Flyback SMPS OK for 320W? 1

[grigson]

Hello,

We need to design an SMPS where efficiency is not that important.
(its for a guitar amplifier)

Max Load = 320W
Vout = +/-40V at 4A
Switching frequency = 66.7KHz.
Vin = 385VDC.


Can we use a Current Mode Flyback in Continuous Mode?

The maximum duty cycle would be 0.25

Comparison with Forward converter:-
Here are 1 Transistor Flyback (CCM) vs Two Transistor Forward currents for supplying the same load of +/-40V, 4A from Vin = 385VDC:-

1. FET Current RMS = 1A95 for Flyback; 2A for Forward
2. FET current peak = 5A12 for Flyback; 5A9 for Forward.
3. Diode Current Peak = 7A6 for Flyback; 6A4 for Forward
4. Output cap ripple = 2A63 for Flyback; 1A3 for Forward


.....each switched at 66.7 KHz.

These numbers are pretty much the same.......there seems to be no reason why a Flyback cannot be used for the 320W power if a Two Transistor Forward can?

Do you know why i am considered virtually insane for even suggesting a Flyback for 320W?
-i can't see a reason a flyback can't do this

 

[Comcokid]

I'm no expert on flybacks - most of what I do is push-pull for full bridge with lower voltage DC input and not off-line like yours.

The general design info out there indicates that the maximum practical power range of flybacks are 150 to 250 watts. That doesn't mean that you can't do it, but there are probably better solutions.

You've examined the requirements of the semiconductor power devices but may have overlooked the most important power device - the transformer. Flybacks store their energy in the gap of the core unlike most other topologies where the transformer does not store energy but couples it from input to output. The core gap results in a large fringing-flux and with the high power - you're likely to only create an induction heater that will try to melt the windings near the core gap.

Not only that, but it will require a BIG core. A quick check found AN4137 from Fairchild semi. This app note shows a range of suggested cores for flybacks up to 200 watts, and in some of these core families the cores listed at 200W are the largest made in that shape - and you need something with a third more power!

So, before you save on your semiconductors and not needing a output inductor be sure you've not pushed yourself into a very large, expensive, and custom ferrite core wound with 220 C heavy polyimide wire with Kapton insulation.

 

[benta]

Agree with Comcokid, you'll spend more money on copper and ferrite than what you save in semiconductors.

I don't know where you are in the world, but be aware that you might need power factor correction as well at that power level.

Regards,

Benta.

 

[itsmoked]

Benta.. You shouldn't have mentioned that.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MacGyverS2000]

<chuckle>

Dan - Owner

http://www.Hi-TecDesigns.com

 

[benta]

Ahaaah! Someone's got a guilty feeling about PFC?

Benta.

 

[dpc]

Don't go there. Really.

 

[Skogsgurra]

OP grigson has entertained us several times already when exploring ways untrodden by mainstream EEs. This could be another exploration of alternate paths. A 300+ W flyback without PFC.
Those low loss HF cores are costly. Why not try something innovative, simple and not so expensive? Like an ordinary 60 Hz transformer core?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LionelHutz]

Can we use a Current Mode Flyback in Continuous Mode?

Wrong question. The answer to this one is "Yes." even if it's a bad idea.

The proper question is "Should we use a Current Mode Flyback in continuous mode?"

Answer = No.