Hi John,Dielectric heating may also apply to the former material in the secondary, so I'm not sure it's only the cap. But as the 942C is PP, dielectric absorption is very low. However, even with that being the case, it's frequency dependent and dielectric heating will increase. We are also pulsing the tank and not running in continuous mode (so that helps).
For polypropylene 942C caps, the dissipation factor can be estimated: DF = F/110000000 (F is frequency). Xc = 1/(2*PI*F*C) ESR = DF * Xc Q = 1/DFWith say a 0.01uF cap size at 100kHz, then DF = .00091, Xc = 159 Ohms, ESR = 0.1447 Ohms, Q = 1100. With say a 0.01uF cap size at 1000kHz, then DF = .0091, Xc = 15.9 Ohms, ESR = 0.1447 Ohms, Q = 110. Note that reactance changes but ESR remains the same. The closer reactance nears ESR, the more heating that will occur in the dielectric.
Frequency does make difference. You can also look at power loss: Power loss = (2*PI*F*C*V^2)*DFSo say you have 500V across a single 0.15uF cap: At 100kHz, 21.42W. At 1000kHz, 2142W. However, this is a continuous case and we are pulsing (beginning of discharge until quench with a low frequency rest between breaks).
I think the 942C caps will do fine at 1MHz due to the application, but there will still be higher losses at high frequency in comparison to low frequency, and there's no way around that. The question is will those losses make a considerable difference overall.
Bart johnbrooks@xxxxxxxxxxx wrote:
....so basically it boils down to the capacitor? Is there a configuration to use the 94C20P15K,,, as this is the current most used cap? j. On Sat, 21 Mar 2009 18:17 -0700, "bartb" <bartb@xxxxxxxxxxxxxxxx> wrote:Lau, Gary wrote:I want to point out that in the example coil I handed out, AC losses (both skin and proximity were accounted for at 1MHz) are 10 times 'lower' than a 1000 turn coil of the same geometry. If one were to use the same wire size as a 1000 turn coil and simply make the coil smaller to get the coil at 1MHz, then yes. But those losses can be canceled should an experimenter desire to eliminate them from the data. Dielectric losses are different and cannot be eliminated.Capacitor dielectric losses and inductor skin-effect and proximity-effect losses will be significantly higher at 1 MHz. Regards, Gary Lau MA, USATake care, Bart _______________________________________________ Tesla mailing list Tesla@xxxxxxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla
_______________________________________________ Tesla mailing list Tesla@xxxxxxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla