Singing Tesla Coil

[Phil_Jahelka]

Hey everybody!

I'm working on a Singing Tesla Coil for a 4-H electricity project and I just finished my first draft of the circuitry. I was wondering if anyone would be kind enough to give me some tips or point out any glaring mistakes I made. Thanks for your help.

Relevant info: the voltage regulator is a LM317, the Op-Amp is a LM741, the PWM is a UC28025, and the MOSFETs are FDA20N50F's, the 18.5V power source is an old laptop power supply, and the audio is coming from an old Walkman CD player.

 

[Resqueline]

The first mistake that glares is the output stage. You can't drive it like that (why, you'd get only 15V ampliude to the coil). Check out high-side/low-side driver IC's.

 

[Phil_Jahelka]

I'm afraid I don't understand. I thought one of the great bonuses of MOSFET's are that the voltage they switch isn't limited by the gate voltage, unlike a BJT.

 

[Phil_Jahelka]

Also, I just realized that the bridge rectifier is backwards. Please pretend I never made that horrific mistake. In addition, I would really like to avoid using GDTs. They're prohibitively expensive and difficult to make. And if I need them, it means I fried the MOSFETs which are the majority of the expense of this project, so I don't really care about cost after those fry.

 

[blocka]

To turn the FET's on Vgs needs to be positive 3-5 volts.

When M1/M2 are on the voltage at the source of M2 is going get to the gate drive voltage of the PWM chip minus Vgs threshold and turn off.

 

[Phil_Jahelka]

Ughhh... I finally get it. Sorry it took me so long to figure out something so painfully obvious. Thanks for your guys' patience. Anywho, what about the audio input? Did I get that part right? Also, I realized how much of a murder machine this thing could be so I added a Isolation Transformer between the rectifier and the mains. I'll put up a new schematic including it when I figure out the driver ICs.

 

[blocka]

Audio part. Why not use the opamp in the PWM chip.
As it is you have connected the output of two opamps together. And left the input to the PWM error amplifier floating. So dont think this will work.

The reference voltage of the chip is 1.25V, so your bias level needs to be this, but this is already done if you use the pwm chips error amp.

 

[Phil_Jahelka]

OK, here's the next draft. I included the drivers, but I'm not sure what value of resistors to use on the MOSFET connections. Also, I'm not sure if I need the boot-strap circuitry, but I included it anyways. It also includes blocka's idea of using the PWM for the audio, which is awesome because it drastically simplifies the circuit. Sorry for not showing all the connections on the far right. Resolution is starting to become an issue and I promise I wired it right. Again, thanks for your continued patience and help.

 

[Phil_Jahelka]

A lot of work has been done on it with people from another forum. Here is the most recent draft.

 

[Resqueline]

That looks better. Now, I haven't studied the datasheets or every aspect of the circuit but I question why you are using a separate chip for reference voltage, and why you are injecting the audio directly into it's output (what are the impedances?) when there's an internal reference in the pwm chip you could divide down to whatever - and thereby getting control over the audio level too.. Also, wouldn't the switching frequency need to be adjustable to suit the coil's resonant frequency?

 

[Phil_Jahelka]

That's a great idea to use the chips internal reference and some resistance to get the right bias on the audio; it would really reduce the parts count. Also, I already have the coil designed and it resonates at just about exactly 100kHz. The current set up is set resonate at that, but just in case I'm making the 5K RT a pot. I should have made it explicit, but I couldn't find a pot symbol in the software. Here's a new schematic incorporating the Vref bias and using the PWM's Op-Amp with adjustable gain.

 

[Resqueline]

I'd figure R4 should go straight to pin 1, and R5 should be connected between pin 1 & 3. There may be something more too but I still haven't consulted the datasheets.
Also, make sure the circuit ground is connected to real ground or else the isolation transformer may be exposed to flashovers.
Tesla coil resonances are very much affected by their environment (& top load) so I guess it would be useful to make the pot external & easily adjustable.