On 10/2/05, Philip Pemberton <philpem@dsl.pipex.com> wrote:
> In message <5eeda4c20510020944u53a561dcgd8da86737fe2772e@mail.gmail.com>
>           Vasile Surducan <piclist9@gmail.com> wrote:
>
> > Your pseudo sinusoidal signal it's a variable amplitude rectangular
> > one (multiple rectangular steps with different amplitude and constant
> > step size ?) or other sort of signal with constant amplitude like PWM
> > ?
>
> It looks like this on a scope:
> +    ___
>     |   |
> 0 __|   |__     __
>            |   |
> -          |___|


This ASCII looks like hell, but I think I've understood, you have
three different voltage levels counting your 0 level too.


> I think I've worked out the timing - the 0V segments last for t/6, the +
> and
> - pulses last for t/3. "t" taken to mean "time for one cycle"
>
> What I'm trying to work out is a way to find the resonant point of the
> transformer.

 You are on the right direction.


The usual method is to feed a sine/square wave to the primary

Unfortunately a square signal on input will generate a lot of noise on
the output. The noise is proportional with the signal speed variation
(the time of the rising and falling edges) Sometime (usually) there
are also multi-resonant spikes on both rising and falling edges of the
input square signal.


> and watch the secondary on an oscilloscope. The problem being I have a
> scope
> but no signal generator. I was thinking something along the lines of a PIC
> with internal ADC, a diode/capacitor filter, then some code to sweep the
> entire frequency range and store the timing figures that produce the
> highest
> voltage. Which means I also need to find some resistors to build a voltage
> divider for the A/D input, and maybe a voltage reference chip. This is
> going
> to take some planning.
>
> I can get 7Vish out of the transformer with normal square wave driving, but
> the transformer is as noisy as hell when I do that. Using a modified square
> wave gets the noise down, but the voltage goes through the floor.
>
> The transformer primary is 270 turns of 28SWG Enamelled Copper wire on an
> RM
> core, secondary is 150 turns on the same RM core. Design frequency was
> 50Hz,
> though that produced a rather lame 2V output. It should - in theory - work
> up
> to 500mA. Of course, theory almost never matches real life.
>
> What I'm trying to do is convert 12V DC into 6.3V RMS AC

 Phil, the simplest way for getting RMS is to use a simple square
driving signal and to found the secondary transformer resonance.
You'll get a perfect sinusoidal signal ( would be harmonics also but
much smaller ).


or 6.3V DC

OF COURSE, this is much simple: Square input, square output, fast
rectifier (two diodes) a capacitive filter and an inductive dual
filter (C+differential inductance+C)
This will kill most of the switching spikes to a low level ripple bellow 10 mV.

, with a
> 10% tolerance (about 5.6 to 7V) for a CRT heater. One end of the heater is
> connected to the cathode, which floats at around -600V (IIRC). Much fun.

Absolutely no problem, use a rectangular switching transformer on ferite core..
Do you remember the Faraday law isn't it ? You don't need more.

If you need smalles noise (I have doubts, usually 100mV ripple on
heater is excellent) then you need a voltage/current slew rate
controlled driver. Take a look for example at LT1533/LT1534 family.


cheers,
Vasile

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist