SMPS: Flyback Vs. Tapped inductor.

[Blueteeth]

Hi,

Bit of a random question for anyone who is up on switched mode power supply topologies.

I'm designing a high voltage power supply to run from 9-12V. Ouput should be 200-350V at low current. Total power input sohuldn't reach more than 15W. There are two different apps, one is for capacitor charging (high power Xenon strobe) and the other is for a portable valve preamp.

Now, given the two different apps I realise that noise will be a problem on the valve preamp, so I'm going for a resonant royer design. Its output would be fairly well regulated and given the almost sinusoidal waveform, noise and EMI sohuld be low.

But for the xenon strobe, the output is effectively going from 0-300 as the cap charges. The most common type of inverter for this it seems is a flyback configuration. But I have seen some documentation on the use of 'tapped inductor' designs. Or 'autotransformer'. Which are essentially boost converters that allow a greater voltage output for a given input.

Does anyone have experience with designing flybacks or tapped inductor converters? Because I'm a perfectionist I'm after high efficiency (>80%) but trying to avoid specialist IC's such as those from Linear Technology and Maxim.

Most xenon strobe inverters on the web either run from 1.5V and are self resonant, or direct from mains. I've yet to see a 'half decent' 12V medium power inverter that achieves good efficiency. I'm rolling my own transformers/inductors and have plenty of MOSFET's to play with. just after some advice/experience.

Thanks.


Blueteeth

 

[glutnix_neo]

I'm not familiar with the tapped inductor topology, are you referring to boost topology with secondary winding on the inductor?

It is difficult to get >80% efficiency with flyback, I doubt if you can on tapped-inductor.

One thing I can advice is you should use a mosfet with a very high voltage rating, voltage spikes when it is off could kill it. You can consider putting some snubbers (but again this would dissipate energy)

 

[ronsimpson]

Blueteeth
Your question "Does anyone have experience with designing flybacks .."? Answer "Yes"
If the input is 1.5 volts your efficiency will be low!
If the input to output ratio is large your efficiency will be low.
Most xenon strobe inverters use tapped inductors. You can think of a tapped inductor as a transformer.

 

[Blueteeth]

Hi Glutnix, and Ron

Yes, a tapped inductor is really just two inductors in series, effectively having a 'tap' somewhere in the middle of the inductor with a ratio of turns. It's also refered to as a 'coupled inductor'. Essentially an autotransformer. Perhaps I'm off but I see it as a flyback with the hot side of the secondary connected to the cold side of the primary. Such as in this link:

**broken link removed**
EDN PDF
**broken link removed**
A standard boost converter is really not up to the task of >150V from a 12v input, as glutmix pointed out, the MOSFET not only needs a high voltage rating for flybacks but for boost converters with a 200v output, one would need a good 600V MOSFET, which increases on resistance, hence, lowers efficiency.

Ron, I don't want to power this with 1.5V, m ypoint was many schematic on the internet use designs from dispoaible camera which use single AA's. I'm after much more power, and greater efficiency. Input should be 12V, with perhaps a maximum of 1A continous. I know strobes use autotransformers for the trigger, but rarely for the actual step up converter.

Really, I'm completely confused and somewhat indecisive as to the best topology for this sort of system: Charging 300V caps. Since the output is a current source, not regulated. So continuous frequency PWM would probably be a bit inefficient.

Flybacks have their place, and I can wind my own, controlling the airgap, inductance and turns ratio (perhaps using peak current limiting) but tapped inductors, although pretty much modified boost converters, appear to ease some of the high voltage burden on the mosfet switch.

Perhaps I shall just need to wind some transformers and do some measurements As I often over think things in my quest for perfection and easily get swapped in idea's and possibilities.

 

[ronsimpson]

I would use a current mode PWM like the UC3842 + FET. Set the current limit resistor so the input current is limited at the 1 amp you want. Use voltage feedback so the voltage cannot get above the 300 volts you want.

Because you have not done this before: I would start out and build a 12 to 50 volt boost up circuit that handles 12 watts. (100 volt FET) then when it is working well, add turns to your coil to auto-transform up to 300 volts.

The 50 volts will not hurt you much but the 300 volts will kill you.