[Prev][Next][Index][Thread]
Re: "Enhanced" Multi Mini Capacitors Part 2
Hi Terry, MMCīers,
My comments interspersed.
>I have been working for the last few weeks on "EMMC" caps.
>I have been digging through data sheets, talking to "gurus",
>and doing some testing. My goal is to find out just how hard
> they can be pushed and see what could be done
> to improve them.
Pushing electronic parts to the limit seems like a second
nature to some of us coilers (like me and you) .... ;o))
> Here are the specs (theoretical)
> Capacitance = 5.6nF +-5%
> Configuration = 10 x 56nF 1600 volt metalized polypropylene capacitors
I hope these caps use metal FOIL for the endplate connections. I have
found that caps with metalized plastic endplates DO NOT have a very
high lifetime. You, too, are using (recommend?) "high" capacitance
for the single cap, right? This keeps Imax (being equal to Imax per cap)
of the string up nice and high (=safe value). I think there are two basic
important factors, which need to be adressed when building an MMC:
a.) The dv/dt times the capacitance of each single cap governs the
maximum current the string can take and should NOT be exceeded
(at least not by factors of >>1). For higher currents, more parallel
strings are necessary, which also gives us (okay, me) low voltage
coilers the needed higher capacitance as a "bonus" for equal Joules.
b.) The rise voltage per cap (dv/dt/n=number of caps per string)
should also not be exceeded by HIGH factors. Various cap
manufacturers I have "visited" (on the internet) test their caps with
3-5x the rated dv/dt before they pass inspection, but they limited
the number of pulses to ~10,000. Exceeding the dv/dt by a "casual"
factor should not kill them. For Telsa coil usage the following
formula should (I hope) give safe values an be of help when building
a MMC:
dv/DT = maximum rated value
Coil = Xkhz => T=1/F.
Of interest is the first 1/4 period.
((V*sqrt(2)* pi/2)/T) = SVr ("Seen voltage rise")
SVr / n (number of caps) = SVRpC (Seen voltage rise per single cap)
===> SVRpC <= rated dv/dt.
The AC rating of the capacitors is of little importance to us coilers
(the exception, which I still donīt understand {the derating factor},
might be the "WIMA 6kVdc, but only 700Vac rated" FKP1 units).
PP caps are designed to withstand the steep rise of square
wave pulses, which we donīt see in coiling usage, so exceeding it
a little (!) here shouldnīt hurt. Looking at the equation above shows
very clearly why the MMC should easily survive running on a bigger
coil (meaning lower secondary Fres). As the frequency goes down,
the ("T" gets larger) cap is stressed much less. As we are running
pulsed mode, the stress is also less than in CW coils. Building a
MMC for a high frequency VTTC, using the criteria described above,
will (Iīm pretty sure here) kill the MMC fairly quickly.
> Type = Panasonic ECW-H16563JV (Digi-Key P10516-ND)
> Voltage (peak) = 22000
> Current (RMS) = 3 amps (peak)=300 amps
So, you basicly did the same thing I did (PeakACV = DCV).
Well, you actually took it one step further, but:
> Burst Frequency = 50KHz to 250KHz
> BPS = 120
And this is the reason why I didnīt go past peakAC = DC rating.
Running them at or slightly below peakAC = DC rating should
allow me to go up "1 step" (meaning the 200bps SRSG idea I
posted a while ago) in the break rate w/o causing unnecessary
worries. High current, low voltage powered coils need ridicuously
large caps and the 200bps SRSG has a higher margin of operating
safety (I think).
> Series Resistance = 0.11 Ohms
> Drain Resistance = 100M ohm
> Drain Resistor Time Constant = 0.56 Sec.
> Maximum Temperature Rise = 5 C
> Thermal Dissipation = 4 C/watt
> Temperature Coefficient = 0.44 C/(Irms^2)
> Un-Cased Dimensions = 1" x 1.2" x 5"
> Construction time = ~30 minutes
> Cost = ~$22
I noticed you are once again using eq. resistors. My only concern,
in your new concept, would be the peak voltage (15kV*sqrt(2)/10)
of 2.2kV that each resistor has to withstand. Even good metal
film resistors are only rated for about 500-600V. A nice (;o) kickback
or strike to the primary might wipe these out, as you are already
overstressing them by a factor of 4-5. I donīt know if a oil/jelly
"insulator" will be of much help here. Iīm not sure if eqs are really
necessary in terms of reliability. I have measured each single caps of
my MMC after shut off (disconnecting the MMC from the primary
directly after shut off before measuring) several times. The residual
voltage across all caps is always less than 150V. The voltage across
the single caps is between 10-15V (depending on total residual voltage
of course). The cap to cap voltage variation was never more than 4-5
volts.
>I ran controlled current tests through it tonight at full power using
a primary circuit without a secondary in place. It didn't blow up! :-) The
> thermal heating was right where all the data said it should be.
> control and measure the RMS current going through it in the primary circuit.
> In a real coil, it should run just barely warm to the touch. I over load the
> voltage rating but these things are pretty underrated and everyone say the
> will be fine. The capacitance didn't change even though I ran it at
twice its
> dissipation for about 15 minutes. It is limited to 120BPS unless the circuit
> will be designed to keep the dissipation below a watt. However, it
really has
> plenty of headroom before it should blow up.
Although my tests are far(!) from controlled, I have never seen any
detectable heating in my present MMC design. Primary current is
around 360A and I am running 6 strings in parallel, so my current
per string is a low 60 (peak) amps. In my final 5.5-6kVA setup the
current per string(!) will not be any higher, so I expect the MMC to
last with ease. For these reasons, (and all mentioned above) I still
canīt understand why Richard Hull says MMCs are only good for
coils <<2kVA input. Richard, would you care to comment? How were
your MMCs built? Coils with less than 2kVA input will (usually) be
running at higher frequencies (200khz range). This will put more stress
on the caps than the higher current (esp. if the total is divided up using
more strings in parallel) will.
> Of course, the last two specs are the shockers! 30 minutes to build at
> 22 bucks! These are going to be what my new LTR coil uses if all goes well.
> So far, they seem rock solid! It is amazing that this little thing will do
> what my much larger MMC will do. It replaces 80 caps and has far less series
> resistance. Of course, compared to a big oil filled rolled cap, the
advantage
> is tremendous...
Not to forget the ease of tuning, repair, weight, voltage versatility
(!!) and less mess should something blow (which I doubt unless
you REALLY push it).
>I still need to cut caps out one by one and see where the voltage
>blows them.They really should be in some oil filled container or something.
>Corona and simply arcing across the leads is a real concern in something
>this small. Of course, they need a lot more abusive testing in a full Tesla
>coil system to be sure...
I will be awaiting your results, but it wouldnīt surprise me if you could
cut the number of caps down to six or seven (in your setup) before
they blow. The real question is (once you know the number "x") is what
safety margin is necessary. I think between +30-50% is a good value
to shoot for. Caps are designed for 10 of thousands of hours. If our
MMC "only" lasts for 2000hrs, we can run our coil every day (!) for 30
minutes for the next 11 years (2000/0.5/365). So, even if the MMC
costs $110, this would still only be $10 per year or 2.73 cents per day!!
If one canīt afford that, one shouldnīt coil [;o)))), because even your
electrical bill will be higher than that.
>If these EMMCs work out, the state of art in Tesla capacitors will be
> "changed"! So far, I am completely pleased with them!!!
(E)MMCs ARE the state of art in TC usage, in my mind. As I have stated
before, a MMC is not more difficult, more work, or more expensive to
build than a rolled poly AND it IS considerably less expensive than a
professional pulse cap. The MMC is really nothing less than such a
commercial PC, but it comes minus the huge price tag. Rolled poly caps
have to live with certain disadvantages which are excluded using the MMC
technique (like non uniform thickness and dirt inclusions, etc). Time will
tell, of course, but I think rolled poly, just like salt water caps and
dinosaurs, will become things of the past.
Coiler greets from germany,
Reinhard