|
|
re: cheap choke design
6 feb 2002
rich grise wrote:
>well, clothes hangers are free, and come already varnished...
those are advantages. my home depot pipe loop has 4 $1.21 3/4" t's
and 2 $0.81 3" nipples and 2 $0.67 2" nipples. the loop od is about
6"x8" i'm planning to cut one of the t's in half to make a variable
air gap by scewing its halves in an out of its t's. the t's allow
cooling air to flow in at the bottom and out at the top of each side
of the loop. the pipe wall thickness is about 1/8".
i also bought a 500' roll of #12 stranded thhn wire with pretty green
15 mil 105 c 600 v insulation for $20.60.
>disadvantage i see with using iron pipe (which is probably more
>like cold rolled steel, but galvanized, or zinc plated) is the
>eddy currents.
i just got a 1993 schaum's outline on "electromagnetics" which has some
b/h curves. "cast iron" has b = 0.74 teslas (wb/m^2) of magnetic flux
density at a magnetic intensity h = 5000 a/m, ie the relative permeability
mur = b/(muoh) = 0.74/(5000x4pix10^-7) = 118. at h = 400, b = 0.18, so
mur = 358. "cast steel" has a higher permeability (almost as good as
silicon steel), with b = 1.62 t at h = 5000 a/m and b = 0.73 at h = 400, ie
mur = 258 and 1452. i assume that iron pipe is "cast iron" vs "cast steel."
page 174 of the outline says "figure 11-14 shows the extreme nonlinearity
of mur vs h for silicon steel. this nonlinearity requires that problems
be solved graphically."
>for a choke, i can see where this might actually be an advantage.
maybe so, in this application, altho i'd also like to minimize
real power loss, which makes heat and costs money.
>what is it you're trying to accomplish?
one purpose is to (re)learn more about magnetics. so far i've discovered
ohm's law for magnetics:
f = phi r, where f ("e") is the magnetomotive force in ampere turns (ni),
phi ("i") is the flux in webers, and
r ("r") is the reluctance.
reluctances add like resistors in series. a section of a magnetic circuit
with length l and permeability mu and cross-sectional area a has reluctance
r = l/(mua) in units of 1/h.
this outline contains many other cheerful facts, like l = n phi/i.
further suggestions welcome.
nick
|
|
