I intend to build a high-power series LC resonant circuit for the purpose of generating strong magnetic fields at a wide range of frequencies, mostly audio range (1-5000 Hz). The inductor coil is about 5-10 mH and 3.3 ohms of 15-gauge copper wire, so I will need about 60-80 uf maximum capacitance to achieve low impedance at the lowest frequencies in that range. The circuit will be driven by a power H-bridge at 100 Volts amplitude, so I calculate that the peak current will be at my target of 30 amps assuming no net reactance.

I wish to be able to sweep the frequency in an automated fashion, so I would like feedback on my design for a high-current high-capacitance programmable variable capacitor.

I will have about ten 6.0 uF power capacitors and a few lower capacitance capacitors decreasing from 6.0 uF in powers of 2 all in parallel. My idea is to have a high-current mechanical relay on each side of each capacitor, with the switching leads of every pair connected and switched simultaneously by a logic-level MOSFET, which is controlled by the same MCU that controls the H-bridge. There will be an option to bypass the capacitor bank altogether with an extra relay for very low frequencies (<100 Hz). I will calibrate the design ahead of time to determine the lowest impedance capacitor combination for any given frequency. The relays have a contact current rating of 40 amps and the capacitors are rated at 50-70 amps RMS. I will have a few separate fans blowing air onto the coil, capacitor bank, and H-bridge (MOSFETs passively cooled at 10 C/W).

I don't have much experience with power electronics, so I was wondering if any of you had design suggestions or can see some obvious pitfalls? Are there any already existing designs for high-power variable capacitors like this that I am not aware of?

EDIT:

I understand that my design idea will certainly not generate perfect resonance at all frequencies, with a noteworthy high-impedance region below about 200 Hz. Using the additional capacitor bank bypass relay (just the coil without capacitors) I could significantly improve current output at that region (see calculations). Some imperfection is definitely acceptable.

My end goal is to explore the effects of electromagnetic frequencies on biological and crystalline systems, and I would like the final coil current to be 20 Amps RMS at any frequency between 1-5000 Hz. To be able to achieve this either requires an enormous amount of voltage/power at 5000 Hz or some form of adjustable capacitance. I apologize if that is too vague, perhaps I need to reflect more on my design goals.

Since the magnetic field generated is proportional to current, I figured that using capacitors for resonance would allow the maximum current to flow in a more frequency-uniform fashion, especially at the higher end of the 1-5000 Hz region. I suppose I could use a class D audio amp alone with no capacitors, but the current and magnetic field would be lower at these higher frequencies. A 3 kW class D audio amp is likely cheaper than an H bridge + 3 kW DC power supply, so it is certainly an option if used with the variable capacitor bank. Would you recommend this?

Calculations:

Assuming an inductance of 6 mH and maximum capacitance of 66 uF (minimum resonant frequency 252.92 Hz), the frequency at which the purely inductive reactance (capacitor bank bypassed) matches the net reactance of the inductor in series with 66 uF is 178.837 Hz.

XL = XC - XL
w*L = 1 / (w*C) - w*L
2*w*L*w*C = 1
w = sqrt(1/2LC)
2*pi*f = sqrt(1/2LC)

f = 178.837 Hz

Assuming that I use enough smaller capacitors such that this is the highest reactance frequency, the maximum impedance is:

Xmax = 2 * pi * 178.837 Hz * L = 6.742

Rohmic = 3.3 ohms

Zmax = sqrt(Xmax ^ 2 + Rohmic ^ 2) = 7.506 ohms

I need an RMS power of:

P = I^2 * Z
P = 20^2 * 7.506 = 3002.52 W

and an RMS voltage of:

V = IZ = 150.12 Volts

At 10 mH, redoing these calculations we would need 3723.38 W max, so 4000 W RMS is probably a safe power to go with to get at least 20 A at any frequency in the range.

posted over 1 year ago

CC BY-NC-SA 4.0

1y ago

dcsuka‭

11 reputation 1 0 2 14