what you have, but your 2Kv .15u rating seems to imply that is it. 

The chart is found here
http://tb3.com/tesla/capacitors/mmc_cap_chart.pdf
It is my understanding that the capacitance's do have some science behind their values which matches them to the desired NST values. I believe it relates to the ability to charge the cap to the fullest extent in ~ 4mS. The time it takes a 60Hz sine wave to go from 0 to its max value. (ie a quarter of the wavelength) at the current output of the NST. But it's a little different as you don't want to resonate with the NST (as it will damage the NST if you are not careful (note, don't use values from the 1st section of the chart)). Typically we shoot for a slightly higher capacitance, which means a slightly lower resonant frequency. This way even though the capacitor is not fully charged at least all the power from the transformer went into it. (imagine a smaller value, where the cap fills up too fast and then the transformer isn't sourcing current for that portion of the AC waveform because the cap is full, that's wasted power!!!) 

Based on the chart,
If you have a static gap
I would stick with 1 string of 15 Series by 1 Parallel for a rating of 10nF and 30KV and have 2 extra sticks of 15 to boot! There used to be a nice set of plans produced by Resonant Research that did exactly this for a 12/30 NST, if i recall correctly. I havent seen these plans online in a while, but i'm sure you can find references in the TCML archives if you look.
If you really wanted to use all of your caps regardless of what the chart says, shoot for a slightly more durable MMC but close in value to what the chart recommends. I would disassemble your 3rd string into 5,5 and 1 sections and then add 5 to each of your 15 sticks and create a 22 series by 2 parallel MMC for a rating of 13.6nF at 44KV, not too shabby. 
Its not all that far beyond 10nF.
(Remember the tesla coil is all about tuning, so you cant really have a wrong cap value, (unless you pick the resonant one, as discussed earlier), you simply adjust other coil parameters to make it work and go about your day) But is this kind of durability necessary on a 12/30. It couldn't hurt, but probably not. 

If you have a rotary gap, (which is sort of the next level, so to speak)
I would disassemble your 3rd string a little and shoot for an 18 series by 2 parallel strings for a rating of 16.7nF at 36KV. As recommended by the chart. 

I hope you find this informative.
-Jay




    ---------- Original Message ----------
    From: Bert Hickman <bert@xxxxxxxxxxxxxxxxxxxxx>
    To: Tesla Coil Mailing List <tesla@xxxxxxxxxx>
    Date: February 21, 2017 at 3:02 PM
    Subject: Re: [TCML] MMC

    Hi Doug,

    Tesla Coil caps see a very nasty oscillatory, high-current, high
    rep-rate, high-voltage environment that is extremely challenging to a
    capacitor's dielectric and metalization system. Fortunately, the
    self-healing feature of these particular caps allows you to overstress
them without suffering immediate catastrophic failure. It allows you to trade off capacitor run-time life versus initial cost for the capacitor bank. Excessive voltage stress on your self-healing caps results in the 
    eventual death of your tank cap, and the greater the overstress, the
    shorter the expected life.

    For TC use, moderately conservative design practice suggests that each
MMC string should have a total DC rating no less that 2.5X - 3X your HV source RMS face plate voltage rating (30 - 36 kVDC for your 12 kV NST). 
    However, you CAN choose to use a lower factor - with more risk and
    reduced lifetime. Some MMC design charts even show a factor as low as
    1.33X (i.e., only 8 caps in series for a 12 kV NST!). Using fewer caps
    in each MMC string significantly increases the voltage stress on each
    cap. In TC caps, overvolting is most often caused by voltage reversals
in the ringing tank circuit. The dielectric system of a cap that rapidly 
    reverses polarity "sees" a voltage stress that's 2X as high as the
    initial capacitor voltage.

    Initial symptoms of overvolting are typically silent, showing up as
    small sparks (partial discharges) along the boundary between the
    capacitor's metalization and adjacent dielectric. When the dielectric
    fails, a short-circuit and self-healing event occurs. Self-healing
    events near the outside tape layer can sometimes be seen as flashes of
    light. Once this destructive process begins, it progressively chews at
    the polypropylene dielectric, causing large numbers of
short-circuit/self-healing clearing events. These progressively damage, 
    and eventually destroy, your tank caps. Depending on the degree of
overstress, this can take minutes, hours, or days. By using more caps in 
    each string to reduce voltage stress, the degradation process can be
avoided and the usable lifetime of the caps can be extended indefinitely. 

    Looking at the physics and of the internal structure of these
    self-healing capacitors, I would recommend using a factor no less that
    2.5x to 3x Vsupply(RMS), or 15 to 18 caps/string for your 12 kV NST.
    If you only need a few hours of run-time life, you can further reduce
    the number in each string to perhaps 2X or less. I wouldn't, but you
    can... :)

    Bert

    doug wrote:

        > >
> I have 3 MMC’s each consisting of 15 .15u X 2 Kv Caps. [10nf X > 30Kv] I can connect them P or S. Which would give me the best setup > using a 12X30 NST. 
>         Doug
>
>         _______________________________________________
>         Tesla mailing list
>         Tesla@xxxxxxxxxx
>         http://www.pupman.com/mailman/listinfo/tesla
>
>         _______________________________________________
>         Tesla mailing list
>         Tesla@xxxxxxxxxx
>         http://www.pupman.com/mailman/listinfo/tesla
>
>     >

_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla 
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla