Buck converter and negative output not possible?

[Alti]

The topic is about a non-isolated switched mode converter with negative output.
To be more specific, +400VDC supply that outputs -5VDC, about 1W.

Whenever I needed a small negative output in projects, buck was the obvious choice.
What surprised me is this application note from ST [1] and the quote (p.15):

The buck-boost converter is the simplest converter to implement a negative power supply.
It should be noted that only positive power supply can be implemented with a buck converter.

What is the problem with buck? It always worked fine, see [2].
I admit I never did that with +400VDC to -5VDC so this find just puzzles.

I am aware buck-boost has the extra advantage that output can go way higher than input voltage in magnitude but I do not need that, ST AN does not need that. The disadvantage of buck w.r.t. buck-boost is that the switch needs to work accross 405V instead of 400V for buck-boost but that is just no problem.


References:
1. "Non-insulated SCR / Triac control circuits" AN3168 ST (Doc ID 17193 Rev 1)
2. "Creating a Negative Output Voltage Using a Buck Converter" AN76946 Vishay (25-Feb-16)

[Marco]

[2] is an inverting buck-boost converter.

[Jay_Diddy_B]

Hi,

This is about using a device designed for buck operation in the inverting configuration by grounding the output and let the buck regulator drive the normal ground pin negative.

When you configure the buck regulator like this, the circuit becomes buck-boost.

In a buck regulator:

Duty-cycle = Vout/Vin

In a buck-boost:

Duty-cycle = Vout /Vin + |Vout|

So, reference two is buck-boost.

Regards,
Jay_Diddy_B

[Alti]

When you configure the buck regulator like this, the circuit becomes buck-boost.

That came as a surprise.

Somehow I missed the fact these two topologies differ only in grounding point. 
Thanks!

[Siwastaja]

Yes, this may be surprising, but realizing that buck, boost, and buck-boost are basically all the same circuit is mind-blowing first. Draw the schematic in the synchronous form with two switches, switches on the top of each other (current flowing from top to bottom) and you'll see it. Synchronous buck and synchronous boost are exact same circuits (just input/output labels swapped).

Buck-boost has the "ground" label at a different place, in what would be the buck's output (or boost's input).

Additionally, what would be the input cap of the buck (or output cap of the boost), is now two capacitors in series (input + output) in the buck-boost.

[ocset]

To get -5v,1w from +400vdc.......

1...do a flyback

...it has a low side fet drive, which makes things easy.

Either wind your own transformer or pick something off the shelf from kaschke or wurth , say
....................
or you could buck down to say 20V (from 400v)  using a DCM Buck, using a  pulse transformer fet drive fort the high side fet...then use an off the shelf buckboost chip to get you from 20v to -5v

[Alti]

To get -5v,1w from +400vdc....

The buck-boost with Viper013 is literally 12 components + bunch of 0603 passives. Flyback would have been my next choice but for >1W.

[Circlotron]

Synchronous buck and synchronous boost are exact same circuits (just input/output labels swapped).

Would make for both easy speed control and adjustable regenerative braking in an electric car that had a DC motor.

[Siwastaja]

Synchronous buck and synchronous boost are exact same circuits (just input/output labels swapped).

Would make for both easy speed control and adjustable regenerative braking in an electric car that had a DC motor.

Yes, in which case the motor replaces the inductor and the output capacitor (i.e., the motor is a voltage source (V ~ rpm) + series inductance). It can be called a synchronous buck, synchronous boost, bidirectional dc/dc converter, or simply a motor drive. The circuit is the same. This is mind-boggling simple once you get it! Although with series wound DC machine, it gets more difficult.

BTW, the same basic principle in still true with a synchronous AC (or "BLDC") motor. More complexity is added on top, but you can mentally separate the commutation sequence from the synchronous bidirectional converter action, which is still exactly the same as with permanent magnet brushed DC motor.