Toroidal 60-Hz transformer

[vtpeaknik]

I was looking for a transformer to convert my inverter's 115VAC output to 230VAC for some purposes, e.g., to potentially run a small heat pump.  Bought one via fleabay that's nominally rated 2000 watts (or volt-amps, rather).  It's a chinese-made toroidal model.  I can't imagine how they manufacture these things by machine?  Supposedly it's more efficient than a traditional style transformer?  It has significant heft, but the wiring inside the box is AWG18, which seems too small for 2000 watts.  Nevertheless, I ran a 500-watt resistive load on it for 10 minutes and it hardly warmed up enough for my hand to notice.  (Disconnectd power before touching! :-)  So perhaps it'll be ok for around 1000 watts of a real load, which is enough for a small heat pump.  Would be nice to get 3000 watts worth of cooling out of 1000 watts of PV input on a hot summer afternoon.

[joestue]

Toroidal transformers are more efficient because the steel is grain aligned and they can run them at 1.6T or more (they do make iron that can handle 1.95T) where as its hard to run more than 1.3T through a E-I core.

A transformer's copper losses follow the inverse of flux density squared. meaning you get 2.1 times as much power through a 1.9T transformer as you can through a 1.3T transformer... at the expense of more core loss.

It is hardly ever cost effective for small transformers to operate at core loss= copper loss.
the reason why is because a transformer's power density is non linear. a 50 megawatt transformer might have a power density of 1 kilowatt per kilogram, at 98-99+% efficiency

here's a datasheet of a 1 KVA transformer.
http://www.toroid.com/standard_transformers/isolation_transformers/DataSheets/798.1202.pdf

the core loss is 5 watts. optimum efficiency would be perhaps a 316 watt load, at which there would be 5 watts of copper loss, for 99% efficiency. power density would be 41 watts per kilogram.

they rate it at 1000 watts, at which there would be 50 watts of copper loss if it is indeed only 5 watts at 316 watt load.
--95% efficiency.

in practice, with substantially increased airflow you could increase the load to 1.41 KVA and there would be 100 watts of loss and you'd have 93.3% efficiency.

btw, i believe that company specifies their transformer's rated load when its mounted horizontally on top of a table with a rubber cushioned mounting bolt running through the middle.

supporting it vertically would increase airflow substantially and it can indeed be practical to run them at double their "Rated" load.

[vtpeaknik]

Thanks for the knowledgable info, Joe.  The transformer I have is enclosed in a rectangular steel case, so no air flow at all, unless there are some small holes.  Would it be useful to make more holes, and mount a small fan onto it to force an air flow through the case?  If it were to dissipate 50 watts internally it'd get pretty hot.  At my test run of 500 watts it may have been similar to your 316W example, running near peak efficiency, dissipating about 10 watts.  If even that - it got barely noticably warm to the touch, after 10 minutes of operation - but that was with the case opened up (to hook up my test wires).

[hydrosun]

There are many ductless heat pumps that work on 115 volts. I have a Sanyo inverter(variable speed) and a Bonaire one speed heat pump that work of a single outback inverter. There are other brands available but you really need to do your homework to find quality. I thought I would have to use a transformer on sone of the highest rated heat pumps like the mitsubishi brands but figured the use of a transformer would negate any additional efficiency.
Chris

[vtpeaknik]

The Sanyo was a nice model but they don't make them any more, since Panasonic bought them out.  The heat pump I have in mind (Fujitsu 9RLS2) is so much more efficient than any 115V model that I know of, that a 5% or so loss in the transformer is still very much worth it.  Moreover, when I run it on grid power, it'll be direct, without the transformer.  I'd only use the transformer when I have an opportunity to use surplus solar power, and that power (when available) is free, even if a few percents end up just heating the transformer (and 85% of the incident sunlight just heats up the PV panels...).

I am still mystified as to how a toroidal transformer can be machine-manufactured?

[joestue]

check youtube, basically there's a circular bobbin that is run through the center and then loaded up with wire, that's why they aren't very "full" of wire.

[Flux]

http://www.talkingelectronics.com/projects/Inductor/Inductor-1.html

This link shows one form of machine.  In all cases the winding bobbin is split or has a gap to get it through the core.

The machines I met needed the wire to be wound on the bobbin when threaded through the core, then the end of the wire was anchored and as the bobbin rotated it wrapped the wire round the core. The bobbin was supported and driven by a rubber belt passing over a series of pulleys to centre it.

The machine in the illustration looks to use the bobbin as a capstan and the wire may be fed from a spool without the pre winding stage.

Probably find better descriptions on Youtube, I haven't looked.

Flux

[vtpeaknik]

Boggles the mind!

A different question about this transformer: it seems (from my memory from briefly looking inside) that the output (230V) socket is wired so that it's "cold" side is connected to the input (115V) "cold" (neutral) side.  This is a European type socket.  But, for my purpose, of running a 230V American applicance, this may not be the best arrangment?  A standard 230V American socket is 2 "hots", with or without a neutral wire present, but each "hot" is 115V relative to the neutral, in opposite phases.  Using this transformer as it is will put one side at 230V relative to the neutral (and ground), rather than the intended 115V, placing an unnecessary strain/risk on the insulation throughout the appliance.  Therefore, assuming the thing is an autotransformer (one coil with a center tap), I'm inclined to swap the neutral and hot input wires, which would place the center tap at neutral, one end of the coil at the input 115V potential, and the other end also at 115V but in the opposite phase, just as a standard American 230V socket is supposed to be.  Does that make sense?

[Flux]

Yes, the European system is totally different. If it is an auto transformer you can interchange the neutral to centre as you suggest.

If it is double wound you won't have a centre tap but you can still remove the link between input and output. Until you check for continuity between windings it is hard to tell but from the cost point of view it is most likely auto.

We do have totally isolated transformers to step 230 down to 110 centre tapped for power tools but these are not normally toroidal and this wouldn't fit that requirement for several reasons.

Flux