[Prev][Next][Index][Thread]
Re: Building your own HV transformer
Tesla List wrote:
>
> Original Poster: Sam Laur <slaur-at-sekunda.pp.utu.fi>
>
> > Original Poster: Chris Tominkson <internetinbox-at-yahoo-dot-com>
> >
> > I was looking at this page:
> > http://members.tripod-dot-com/~dbrunner/hvxfmr1.htm
> >
> > That tells you how to makea high voltage transformer, that puts out 24
> > Kilovolts, 1.something Amps. Sounds pretty simple, but I am still
> > confused on the position of the iron rods. It sounds like in this
> > transformer, they are not directly across from each other, but on the
> > same rod, and across from each other?Something weird like that.
>
> He's doing a two-step stepup, first to 2.4 kV and then to 24 kV. But what
> puzzled me was the part list :
>
> # 2 15" long, 1" diameter steel rods (call your local iron store)
>
> Won't solid steel rods cause enormous losses in eddy currents? Well, it
> might be interesting to see (for a while), but the fun won't last for
long...
>
> And another thing is insulation. "10 AWG uninsulated wire" for the primary??
> Ouch... Even though you _can_ varnish the wire so that it feels "insulated",
> a proper wire would be safer and easier. And then you'll be needing some
> insulation between the windings and the core, especially in the second
> transformer.
>
> Anyway, does anyone have some simple (or not so simple) formulas to
> approximate core thickness (laminated silicon steel), primary windings
needed
> to not saturate the core at the wanted VA level, etc. I've found a couple
> places on the net but the results were so wildly in disagreement that I feel
> I need a third (or fourth) opinion. For example, I think I can get a cheap,
> used 3-phase transformer from the junkyard which has about 3" by 3" core
> thickness. How much abuse would this take? And I also already have a smaller
> typical EI-type laminated core, which has a thickness of about 1" by 3".
The proposed design is STUPID and CANNOT work! Waste of time for
several reasons, not the least of which is the use of solid steel cores
"from your local iron store". Who has a local iron store? Second, if I
can read the page right, he proposes to wind the primary out of 1/8"
aircraft cable (steel??) or #10 UNINSULATED wire!!!!!!! Is the core
2.15" long or 215" long? Can't read it. Seems to be an open-core
design, for which the magnetizing current would be large, even without
core saturation. No mention of insulation between turns or anything
like that.
As for formulae to approximate core thickness, here's how it goes.
The basic formula governing things is this:
Vw = 1.718 x 10^-5 * B x Ac x N
where the Vw is the winding voltage in volts rms, the frequency is 60
Hz, B is the core flux density in gauss, Ac is the core area in square
inches, and N is the number of turns. This is the voltage which will be
induced in the core when the flux density in the core is B gauss.
Inverting the expression, you get the number of turns per volt:
(N/Vw) = 58183/(Ac x Bmax)
Where Ac is as above, and Bmax is the maximum flux density allowable to
avoid excessive core loss (core loss gets very large before saturation
is reached). For typical modern transformer steels Bmax is of the order
of 12,000 gauss, and
(N/Vw) = 4.85/Ac turns per volt.
This says that, for a core area of 1 square inch, you need at least 4.85
turns to keep the flux density below 12,000 gauss.
The important governing factors are the core AREA and the number of
turns. The length of the core depends on the wire size you are using
and the number of layers (for the primary, as a practical matter). You
have to pick a wire size big enough to handle the design current, then
allow enough area to wind it on the core. By the way, the core loss is
proportional to the volume of the core, so the bigger the core the
bigger the loss. That's life. For the example above, the core loss
will be about 1 watt/pound, and the magnetizing excitation will be about
6 volt amperes per pound.
You either pick a core area and determine the turns/volt, or pick the
turns/volt and pick the core area, depending on what options you have.
You then have to design the windings using standard wire table values
for wire area, turns/inch, turns/square inch, etc. You than have to
have a large enough core window to hold the turns and the required
insulation between turns. As a practical matter, you may want to shop
around for cores (hopefully with usable primary windings already on
them) and work out the rest. For your two cores, 1" x 3" cross section
and 3" x 3" cross section, the required turns/volt would be (4.85/3) =
1.62 or (4.85/9) = 0.54.
By the way, an old copy of the ARRL RADIO AMATEUR'S HANDBOOK (12th
edition, 1935, page 167, gives examples of transformer cores and
"transformer power". These include:
POWER (watts) PRIMARY WIRE# TURNS/VOLT CORE DIMENSIONS (in)
100 #20 B&S 3.33 1-1/2 x 1-1/2
500 #13 B&S 1.66 2-1/8 x 2-1/8
1000 #9 B&S 0.99 2-3/4 x 2-3/4
These values are more conservative than the ones I gave, and work out to
a flux density of a little under 8,000 gauss. With modern steels you
should do better.
Hope this is of some help. There are immutable laws of physics which
transformers must obey, whether we wish them to or not, and many of the
designs I have seen here violate them. For example, the idea that you
can increase the output voltage of a transformer much beyond that for
which it was designed by upping the primary voltage (I have seen the
suggestion to use 120 volt MOT's on 240 volts!!!) are doomed to failure,
as the cores will saturate and the primary burn up due to very high
current. You need those turns/volt!!!!!
Ed