Instability for Current limited voltage mode half bridge SMPS.

[grizedale]

Will a voltage mode half-bridge, with primary over-current limiting, (via current sense transformer in series with primary) suffer a drifting of its primary capacitor's mid-point voltage?.......and a resultant unstable operation as a result?

-peak Current Mode half bridge suffers this instability...voltage mode types are not supposed to suffer it...however, if the load is always switching completely on to completely off...the primary current will tend to keep hitting the primary current limit...and this may make the mid-point voltage of the rail-spilitting capacitors drift towards a rail.?

 

[Orson Cart]

A half bridge with peak cuurent mode will always tend to discharge one of the rail splitting caps - or fully charge a DC blocking cap in series with the transformer - this is why you don't see a DC blocking cap in a full bridge peak current mode - but do see one in a full bridge voltage mode.
Only average current mode works with a half bridge/Dc blocking cap. Regards, Orson Cart.

 

[grizedale]

A DC blocking capacitor is surely always needed to block DC?...i.e prevent flux-walking, whether voltage or current mode.?

In any case, ALL half-bridges and full-bridges need primary overcurrent protection.......or else they are not immune from start-up into a short circuit......................so therefore in cases of overload, or transient...........any half or full bridge, whether peak current or voltage-mode, or average current mode will become a cycle-by-cycle current limited converter, and will start to flux walk...ie put a dc current in its primary.

This has basically concluded that the half-bridge is not a safe topology to use.........unless this balancing winding does the trick.

 

[mtwieg]

A DC blocking capacitor is surely always needed to block DC?...i.e prevent flux-walking, whether voltage or current mode.?

No, this is incorrect. In fact, when using peak current mode control, use of a blocking capacitor is a very bad idea, and will actually increase the tendency for staircase saturation (flux walking). In a peak current mode scheme, there must be a DC path available to the primary current.

This has basically concluded that the half-bridge is not a safe topology to use.........unless this balancing winding does the trick.

You need some method of flux centering, correct.

 

[grizedale]

OK, but i think it is first necessary to establish that all half-bridge and full bridges are current mode controlled...ultimately.

Because if you overload them, or short their output , they simply act as cycle-by-cycle current limited converters, as they switch up to the primary current limit every cycle.

So your least statement that a DC blocking cap is not wanted in current mode half or full bridges is as good as saying that no-one should ever use a dc blocking cap in the primary of ANY bridge converter.

...because , as i said, when overloaded, even voltage mode converters will behave as cycle-by-cycle current mode converters.

..which would beg the question...just why do people use dc blocking capacitors in bridge converters?

 

[mtwieg]

OK, but i think it is first necessary to establish that all half-bridge and full bridges are current mode controlled...ultimately.

Because if you overload them, or short their output , they simply act as cycle-by-cycle current limited converters, as they switch up to the primary current limit every cycle.

So your least statement that a DC blocking cap is not wanted in current mode half or full bridges is as good as saying that no-one should ever use a dc blocking cap in the primary of ANY bridge converter.

...because , as i said, when overloaded, even voltage mode converters will behave as cycle-by-cycle current mode converters.

..which would beg the question...just why do people use dc blocking capacitors in bridge converters?

There's a big difference between having peak current limiting on the primary and having actual current mode control. When operating in voltage mode control, DC blocking is useful since it prevents staircase saturation due to asymmetrical volt-time products from the inverter. If a converter has peak input current limiting, but it's still operating in voltage mode control, then this benefit still holds. If that converter were to have a shorted secondary and the current limiting kicked in, then it might be operating as a current mode supply, and yes staircase saturation may occur during that time, but that's pretty much a moot point because it's in a fault condition anyways. The core saturation doesn't really matter at that point, I think.

 

[grizedale]

The thing is, when an overload condition exists..not a short but an overload...this especially happens with guitar amplifiers.

When an overload exists, the volt-mode half bridge is in cycle-by-cycle current limit...and as we said......its centre point goes drifting to a rail.

 

[Orson Cart]

A small low cost CT on the pri side can be used to do average current limiting on the pri switches - for a volt mode half bridge - there is no flux runaway or any problems starting into a short with this method - also a CT on the output can be used to the same effect and gives an isolated signal back to the AC side for current limiting (average, not peak).

 

[grizedale]

though i think we must remember that when in short or overload ...then "voltage mode" is not "voltage mode " any more....its "cycle-by-cycle current limiting mode"


...and so off its centre point goes to the rail.....and bang

 

[Orson Cart]

Not if you use an averged current signal (from Pri or sec) to reduce the PWM

 

[grizedale]

ok, but we can rule out the secondary side straight away.....because in the event of start-up into a heavy overload, nothing on the secondary side will ever get powered up.

So for this "average current mode control" we are talking about primary side circuitry....and to achieve that kind of function, we are talking lots of money....is it worth it?

And how does "average" current mode control stop sudden high di/dt overcurrents?...for that , surely we need "immediate ", supe0fast acting overcurrent protection , and not "average" current mode protection?

 

[Orson Cart]

The CT for average current measurement can be on pri or sec (sec better) it does not need power, small CT's are pretty cheap in volume, the ouput choke limits the rate of rise of the current in the mosfets - true average current mode is suffciently fast enough to protect the mosfets in the event of a shorted output (even starting from full power) - Regards Orson Cart.

 

[grizedale]

Hi,

Thanks
Does "average current mode" simply mean that the current sense measurement is heavily filtered....or does this mean an expensive pwm controller with av current mode facility?

 

[Orson Cart]

See

https://www.ti.com/lit/an/slua079/slua079.pdf

for average current mode