Flyback converter / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Flyback converter

1 2 next

Move Thread

LAN_403

Dr. Shark

Fri Apr 14 2006, 09:02AM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Since a lot of people sucessfully presented boost converters here in the past, I though I'd go a step further and play with a flyback - but in the original sence of the word, i.e. a forward convterter, not a TV high voltage transformer.

The input voltage is only 12V, but by driving in discontinous mode, I get the primary voltage to kick up to 400V

The turns ratio is only 1:10 on the inductor, so the output is in the low kV - still enough for a small spark though:

The only thing that I dont understand: I cannot charge up the capacitor in the above picture, and no, I did not forget to put in a diode.

OK, this is not a great project in any way, but I think that it is a great remider on how flyback transformers are supposed to work, and how you can really tayor the peak voltage across the primary by varying the on time of the switch. So selecting the correct frequency to drive a flyback is not about some mysterious "resonant frequency", or about being close to the 17kHz "it is meant to be driven at", but it is really about scoping the drain or inductor voltage and making the on time as long as possible without overvolting the switch.
I hope this is helpful to some people. At lest I have never though of a flyback transformer as a boost converter, but now I think it is really helpful if you do.

ragnar

Fri Apr 14 2006, 09:26AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

love it, joe!

I'm glad I'm not the only person tinkering over easter =)

Dr. Dark Current

Fri Apr 14 2006, 10:16AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

wow.. how cute

I see that the transformer has only ferrite rod(???) as a core - how is it supposed to run at 17khz???

Marko

Fri Apr 14 2006, 12:36PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

hi joe,
I did something similar with my flyback converter-flyback driver.
I wraped few turns around core of noise filter choke and seriesed two windings.
Turn ratio was low and spark started at less than 1 mm but I could draw it more than one cm.
I played with few small transformers, some gave really nice sparks and even streamers (this one didnt live for long after that).

For your converter you should use fully closed core and add an air gap if needed.

Also, you didnt show schematic nor some explanation of the driver, as I could figure out there is only one transistor and ne555 on the board

Steve Ward

Fri Apr 14 2006, 06:19PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

but in the original sence of the word, i.e. a forward convterter, not a TV high voltage transformer.

Actually, you mean a flyback converter (not forward)... similar scheme, but different in operation.

Dr. Shark

Fri Apr 14 2006, 07:10PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

OK, here comes the info I left out when I posted this:

>> ferrite rod(???) as a core - how is it supposed to run at 17khz???

Jup, its just a rod, and it runs at a much higher frequency. 17kHz is what TV Flybacks run at.

>>For your converter you should use fully closed core and add an air gap if needed.

Could you explain why? I think the coupling is pretty good as is (the primary is under the secondary), and I am pretty happy with the low inductance values. Well, I get a lot of heating in the copper, so I guess you've got a point.

>>Also, you didnt show schematic nor some explanation of the driver, as I could figure out there is only one transistor and ne555 on the board

It is just that: The 555 is driving a small IRG IGBT directly. This seems to work pretty well, althouth the IGBT gets warm, too.

>>Actually, you mean a flyback converter (not forward)... similar scheme, but different in operation.

I think a flyback converter is actually a kind of forward converter, isn't it? BTW, glad to hear from you, try spending less time for university, we need you here (I am planning to construct a DRSSTC this weekend, so at least I will need you

)

Self Defenestrate

Fri Apr 14 2006, 10:25PM

Registered Member #87 Joined: Thu Feb 09 2006, 01:36PM
Location: San Jose
Posts: 191

joe wrote ...

The only thing that I dont understand: I cannot charge up the capacitor in the above picture, and no, I did not forget to put in a diode.

Try shorting the leads of the capacitor while its running. With such low voltage/current and high frequency, I'd build a cascade for that. Couple 1n4007's and a few ceramics later, you'll have a pretty nice supply.

Sulaiman

Sat Apr 15 2006, 07:28AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

The peak voltage of the primary flyback 'spike'
is not directly related to ON time of the switch (IGBT).
It is related to
1) any clamping or snubbing on the Primary side (e.g. IGBT avalanche voltage)
2) peak Secondary voltage divided by transformer ratio
3) leakage-inductance 'spike'

The reason you're having trouble charging the capacitor is
an initially discharged capacitor 'looks' like a short-circuit on the secondary
so the energy stored in the transformer cannot dissipate before the next cycle
so the current in the primary increases with each cycle
and the primary saturates.
This will cause switch (IGBT) heating and/or death.

A ferrite rod makes a good core but;
There will be a lot of radiated emi (who cares for hobby use)
and unless you are careful with construction there will be quite a bit of leakage inductance
which wastes energy by heating up the switch (or clamp or snubber)
(wind the primary the full length of the rod and the secondary in the middle half of the rod)

Dr. Shark

Sat Apr 15 2006, 09:42AM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Thanks for the input Sulaiman, how is it that you seem to know quite a bit about magnetics that I dont? Hmm...

I think I have seen a very complicated formula (was it here

I dont know) relating all sorts of parameters to the peak voltage, and adding a secondary winding surely does not make matters less complicated. I think trial and error is good enough for me here.

Your argument about the capacitor sounds good to me, that would also explain the heating in the copper, where most of my power seems to go.

Why is it there is so much discussion about this, I just build this as a diversion while going through the pains of building a DRSSTC.

DrZoidberg

Sat Apr 15 2006, 03:39PM

Registered Member #350 Joined: Mon Mar 27 2006, 05:14PM
Location:
Posts: 106

What Sulaiman said is right of course but the capacitor should get charged anyway. The charging is just inefficient.
You got sparks between the leg of the cap and the transformer. That means it must be charging or the diode is broken. The voltage might have been too high for the diode.

But that gives me an idea how to charge a cap with such a transformer more efficiently and avoid a saturation of the core.
Use a 555 in monostable mode instead of astable. i.e. as soon as the voltage at pin 2 changes from high to low (falling edge) the output will become high for a set time.
Then you use that to turn the primary current off whenever it becomes too high.
Lets say you want the off time to be 10us then you set the 555 to output a 10us pulse. Then you figure out at which current the core becomes saturated. Lets assume it's 10A. Then you connect a resistor of 0.07 ohm in series with the primary winding. At 10A the voltage drop will be 0.7V. That voltage is applied to the base of an npn transistor. The collector is connected to pin 2 of the 555. Pin 2 is also connected to +Vcc over a resistor.
That should then always give you the optimum turn on time.

1 2 next

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