Dividing down the mains efficiently

[treez]

Supposing you want to supply a 3kW resistive load at any power between 0 to 3kW
You need to power it from the 240VAC mains ultimately.
You want very good efficiency, so diode bridge followed by Class D amplifier style PWM'ing is going to be a bit lossy in the upstream diode bridge.
You need it to be PFC compliant, and of a fairly small size generally.
Also, as you know filtering will be needed if HF PWM used, which is likely, due to size constraints.
No mains 20ms cycle skipping as the flicker is deemed unacceptable.
You cant put DC into the resistive load....(unless its pulsating down to zero amps every 10ms...for obvious reasons regarding contactor arcing etc.)
No need for isolation whatsoever.

What's your "fave" method of doing this?

I was initially thinking totem pole (bridgeless) PFC stage but it would mean DC into the load...or would it....

 

[KlausST]

Hi,

Just a PFC stage with variable output could do the job.
Or a autotransformer.

Klaus

 

[treez]

Thanks but autotransormer too big. (its 5kW)
The PFC stage is interesting....but if it was boost pfc then its boosting up the voltage and we dont want that.
Also, if the output of the Boost pfc is DC, then its no good for us. (however, undulating dc, which goes to zero current/voltage every 10ms, is fine)

So far, the "standard" and well trodden way to do this is a mains diode bridge followed by a Class D amplifier style full bridge at say 30khz........the "bridge" containing an inductor in series with the resistance, and a cap in parallel to the resistance.

When "dimming" You switch each diagonal pair for 10ms at a time...........and when/if the diagonal pair is switched off, you let the current freewheel in the two top fets... when you want full power you just leave a diagonal pair ON for 10ms, then off for 10ms, then on for 10ms...etc etc...for both diagonal pairs.
This is the standard "noddy" way to do this.....whats the "quality" way to do it.?

 

[Easy peasy]

For a purely resistive load, an AC buck converter would work just fine - meets all your specs and is very small and efficient

 

[treez]

Thanks, i am trying to do the AC buck converter differently from this...

Fig. 1: AC to AC Buck converter.

Download scientific diagram | AC to AC Buck converter. from publication: AC VOLTAGE REGULATION BY SWITCH MODE BUCK-BOOST VOLTAGE CONTROLLER | Voltage sag is an important power quality problem, which may affect domestic, industrial and commercial customers. Voltage sags may either be decrease...

www.researchgate.net

...to cut down the number of diodes in the power current path

--- Updated Jun 7, 2021 ---

----------------------..................................--------------------------------------...............................-----------------
Attached is the "noddy" way to do it.
Any improvements on this?
LTspice and pdf schem attached

 

[Easy peasy]

it's better without the bridge - if you have a volume application we can supply a pretty good circuit for AC / AC buck ...

--- Updated Jun 8, 2021 ---

By the way - the link you posted is not a buck

 

[treez]

Thanks, i take it you mean literally a buck, and with a back to back fet upstream to switch it out when its input goes negative for 10ms?

 

[Easy peasy]

literally a buck that is happy to have ac in and supply the same freq ac out but at a controlled level from zero up to Vin

 

[BradtheRad]

Coupled inductor reduces AC voltage while increasing AC current.
This concept might be practical if there were an easy method to vary the inductor value (possibly with a movable slug). Or as an alternative, switch in various LC combinations.
It's not the same as an autotransformer. Its Henry value can be smaller although the component probably is not compact.
It creates power factor error which means a capacitor constructed to carry several Amperes.
This simulation is a single example suited to a particular load.

 

[Easy peasy]

it's a low freq buck, neat, but quite bulky at 3kW, not too good with a load short

 

[treez]

Hi
Thanks for these ideas.
Also, there is the case of using the cct from post #5 above (as attached LTspice sim and pdf), and just shorting out the bridge rectifiers as follows...

 

[Easy peasy]

if your load is ok with a net dc component, then yes, that BR followed by buck would be ok.

 

[Easy peasy]

Let us know how it goes - driving the high side switch from 0- 100% pwm can be a small issue - also protection from mains spikes - pk current mode a good idea - but you could get away with fixed pwm too - as it is very nearly a linear system ( with a pk curr mode limit ). 1200V 65 m-ohm SiC fets a good idea.