[Prev][Next][Index][Thread]

Re: Self-built power transformer

HI Doug,

> Original Poster: Doug Brunner <dabrunner-at-earthlink-dot-net> 
> 
> First off, I'd like to thank everybody for helping me with this transformer.
> 
> Questions: what do you mean by end-to-end airgaps? What should the core be
> laminated with--will poly work? I just covered it with that today, 
to prevent > arcovers.
 
I mean that the core is not a closed magnetic circuit. There is at 
least one arigap in it.

> Also, how many turns would you advise I use? I've found some unused
> transformers
> to rip apart, so I have plenty of wire, as well as lots of poly to
insulate it
> with.

You have to know the maximum flux density your core can withstand
before you can determine the number of turns to use. In fact it's not 
even that simple. As flux excursion increases, so does core loss. 
You really need to measure and/or plot the B-H curve for your core to
determine a level of ampere-turns to apply to it for an acceptable 
loss. Sorry, transformer design is not just a matter of winding turns, 
nor is it a black art. People familiar with audio amplifiers often 
claim RF design is a black art. Same thing applies. They simply 
ignore parasitic L's and C's that don't make any difference at AF and 
then discover they do matter at RF. It is a matter of taking all 
parameters into account. You have to think it out to the nth degree.
     I've learnt most of what I know about transformer design from
designing switchmode power supplies which are simply scaled up in 
frequency. Most issues apply at 50Hz as far as the transformer goes
with the exception of conductor and core behaviour at 100's of kHz.
I am currently designing a 2kW+ offline flyback switcher for TC use 
and a switchmode power factor correction switcher to precede it. 
Object of this exercise is get accurate figures for TC output vs 
accurately determined primary power at the best possible wallplug 
efficiency.

Malcolm

> Last, an explanation for my signature--I spend a lot of time playing
> Unreal, so
> it's convenient to have my netname on my signature.
> 
> Tesla List wrote:
> 
> > Original Poster: "Malcolm Watts" <MALCOLM-at-directorate.wnp.ac.nz>
> >
> > Hi Doug,
> >           Some important things about your proposal:
> >
> > > Original Poster: Doug Brunner <dabrunner-at-earthlink-dot-net>
> > >
> > > What I meant is, it's wound on a solid iron core and placed in a pipe to
> > > contain
> > > the oil and provide magnetic coupling.
> >
> > First off, eddy current losses will be high as the core is one big
> > shorted turn. So is the pipe you are going to surround it with. It
> > should be laminated. It's also a rather lossy shorted turn as iron is
> > not a good conductor so expect some heating. Next, plain old iron
> > will exhibit rather high hysteresis losses as it is easily magnetized
> > and tends to stay that way. Thirdly, coupling will be very poor as
> > there are huge end-end airgaps. Expect a low induction factor (= lots
> > of turns required for primary and secondary windings to avoid
> > saturation). It will also exhibit a lot of leakage inductance. I
> > agree with Erik. I don't think this is a good proposition.
> >
> > Malcolm
> >
> Snip.........
> --
>            --Mr. Postman (Doug Brunner)
>                 <dabrunner-at-earthlink-dot-net>
> 
> 
> 
>