If you interested in, here is a good page to create really simple and good PWM generator:

http://www.cpemma.co.uk/pwm_erg.html

You can find elements which is responsible for freq and PWM.

" ... frequency = R2 / (4 x R3 x R1 x C1) ..."

Todo: eliminate 2N2222 and motor driver then put 74HCT14 Schmitt-trigger (with 6ns raise time) to the output of final OpAMP. Then connect to Mostfet's driver the 74HCT's output. Note: 74HCTxx require 5Volts ,so we need additional 5V stabalizer block.)

Regards,
Replicator

Jason,

your FET driver circuit has a big error.  The manufacturer's spec sheet shows a test circuit which they use to determine the switching speeds.   They simulate the large FET input capacity by using a 1nF on the driver output.

In the real world, when you use this FET driver, you must remove this 1nF capacitor; the FET gate already has 1nF or 2nF capacity and you don't want to have an additional capacitance in the form of a 1nF capacitor!

As additional comments,

the 4uF7 capacitor should be labeled tantal
the 0uF1 capacitor should be marked ceramic
it should be mentioned that additional parallel ceramic capacitors never hurt.  For example, additional 1nf and 10nF bypass capacitors.

Pay a lot of attention to the polarity of the tantal capacitor, if wrong polarity the tantal will catch fire and turn into a block of carbon.  Quite a spectacle.

Regards, Earl

Quote from: Jdo300 on June 19, 2007, 07:57:02 AM

Hello All,

I'm finally back from my weekend trip and thought I would take the time to draw up the the circuit diagram for my control circuit.

Hi Earl,

Thanks for the heads up about the 1 nF caps on the driver outputs. This is exciting news since the drivers have been working extremely well (achieving 38ns switch-on times), so it should be VERY good without them! I was experiencing some problems at one point where the FETs were not turning off all the way (when loaded) which may have been due to the cap sitting there charged up. I'll also add the labels for the cap types to the circuit diagram. I am currently using small electrolytic caps for the 4.7 uF filter caps so I'll see about getting some tantalum ones to replace them.

Again, thank you for your suggestions .

God Bless,
Jason O

Hello all,

@Darren,

I totally agree with you. The point is, there are a lot of new people....not in electronics...

With 3 synchronised frequencies its muuuuch better and easier to work but if you have such a success as I had then there is no equipment anymore.
I wanted that the people first learn how to mix, whats going on with the current and voltage from the power supply and then synchronise the frequencies.

@Roberto

be careful.

Otto

Hi Otto,

I'm already headed down the harmonic road as well. I have a bunch of 12-bit ripple counter ICs that I am using to make a 12-harmonic frequency divider board. The board is already made but it has some glitches that still need to be worked out before I can use it. (I originally made it to test out Roberto's original TPU setup with the two harmonics locked in). For anyone else interested, the IC is a 74HC4040 and would be a cheap and easy way to make this. My board has three 12-pin dip switches that route any of the 12 harmonics to the three outputs to drive the MOSFETs. I'm not home at the moment, but I'll post a photo of it later.

God Bless,
Jason O

http://www.hauntedfrog.com/gt/movies/2007/duckon/SingingTeslaShow.html

thanks for the link, reminds me of a plasma speaker,  on steroids lol

http://www.plasmatweeter.de/

Dirt

GK,

CC - Control Coils - 1" width

Tubing for collector = 1/2"


Tim

@Roberto,
The divide by "n" device CD4059A you have mentioned is an excellent idea, BUT I think you need to still need to feed it with a function generator to get desired master frequency to start the divide by.
Feeding it with a xtal oscillator will limit the number of output frequencies.
I plan to do the same thing and this device is just what I have been looking for.

@Jason,
Although your function generator boxes do not support duty cycle control, you can however simply have an op-amp connected to the triangle wave output.
You adjust the -ve bias using a trimmer on the op-amp to determine at what point the op-amp turns on and off. Then using the amplitude control on the function generator you can set the duty cycle.
Simple.
I still have to test this with my DDS 20 function generator, except I have to use the sine wave output as the feed to the +ve bias of the op amp, not quite as good as using a triangle wave output, but it will do.

Not sure about the ripple counter idea, I think you would be better with what Roberto has suggested with 3 X CD4059A.

I think this is the way everyone should head towards, as this guarantees a synchronised signal in all 3 control coils.


@All
I have ordered up two more DDS 20's in kit form this time, so its going to be callenging soldering up the surface mount AD9835 chip with 0.5mm pin spacing.

I was thinking what WaveWatcher was saying about the clues in the post from SM, maybe it was SM himself???
I was trying to figure out the tie-up between 35.705KHz harmonic and 245KHz output.
So what combination will fit them together exactly?
Maybe 245KHz was a ballpark figure so as not to give the farm away?
 
If you have:
249.935KHz / 7 = 35.705KHz
249.935KHz / 5 = 49.987KHz
249.935KHz / 4 = 62.48375KHz

These would all be in perfect sync if you use 3 x divide by n devices all fed from a common function generator.

Also, there has been no mention of a tuning capacitor so the coils can be tuned to their harmonic frequencies.

Regards
Rob