drive flyback transformer with a sine wave? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

drive flyback transformer with a sine wave?

previous 1 2 3 next

Move Thread

LAN_403

Mattski

Mon Jun 06 2011, 09:12AM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

There are two loss mechanisms in switching transistors: switching loss and conduction loss.

IGBT's will generally have higher switching loss, but at very high currents they will generally have less conduction loss. At low frequencies switching loss might be small, since it's basically a fixed penalty which is paid every time the device switches on or off. So a lower frequency high current converter could easily be more efficient with an IGBT. That's why you can find IGBT's rated to over 1kA of current, but not MOSFETs (to my knowledge), because the MOSFETs are not as good at that current level.

Ash Small

Mon Jun 06 2011, 10:04AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Yep, that covers transistor losses, but in a recent thread people advised that the most important factor for flyback topology was fast switch off time. It 'appears' that you are advising otherwise. I'm just trying to clarify a point.

Mattski

Mon Jun 06 2011, 05:01PM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

Ah, well fast switch time is always desirable for high efficiency, but to my knowledge it does not significantly perturb the circuit operation.

Ash Small

Mon Jun 06 2011, 05:35PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Thanks for your replies Mattski.

One more point, though. (at least, I think it's the last one!)

You are saying a fast switch-off time simply reduces switching losses.

Won't a faster switch-off time enable a faster rate of change in magnetic flux? (ie the flux can't drop to zero while current is still flowing, surely?)

Mattski wrote ...

.Better to think of it as rate of change in magnetic flux over time, since the current under consideration can come from the primary OR secondary winding, but it's one flux that they share in the ferromagnetic core.

.

Proud Mary

Mon Jun 06 2011, 07:25PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I'm sure a good deal of confusion arises in these discussions when people refer to a TV line output transformer as 'a flyback' even when the transformer is not being used in flyback mode.

Ash Small

Mon Jun 06 2011, 08:04PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Proud Mary wrote ...

I'm sure a good deal of confusion arises in these discussions when people refer to a TV line output transformer as 'a flyback' even when the transformer is not being used in flyback mode.

You're not 'Yank bashing' are you, PM?

I agree the 'English/British' 'TV LOPT' causes a lot less confusion than the 'American' ' TV flyback'.

Mattski

Mon Jun 06 2011, 08:22PM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

wrote ...
You are saying a fast switch-off time simply reduces switching losses.

Won't a faster switch-off time enable a faster rate of change in magnetic flux? (ie the flux can't drop to zero while current is still flowing, surely?)

That might be true if there was only one place for the magnetic energy to go, but the flux can be maintained continuous by the secondary winding. A quick simulation shows this: in the first case a FET is switched with a low impedance gate driver (Rgate=5), and again with a high impedance one (Rgate=50) which will take longer to switch the transistor off. In the first case the behavior is basically ideal, the flux is proportional to the sum of the currents times their relative turns ratios which is the top plot in each case, and in both cases it is continuous at turn-off. I(Lp) is primary current which ramps up linearly as expected for square wave voltage, I(Ls) is secondary current which decays exponentially as typical of a inductor-resistor circuit. And when the FET turns off the secondary current starts flowing at it starts out twice as large as the peak primary current because in this case it has half as many turns so needs twice the current to maintain the same flux.

In the second case where the gate resistor is large the FET takes a while to turn off, during which time a reduced voltage continues to be applied to the primary winding. Because of the slow switching time the primary current decays while the secondary current ramps up, and the peak flux is slightly higher since the FET stays on for slightly longer, but the current out of the secondary winding is practically unchanged, and for a resistive load the output voltage is almost identical.

Edit:

wrote ...
I'm sure a good deal of confusion arises in these discussions when people refer to a TV line output transformer as 'a flyback' even when the transformer is not being used in flyback mode.

Quite right Mary, I think that was certainly a part of confusion here when we use two words for the different things :)

Ash Small

Mon Jun 06 2011, 09:02PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Those simulations are quite interesting, Mattski, especially the slope on the leading edge of the voltage spike where R=50.

Thanks.

Mattski

Mon Jun 06 2011, 09:45PM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

These simulations are done in LTSpice which is my favorite free SPICE type simulator. There is an active Yahoo Groups mailing list for LTSpice which has collected many models and tutorials. The main drawback is that it is not a fully featured schematic capture with integrated BOM and PCB layout program.

wrote ...
Those simulations are quite interesting, Mattski, especially the slope on the leading edge of the voltage spike where R=50.

Are you looking at the little red spike in I(Ls) in the bottommost plot at t=0? That's because the flux doesn't quite decay all the way to zero so at turn-on there is an overlap as I(Lp) ramps up while I(Ls) ramps down, same as happened in reverse at turn-off. I should point out that I simulated 100 cycles and only had it save the last one, so t=0 is really about t=5ms, there is stuff happening before this plot.

Ash Small

Mon Jun 06 2011, 10:00PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

I'm looking at the bottom plot between t=13us and t=14us (Scale transposed from second plot)

I downloaded LTspice recently, but couldn't get it to do anything. I'll go back and try again at some point.

previous 1 2 3 next

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