high frequency coil very sensitive to nearby objects / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

high frequency coil very sensitive to nearby objects

1 2 next

Move Thread

LAN_403

Crunchy Frog

Sat Jul 24 2010, 02:35AM

Registered Member #2422 Joined: Tue Oct 06 2009, 02:41AM
Location:
Posts: 85

I'm building a class e 4Mhz coil, based on

but with no GDTs. I'm running it at 12V now so I can make sure it all works before turning up the power.

What I've found is the drain waveform on the IRFP450 (which i'm using instead of the mtw14n50) is *very* sensitive to conductive objects (like my hand) near the coil. (picture of this attached) So it would seem like the capacitance is affecting the Fres of the coil. But, if I stick an unconnected probe near the coil, the signal it picks up is basically pinned on 4 mhz.

So my best guess is my hand is creating a capacitatively coupled ground, increasing the loading on the class e amp, and changing the class e Fres. I also found that with the smaller mosfet, its class e Fres would be a lot higher if it was connected to the gate of the irfp450 as opposed to left floating.

So what I'm wondering is:

What is it about a bigger load that increases the Fres?
Is this common/normal?
Is there any way to prevent it?

And the rest of the setup:

Arcstarter

Sat Jul 24 2010, 03:12AM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

It adds topload capacitance. That makes the resonant frequency drift until the primary and secondary are nowhere near resonance.

Crunchy Frog

Sat Jul 24 2010, 03:40AM

Registered Member #2422 Joined: Tue Oct 06 2009, 02:41AM
Location:
Posts: 85

But, when I put a probe connected to a piece of tinfoil near the topload, the signal it picks up stays at 4 mhz no matter what.

cjk2

Sat Jul 24 2010, 04:22AM

Registered Member #51 Joined: Thu Feb 09 2006, 04:17AM
Location:
Posts: 263

"But, when I put a probe connected to a piece of tinfoil near the topload, the signal it picks up stays at 4 mhz no matter what."

Correct. I assume you are not using feedback or auto tuning? If you use a fixed clock source, for example a crystal, your driver will output that frequency regardless of if your coil is in tune or not.

HVgeek

Sun Jul 25 2010, 10:17AM

Registered Member #2998 Joined: Tue Jul 13 2010, 08:34PM
Location: Swedish forests.
Posts: 26

...And thus your performance will plummit since you're not in tune anymore with the capacitive grounding-effect of your hand.

GeordieBoy

Mon Jul 26 2010, 12:24PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

For proper Class-E operation the TC in the output network needs to be resonant at a frequency somewhat lower than the drive frequency. This provides the necessary inductive reactance to make the Class-E amplifier switch smoothly.

So your TC might be resonant at about 3.8MHz when driven via a link coupled primary. When driven at 4.0 MHz from say a crystal oscillator there is enough resonance to develop sparks but it reflects a small amount of inductance back to the PA at 4MHz. This is what enables the drain voltage to swing back down to zero before turn on.

If the amount of inductive reactance is just right you get the smooth landing at zero volts with zero gradient right where the device turns on. If the amount of inductance isn't spot on then the drain voltage either fails to reach zero, goes through zero too early, or tries to go negative and bottoms out as shown in your pictures. In practice getting within a few volts of ground at turn-on is good enough to eliminate most of the turn-on losses. It doesn't need to be perfect.

Whatever way you look at it the driver runs at a fixed frequency and you vary the tuning/coupling of the TC resonator to control the nature of the load reflected back to the Class-E power amplifier.

It is this "off resonance" tuning of Class E amplifiers that makes them narrow-band devices and quite sensitive to changes in load impedance. That is why they are not very well suited to audio modulation by frequency-modulation. FM'ing the carrier quickly moves the Class-E amplifier away from it's ideal operating point where efficiency is maximum.

-Richie,

Crunchy Frog

Mon Jul 26 2010, 06:51PM

Registered Member #2422 Joined: Tue Oct 06 2009, 02:41AM
Location:
Posts: 85

Thank you! That explains everything!

doctor electrons

Mon Jul 26 2010, 09:54PM

Registered Member #2390 Joined: Sat Sept 26 2009, 02:04PM
Location: Milwaukee Wisconsin
Posts: 381

For Geordieboy!
Question! If the coil is not switching in class-e operation, or very close to it. This would lead to hard switching of
the mosfets and eventually premature failure from heating?

Arcstarter

Mon Jul 26 2010, 10:01PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

doctor electrons wrote ...

For Geordieboy!
Question! If the coil is not switching in class-e operation, or very close to it. This would lead to hard switching of
the mosfets and eventually premature failure from heating?

Yes, just like not having the capacitor in the first place :).

Crunchy Frog

Tue Jul 27 2010, 02:44AM

Registered Member #2422 Joined: Tue Oct 06 2009, 02:41AM
Location:
Posts: 85

Well not necessarily hard switching. If the load is under-damped, the drain voltage will try to swing negative and the fet's body diode will start to conduct. So it would still switch on with zero volts gate-source, but the diode will be put into reverse recovery, which creates extra switching losses. Read

for a good explanation.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob