On Sat, 15 Feb 2003, Olin Lathrop wrote:

*>> The 'easy' way to make a sine converter is to use a class B push-pull
*>> amplifier and a split secondary transformer (2x10V for 12V batteries).
*>
*>Not so easy when you realize that class B is not very efficient, and at
*>his power levels getting rid of the heat will be a significant challenge.

Class B is the most efficienct way to drive a sine wave with reasonable
distortion and without complications (like high power harmonic filters
whose Q must be kept low and dealing with load dumps with several Henrys
of secondary inductance at tens of A/us slew rates produced by
non-sinusoidal drive). The efficiency is ~70% theoretical and the same as
that of an equivalent powered audio amplifier, and it is feasible in
low-tech low-development-expense mode upto 1kW or so.  Also most older AC
inverters work exactly like this.

*>Commercial inverters either alternately drive the transformer primary to
*>full in each direction producing a sort of rounded square wave, or they

That is not a rounded square wave, it is a square wave. What they do is
drive the primary with fast PWM modulated to make a sine shape after
integration by the big transformer and associated filters.

*>use DC-DC converter techniques to drive the primary with a high frequency
*>chopped sine wave as you said.  The rounded square wave approach may not
*>be as bad as it first appears, depending on the loads you plan on driving.

What in g*d s name is a rounded square wave and how do you make it using
straight switches ? The only way I know if a multiply tapped transformer.
It is very expensive to make. That would approximate a sine in steps.

*>Incandescent light bulbs and electronic equipment with switching power
*>supplies (like a PC) don't mind this at all.  However, sine wave output is
*>the only thing guaranteed to work with all standard appliances.
*>
*>> All
*>> the other ways involve a DC/DC converter steered with |sin| waveform and
*>> an active H bridge in the high voltage side.
*>
*>Really, "all" other configurations than class B are like this!?  C'mon
*>now.  For one thing, you can use the same center tapped primary you
*>mentioned above, center to the +DC supply, and each end switched to ground
*>using PWM so as to produce a sine wave out.  In fact, it was my
*>understanding that this was a reasonably common arrangement.

When PWM is used normally there is no tap, and either a half bridge or a
full bridge drive is used. Switches are cheaper than transformers. There
are some designs with tapped transformers but the core space is used
inefficiently like that.

*>> This will give about 70% efficiency, which is low.
*>
*>A well tuned and tweaked system should be able to do better than 70%, but
*>that is still a lot better then class B.

Class B is rated 70% efficient (at 5% or 10% distortion - I do not
remember).

*>> Maybe it would be easier to convert the 115Vac
*>> installation to 24Vdc or 48Vdc and save one major part (or pita). What
*>> would it take to convert the blower to a static switched dc motor ?
*>These
*>> are made to specs like you need for servo purposes.
*>
*>That may be a reasonable approach in some cases, but I think the original
*>poster is doing this partly for the sake of doing and learning this
*>himself.  Certainly a system that produced normal line power would be much
*>more flexible, and a much better learning experience.

The op is trying to buy insurance against ice storms probably. Imho for
his needs the best way would be to get a genny that runs on the same fuel
as the furnace and concentrate on making an automatic fail-over circuit
(with temporary power dropout while switching). At least that would be
what I would do. The kind of failover without power drop in the middle is
very hard to do for mains.

Peter

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu