Tesla Coil Speaker Frequency Response
Mike, Fri Dec 08 2006, 05:14AMSo after the help of Steve W. and Jimmy, I have formed plans for my Science Fair. ( I am doing one this year mostly since i get 10% extra credit for it in AP Bio and its free funding to build a TC from my parents;).
I plan to build a SSTC using Steve Conner's DWSSTC driver that has an audio input at this page:
Schem pic here ( 
)I will be audio modulating it with a range of frequencies in a higher band for comparison towards normal cone speakers since they tend to deform at higher frequencies. I will be measuring the amplitude using a microphone.
So, Anyone have any comments/suggestions/criticism? Also, is Conner's DWSSTC circuit the right one to use for this project. I figured it was good since it already has a simple Audio in and looks pretty simple to build.
Thanks for the advice.
Mike
Re: Tesla Coil Speaker Frequency Response
Reaching, Fri Dec 08 2006, 06:01AM
steve conners circuit works good for a small audio modulated sstc so if you dont like complicated things you can use it. i dont know how high the frequency could be with this circuit but for a good audio output without the hissing from the sparks you need a high resonant frequency in a range of 1mhz or above with a high Q secondary
Reaching, Fri Dec 08 2006, 06:01AM
steve conners circuit works good for a small audio modulated sstc so if you dont like complicated things you can use it. i dont know how high the frequency could be with this circuit but for a good audio output without the hissing from the sparks you need a high resonant frequency in a range of 1mhz or above with a high Q secondary
Re: Tesla Coil Speaker Frequency Response
..., Fri Dec 08 2006, 07:03AM
I would stay away from the pll coils and go for Dan's plasmasonic tl494 based ones. My reasoning behind this that that the pll coils can be tricky to get tuned right (although not necessarily) and it is hard to get a high amount of modulation into them before they start to get huge amounts of distortion. I built one using that schematic running at 700khz with ~1" of spark and while the sound was audible it was nowhere near a decent listening level before the distortion starts to kick in.
Not saying that all coils will have that problem, but...
Also, with the plasmasonic topology you only have 2 fets on the GDT's which makes the gate drive a heck of a lot easier. It isn't easy to swing 4 2nf gates at 1mhz (trust me.)
..., Fri Dec 08 2006, 07:03AM
I would stay away from the pll coils and go for Dan's plasmasonic tl494 based ones. My reasoning behind this that that the pll coils can be tricky to get tuned right (although not necessarily) and it is hard to get a high amount of modulation into them before they start to get huge amounts of distortion. I built one using that schematic running at 700khz with ~1" of spark and while the sound was audible it was nowhere near a decent listening level before the distortion starts to kick in.
Not saying that all coils will have that problem, but...
Also, with the plasmasonic topology you only have 2 fets on the GDT's which makes the gate drive a heck of a lot easier. It isn't easy to swing 4 2nf gates at 1mhz (trust me.)
Re: Tesla Coil Speaker Frequency Response
Steve Conner, Fri Dec 08 2006, 03:02PM
My circuit works by FM modulating the audio, and the Tesla resonator itself converts it to AM by slope detection. So there's bound to be a good deal of distortion, since the slope is non-linear. The clearest sound happens when the frequency is tuned a wee bit above resonance, until the spark is about half of the maximum size. If you tune it for maximum spark length, the audio will be quiet and distorted. I'll leave it to you to figure out why this happens.
Don't tune below resonance whatever you do, as this stresses the MOSFETs a lot by making them switch a leading current.
As for frequency, blackplasma is the expert on this, I believe he got my circuit working with audio modulation at 6MHz or something crazy. He used a 74HC4046 instead of the regular 4046.
The PlasmaSonic works by directly amplitude modulating the signal using PWM. I don't know the advantages and disadvantages, but I imagine it works better.
Steve Conner, Fri Dec 08 2006, 03:02PM
My circuit works by FM modulating the audio, and the Tesla resonator itself converts it to AM by slope detection. So there's bound to be a good deal of distortion, since the slope is non-linear. The clearest sound happens when the frequency is tuned a wee bit above resonance, until the spark is about half of the maximum size. If you tune it for maximum spark length, the audio will be quiet and distorted. I'll leave it to you to figure out why this happens.
Don't tune below resonance whatever you do, as this stresses the MOSFETs a lot by making them switch a leading current.
As for frequency, blackplasma is the expert on this, I believe he got my circuit working with audio modulation at 6MHz or something crazy. He used a 74HC4046 instead of the regular 4046.
The PlasmaSonic works by directly amplitude modulating the signal using PWM. I don't know the advantages and disadvantages, but I imagine it works better.
Re: Tesla Coil Speaker Frequency Response
thedatastream, Fri Dec 08 2006, 03:59PM
If using a rectified mains input you will need some decent decoupling on your rectified voltage otherwise you'll end up with some background 100Hz hum in the audio. Stick plenty of uF in there.
James
thedatastream, Fri Dec 08 2006, 03:59PM
If using a rectified mains input you will need some decent decoupling on your rectified voltage otherwise you'll end up with some background 100Hz hum in the audio. Stick plenty of uF in there.
James
Re: Tesla Coil Speaker Frequency Response
WaveRider, Fri Dec 08 2006, 04:12PM
If you want good quality audio, you will need a smooth DC power supply (no unfiltered 1/2 or full wave rectified DC, I use a 10000uF elco). If you try to modulate the frequency of oscillation (Ã la Steve C), it is important the the PLL loop bandwidth be as small as possible (altho' there is a trade-off associated with being able to track the plasma loading and keeping the PLL locked). The PLL will generally act like a high-pass filter on your audio in that configuration (as it constantly tries to keep the VCO output 90 deg out of phase with the sense probe).
In my PLL plasma audio experiments, I use a PLL with a fairly large loop bandwidth (greater than 20 kHz) so the PLL tracks well the changing loading as a result of the audio. I can do this because I modulate the drain voltage of the power MOSFET that drives the coil. The PLL does its best to keep the coil at resonance as the plasma loading changes with the audio. It is a more complicated method, but the sound quality is outstanding over a wide range of volumes.
WaveRider, Fri Dec 08 2006, 04:12PM
If you want good quality audio, you will need a smooth DC power supply (no unfiltered 1/2 or full wave rectified DC, I use a 10000uF elco). If you try to modulate the frequency of oscillation (Ã la Steve C), it is important the the PLL loop bandwidth be as small as possible (altho' there is a trade-off associated with being able to track the plasma loading and keeping the PLL locked). The PLL will generally act like a high-pass filter on your audio in that configuration (as it constantly tries to keep the VCO output 90 deg out of phase with the sense probe).
In my PLL plasma audio experiments, I use a PLL with a fairly large loop bandwidth (greater than 20 kHz) so the PLL tracks well the changing loading as a result of the audio. I can do this because I modulate the drain voltage of the power MOSFET that drives the coil. The PLL does its best to keep the coil at resonance as the plasma loading changes with the audio. It is a more complicated method, but the sound quality is outstanding over a wide range of volumes.
Re: Tesla Coil Speaker Frequency Response
Mike, Fri Dec 08 2006, 10:59PM
Hi, thanks for all the comments, It does indeed seem like the Plasmasonic II is a good candidate. I think I will probably use that one, any other ideas on methods of measuring the amplitude. I know a mic but anything more specific?
Thanks,
Mike
Mike, Fri Dec 08 2006, 10:59PM
Hi, thanks for all the comments, It does indeed seem like the Plasmasonic II is a good candidate. I think I will probably use that one, any other ideas on methods of measuring the amplitude. I know a mic but anything more specific?
Thanks,
Mike
Re: Tesla Coil Speaker Frequency Response
..., Fri Dec 08 2006, 11:44PM
Decibel meter? I have one that we use to test how loud it is in my house as we add more and more insulation. Think we paid like $40 for it at radioshack.
..., Fri Dec 08 2006, 11:44PM
Decibel meter? I have one that we use to test how loud it is in my house as we add more and more insulation. Think we paid like $40 for it at radioshack.
Re: Tesla Coil Speaker Frequency Response
EDY19, Sat Dec 09 2006, 12:24AM
If you use a microphone, you might want to make sure that it has a level frequency response as well.
EDY19, Sat Dec 09 2006, 12:24AM
If you use a microphone, you might want to make sure that it has a level frequency response as well.
Re: Tesla Coil Speaker Frequency Response
Steve Conner, Sun Dec 10 2006, 09:56AM
Measuring the frequency response of speakers can be tricky to do right. Even if you use a mic with a perfectly flat frequency response, the acoustics of the room you do the measurement in can cause huge peaks and dips, especially at low frequencies. You're really supposed to do it outdoors or in an anechoic chamber.
Real calibrated mics can cost several hundred dollars, but apparently the little electret mic capsules that you can get at Radio Shack or wherever are reasonably flat. You can use these with a PC sound card and some sotware to draw a frequency response plot.
Or you can use ...s decibel meter approach, with a set of third-octave pink noise test tones. You play each tone and write down the sound level shown by the meter, and you can draw a kind of coarse bar chart of the frequency response. You can get a CD of the test tones from quite a few places, as they're used for setting equalizers in PA systems and such like. If you don't have access to proper anechoic conditions, this method is arguably better because it averages out the wildest ripples due to the room. Just don't move the meter in between readings, put it on a stand or something.
Steve Conner, Sun Dec 10 2006, 09:56AM
Measuring the frequency response of speakers can be tricky to do right. Even if you use a mic with a perfectly flat frequency response, the acoustics of the room you do the measurement in can cause huge peaks and dips, especially at low frequencies. You're really supposed to do it outdoors or in an anechoic chamber.
Real calibrated mics can cost several hundred dollars, but apparently the little electret mic capsules that you can get at Radio Shack or wherever are reasonably flat. You can use these with a PC sound card and some sotware to draw a frequency response plot.
Or you can use ...s decibel meter approach, with a set of third-octave pink noise test tones. You play each tone and write down the sound level shown by the meter, and you can draw a kind of coarse bar chart of the frequency response. You can get a CD of the test tones from quite a few places, as they're used for setting equalizers in PA systems and such like. If you don't have access to proper anechoic conditions, this method is arguably better because it averages out the wildest ripples due to the room. Just don't move the meter in between readings, put it on a stand or something.
Re: Tesla Coil Speaker Frequency Response
thedatastream, Mon Dec 18 2006, 10:04AM
I had a thought last night - how much of the frequency response is determined by the Q of the resonator itself?
Assuming a fairly high Q, changing the amplitude of the drive signal would not instantaneously change the output voltage as the resonant circuit would take time to decay. However, if there is a streamer or arc on the output of the resonator, that will de-Q the system and response will be quicker.
If the coupling to the secondary was reasonable, then the energy in the resonator would be "sucked" back into the primary when the drive was lowered, speeding the response time.
Comments \ thoughts?
James
thedatastream, Mon Dec 18 2006, 10:04AM
I had a thought last night - how much of the frequency response is determined by the Q of the resonator itself?
Assuming a fairly high Q, changing the amplitude of the drive signal would not instantaneously change the output voltage as the resonant circuit would take time to decay. However, if there is a streamer or arc on the output of the resonator, that will de-Q the system and response will be quicker.
If the coupling to the secondary was reasonable, then the energy in the resonator would be "sucked" back into the primary when the drive was lowered, speeding the response time.
Comments \ thoughts?
James
Print this page