What are Radio coil types and how to wind them?

[ali6x944]

Hello everybody,
I'm working on making a crystal radio, and i ran into a problem...
 the radio i made was not working, then i salveged one out of an old am radio, it woked fine after that...
Unfortunitly i can not use the coil due to its massive size so i would like to know What are Radio coil types and how to wind them?

[German_EE]

It looks like a lot of people have viewed your post but none have answered, perhaps because it is a little too vague. I will make an attempt.

In electronics there is only one type of coil or inductor, a length of wire wound in a spiral and then connected at both ends. Two things make a difference to the inductor, the core material and the 'Q' of the coil.

Coil Material
The simplest core material is fresh air, an air cored inductor, and an example can be seen here: http://www.eham.net/data/classifieds/images/266326.jpg Unless you use a VERY large coil this will be no good for a crystal radio because the value of inductance will be too low. Most coils use ferrite cores and these can either be in the form of a circular ring or toroid or in the form of a slug which is adjusted in and out. There is also the ferrite rod antenna (normally about 15cm long) that can be seen inside most Medium Wave and Long Wave radio receivers, these would be ideal for a crystal set: http://www.petervis.com/Radios/national-panasonic-rf-1103/national-panasonic-rf-1103-ferrite-rod-antenna/national-panasonic-rf-1103-ferrite-rod-antenna.gif Note that a ferrite core will only operate over a range of frequencies and if you use the wrong core your radio will not work.

Coil 'Q'
Coil Q is a measurement of the efficiency of an inductor, the higher the Q the better the coil. If you wind an air cored inductor maximum Q is obtained when a) the diameter of a coil is equal to its length b) the resistance is as low as possible and c) the distance between the windings is equal to the thickness of the wire. This can mean some large coils (see above) but it can be done http://home.earthlink.net/~belundy/xtal-set.html When a ferrite core (especially a ring core) is used and the ferrite is used inside its working frequency then Q is increased by a very large amount which means that a smaller coil can be made. If the ferrite core is outside of its working frequency then the Q of the coil is lowered and you now have an RF choke, something which is still an inductor but is used to suppress RF signals on power lines  https://www.radioshack.com/collections/connectors-connectivity_instore/products/radioshack-100-h-rf-choke?variant=5717358789 Yes, it does look just like a small ferrite rod antenna but the core material is different.

So, in summary, for your crystal set to work:

1) You need either a large air core inductor or a ferrite rod

2) You need the right number of turns on the inductor

3) You need to wind the coil for maximum Q by taking care to space the turns using wire that is not too thin because of resistance losses.

This post turned out longer than I planned but I hope that it helps. I am aware of the Internet restrictions in Saudi Arabia so some care has been taken with the links which are all safe. Hope this helps!

[Z80]

Hi,  It might help if you explain more about what you are trying to do.  The traditional method to make crystal set coils is either pull one from an old am radio (like you did) or wind your own 'air cored' on an old toilet roll tube or similar.  The purpose of the coil is to form a tuned circuit with the capacitor which allows you to (sort of) tune in individual stations.  For this to work as well as possible, the Q of the tuned circuit needs to be as high as possible which usually means large-ish coils as German_EE explained.  You could probably get a small SMD inductor to tune but the performance is likely to be terrible.

[voltz]

For Medium wave AM broadcasts, its 60 turns of enamelled copper wire on a ferrite rod 3 to 5 inches long.
Just wind it by hand, easy to do. Then a dab of hot glue to hold the ends in place and stop it springing out. Keep the turns close together, side by side.

[Ian.M]

Its better to make up a sleeve for the rod to wind the coil on.  Start with the middle of the rod wrapped in clingfilm so you can get the sleeve off again.  Then wrap a strip of cartridge paper round the rod, securing it with superglue.  Build it up a couple of turns, saturating each with superglue. Let it all fully cure, slide it off the rod (impossible without the clingfilm), clean out as much as possible of the clingfilm from the center of the sleeve, and wind the coil using the sleeve as a former.   The start can be anchored with a narrow strip of paper and superglue.  Tack the end into place with stickytape in case you need to add or remove turns.   Once you are satisfied you can get the range you need, lock the end in place as you did the start and 'dope' the coil with superglue.   

If your set design calls for a tapped coil, use magnet wire with solder-through enamel, and for each tap, make a 4" loop, twist up just the base of the loop tightly, and solder the twist, taking care to get a good joint.  Cut one side of the loop at the twist to get an 8" tail.  Be careful NOT to overheat adjacent turns.  It may be advisable to wind the tap turns ONLY over a narrow strip of cartridge paper placed lengthwise along the rod to protect the other turns from the heat of soldering.

The total inductance of the coil and thus the tuning range can be adjusted by sliding the sleeve bearing the coil off-center on the rod.  Adjust it with the tuning cap's padding trimmer(s) initially set to 50% and the main vanes fully engaged (max capacitance) for the desired minimum frequency.   Use the padders to adjust the maximum frequency with the main vanes disengaged (min. capacitance).  You will probably need to readjust the coil (without touching the padders) and then the padders to get the dial limits correct.  When the coil is correctly set, lock it in place with a light dab of hotglue or beeswax either end.

When mounting the rod, its important NOT to have too much metal near it and absolutely critical NOT to have any metal loops round it or plates across the ends.   Cable ties direct to the PCB or plastic P clips are good.  If there is a metal chassis, use plastic standoff pillers under the P clips.

[ali6x944]

Thanks for the graet info!
Last question is if i have a 500mm Ferrite rod and i want to use it in am radio but i don't know the number of windings or turns to make an inductance of 350uh with a center tab and extra winding like these ones found in old pocket radios...
If any one know how please post...
Thanks for the help everybody

[Ian.M]

Ferrite rods vary and are *HORRIBLY* complex to design coils for. See http://g3rbj.co.uk/wp-content/uploads/2013/10/Web_-The_Inductance_of_Ferrite_Rod_Antennas.pdf

Unless you have good data for the specific rod you are using, you will have to wind a test coil on it and measure the inductance with the coil centered then offset halfway to one end.  Add or remove turns till the reading in the two positions brackets your desired inductance.

You will need a good LCR meter that can work with a low amplitude test signal, or a GDO and a cap to resonate the test coil.

[orolo]

For crystal radios, I've seen people using either cylindrical air coils, ferrite rod coils, or helical coils (there are other exotic variations). If size is a concern, the ferrite rod is the best option.

Ferrite rods vary and are *HORRIBLY* complex to design coils for. See http://g3rbj.co.uk/wp-content/uploads/2013/10/Web_-The_Inductance_of_Ferrite_Rod_Antennas.pdf

Unless you have good data for the specific rod you are using, you will have to wind a test coil on it and measure the inductance with the coil centered then offset halfway to one end.  Add or remove turns till the reading in the two positions brackets your desired inductance.

I have been reading all G3RBJ's articles recently, because I want to experiment a bit with LF and some huge ex-soviet ferrite rods I got from ebay. The articles are excellent, each one worth detailed reading. But the moral is that ferrite coils are not that complex, after all, since for high permeability rods the most important factor is their length/area ratio, which is easily measured. The reasong roughly being that, since in ferrite rods the magnetic flux must return through a long path in free space, their permeability is shadowed by that of the returning path-- something like a huge air gap.

Therefore, for a high permeability rod, the inductance is just a multiple of the air-core coil inductance, which is easily predicted (as long as you are working at low frequencies, otherwise, things get hairier). Then, knowing your ferrite shape, and assuming its permeability is high enough, you multiply the air core inductance by a constant in the range 10-40 to get the final inductance. The details and formulae are in the article, but it amounts to looking up figure 6.2 . Of course, to know the exact inductance an experimental setup is needed (I'd use the ferrite in an oscillator with known capacitance), but G3RBJ's article seems good enough to make an easy, educated guess of the design of the antenna for any ferrite.

One detail worth taking account of in the case of rod ferrites is their Q factor. For a well-designed inductor, values in the 500 range can be attained; I have seen values near 1000 for litz wire versions. A Q of 500 at 1Mhz is a bandwidth of just 2Khz.

Thanks for the graet info!
Last question is if i have a 500mm Ferrite rod and i want to use it in am radio but i don't know the number of windings or turns to make an inductance of 350uh with a center tab and extra winding like these ones found in old pocket radios...
If any one know how please post...

What is the rod's diameter? You need that parameter, in addition to the rod's length. What wire gauge/diameter are you going to use? Are you winding directly on top of the ferrite, or using a sleeve? Note that you shouldn't wind non-enamelled wire directly on a ferrite, since many ferrites are conductive.

[Radio Tech]

If you could provide a schematic of the circuit you are building/working on it would be a lot of help.

[ali6x944]

it is 0.9mm, and what i posted was the rod's length *i just forgot to note that sorry*...
and it is wound directly onto the rod

[orolo]

it is 0.9mm, and what i posted was the rod's length *i just forgot to note that sorry*...
and it is wound directly onto the rod

Okay, so d = 0.9mm, L = 50mm. Then L/d = 5.556.

Covering the whole rod with a single coil would be unpractical, since it is too long. We must assume that the rod will protrude from both sides of the coil, at least 1cm to each side. This will somewhat increase the relative permeability of the coil, compared to a completely covered rod. For a rod of your proportions, L/d = 5.5, fig. 4.6 predicts an inductance gain L/L0 of 15 to 18, depending on the relative permeability of the ferrite. With the ferrite protruding, we can assume that L/L0 will be on the higher end, maybe reaching 20. Since you want 350uH, the air core coil (no rod) should have about 350/20 = 17.5uH inductance.

I have 28AWG magnet wire lying around, with a diameter of 0.321mm, which allows me 10 turns for each 3mm. If you have different wire, just make 10 turns and measure them. With a good online calculator, i get that 80 turns, with a coil length of 2.7cm and a diameter of 9.3mm (from wire center to wire center) gets me about 17.4uH. Let us work out the details with this coil.

First, I wrote a small python script to work out the calculations:

Code: [Select]

from math import *

# Nagaoka's factor, formula A1.2.
# This formula does not change with length units.
def Kn(lc, dc):
    return 1 / (1 + 0.45*dc/lc - 0.0005*(1.0*dc/lc)**2)

# Inductance of air core coil, formula A1.1.
# All lengths in millimeters.

def L0(lc, dc, N):
  return (3.1416 ** 2) * 1e-10 * N**2 * (dc ** 2) / lc * Kn(lc, dc)   


#
# The following formulae are independent of length unit.
#

# Relative permeability, outside, formulae 7.5, 7.6
def k(lf, lc, df, dc):
    canf_d_e0 = 0.5 * 3.1416 * (lf-lc) / ( log(2.0*(lf+df)/df) - 1)
    return (canf_d_e0 + 2*df)/(2*dc)

# Formula 4.1
def x(lf, lc, df, dc):
    return 4.46 * lc / dc / (1 + 1.17*dc/lc)

# Inductance ratio, formula 5.1.1
def Lr(lf, lc, df, dc, uf):
    my_x = x(lf, lc, df, dc)
    my_k = k(lf, lc, df, dc)
    return ((1 + my_x) / (1/my_k + my_x/uf))

#
# Adding it up:
#
def coil():
    lc = input("Coil length (mm)       : ")
    dc = input("Coil diameter (mm)     : ")
    lf = input("Ferrite length (mm)    : ")
    df = input("Ferrite diameter (mm)  : ")
    N  = input("Number of turns        : ")
    uf = input("Ferrite permeability   : ")
    my_L0 = L0(lc, dc, N) * 1e6
    my_Lr = Lr(lf, lc, df, dc, uf)
    print "Air core inductance, L0, in uH: ", my_L0
    print "Inductance gain with rod L/L0 : ", my_Lr
    print "Inductance of ferrite rod, uH : ", my_L0 * my_Lr

So I run this for our model:

Code: [Select]

>>> coil()
Coil length (mm)       : 27
Coil diameter (mm)     : 9.3
Ferrite length (mm)    : 50
Ferrite diameter (mm)  : 9
Number of turns        : 80
Ferrite permeability   : 100
Air core inductance, L0, in uH:  17.5196007631
Inductance gain with rod L/L0 :  18.7215368833
Inductance of ferrite rod, uH :  327.993851867
Hmm, about 330uH, which is almost spot-on, given the 5% error implicit in all calculations. If the permeability of the ferrite goes up to 200, the final inductance of the rod turns out to be 358uH. If the permeability is just 50, it goes down to 280uH. Now, the inductance of ferrite rods is, for some measuring reasons we don't want to dwell into, usually overestimated (eg. my soviet ferrites claim uf = 400, which I don't trust for a moment), but uf=100 seems a safe bet.

Just to be on the safe side, I would make a coil of about 80-85 turns, and try it in a colpitts oscillator, to see if the frequency falls on the desired range.

[bills]

try this link http://www.midnightscience.com/formulas-calculators.html

[Ian.M]

@orolo: It would be worth extending your script to calculate for off-center coils. and writing another one to do the reverse calculation for permeability from test coil results.

[orolo]

@orolo: It would be worth extending your script to calculate for off-center coils. and writing another one to do the reverse calculation for permeability from test coil results.

Thank you, I will do that as soon as I get the time. I shall need the program anyway for my experiments, so it could be of some use to other people.

[orolo]

I'm making progress with the rod inductance calculator. Since there seems to be no good calculator online, I intend it to be complete, precise and open source.

As a first step, I have refined the air coil inductance calculator to include Nagaoka's factor with 4ppm precision, Rosa's corrections to the tens of ppms, internal inductance, and substituting coil length by number of turns and winding pitch. I've followed the magnificent notes by David W. Knight, taking his most precise formulas (to the exception of those that depend on elliptic functions-- I have discovered that the SciPy version 0.13.3 in my python does not compute them well, after some hours of frustration, so I have used the best polynomial formulas provided, and tested against Maple and known data to assure correctness).

The code, as it stands now, is:

Code: [Select]

from math import log, log10, sqrt, exp

#  Ferrite rod inductance calculator.
#
#  Calculations taken from:
#
#  [P] Payne, Alan. "The Inductance of Ferrite Rod Antennas".
#  http://g3rbj.co.uk
#
#  [K] David W. Knight. "Solenoid inductance calculation".
#  http://www.g3ynh.info

pi   = 3.14159265358979323
dpi  = 2*pi
pi2  = pi ** 2
pim1 = 1 / pi
pid2 = pi / 2
sqr2 = sqrt(2)

mu_ir = 0.999995044                 # Copper relative permeability.
mu_i  = 4 * pi * 1e-7 * mu_ir       # Copper permeability.
rho   = 1.7241e-8                   # Copper resistivity (annealed)

#######
# COIL DIMENSION RELATED:
#

# Minimum pitch angle, in case we need it.
# [K] formula 5.5, independent of units.
def psi_min(dc, dw):
    return arcsin(dw / (pi * dc))

# Secant of pitch angle, [K] 5.2.
def sec_psi(lc, dc, N):
    return sqrt(1 + (lc / (pi * dc * N))**2)

# Minimum wire pitch.
# [K] formula 5.4, dimensional but unit independent.
def p_min(dc, dw):
    return dw / sqrt(1.0 - (dw / (pi*dc))**2)

# Current-sheet diameter correction, for thick wires.
# [K] formula 3.1a.  Independent from units.
def D0(dc, dw):
    return dc * (1.0 - (1.0 * dw / dc))

# Length of wire.
# Dimensional, but independent of units.
def lw(p, N):
    return p*N

###
# NAGAOKA CORRECTIONS:
#

# Nagaoka's factor, Lundin's formula, [K] Sec. 8a.
# This formula does not change with length units.
# Tested against exact (Lorenz) with Maple.
def Kn(lc, dc):
    if (lc > dc):
        x = 1.0 * dc / lc
        x2 = x * x
        return (1 + (0.383901 + 0.017108*x2) * x2) / (1 + 0.258952*x2) - (4*x)/(3*pi)
    else:
        x = 1.0 * lc / dc
        x2 = x * x
        t1 = (1 + (0.383901 + 0.017108*x2) * x2) / (1 + 0.258952*x2)
        t2 = (0.093842 + (0.002029 - 0.000801*x2) *x2 ) *x2
        return (2 * x / pi ) * ( -t1 * ( log(0.25*x) + 0.5 ) + t2 )

###
# INTERNAL INDUCTANCE:
#
# THIS CORRECTION IS NOT ENHANCED BY THE MAGNETIC CORE!

# Skin depth. Answer in mm. This formula is dimensional.
def d_i(f):
  return 1e3 * sqrt(rho / (pi * f * mu_i))

# Internal inductance factor. [K], formula 6.4.
ex0 = 1 / 3.74

def phi(dw, f):
    di = d_i(f)
    z = (0.27445 * dw) / (sqr2 * di)
    y = 0.02369 / ( 1 + 0.2824 * (z**1.4754 - z**(-2.793))**2 )**0.8955
    t = ( 1 - exp( -(dw/(4*di))**3.74) )**ex0
    return 4 * di * t / (dw * (1-y))

# Internal inductance in Henries. [K], formula 6.5
def Li(lc, dc, dw, N, f = 0):
    return mu_i * (1e-3 * dc) * N * phi(dw, f) * 0.125 * sec_psi(lc, dc, N)

###
# ROSA'S CORRECTIONS:
#

# Rosa's self inductance correction. [K], formula 10.9.
#
k0 = 1 / (log(8*pim1) - 0.5)
k1 = 3.437
k2 = 24 / (3*pi2-16)
w1 = -0.47
w2 = 0.755
def Pnom(x):
    return k0 + (k1 + k2 * x)*x + w1 / (w2 + (1 / x))**1.44
def Ks(lc, dc, N, dw, f = 0):
    x = lc / (N * dc)  # p / D
    y = dw / dc
    z = log(0.125 * y)
    t  = log(1 + pid2 / x) + 1 / Pnom(x) + z + 2
    t -= y * y * (0.333333333333333 - z) / 8
    t -= mu_ir * phi(dw, f) * 0.25 * sqrt(1 + (lc * pim1 / (N * dc))**2)
    return t

# Rosa's mutual inductance correction. [K], formula 10.18.
#
c0 = log(dpi) - 1.5
c1 = -0.33084236
c2 = -1.0 / 120
c3 = 1.0 / 504
c4 = -0.0011925
c9 = -(c0 + c1 + c2 + c3 + c4)
def Km(N):
    x = 1.0 / N
    x2 = x*x
    return c0 + ( -log(N)/6 + c1 + ( c2 + (c3 + (c4 + c9*x2) * x2) * x2) * x2) * x

##
# INDUCTANCE OF AIR CORE COIL:
#

# Inductance of air core coil, formula A1.1.
# All lengths in millimeters.
# NOTE: Internal inductance __NOT__ added.

def L0(lc, dc, N, dw = 0, f = 0, thin = True, rosa = True):
  if (thin):    # Thick wire corrention?
      my_dc = dc
  else:           
      my_dc = D0(dc, dw)

  if (rosa):
      rosa = ( Km(N) + Ks(lc, dc, N, dw, f) ) * lc / (pid2 * dc * N)
  else:
      rosa = 0
  return pi2 * 1e-10 * N**2 * (my_dc ** 2) / lc * ( Kn(lc, my_dc) - rosa )   


#
# The following formulae are independent of length unit.
#

# Relative flux to maximum, due to finite permeability, formula 8.1:
def rflux(lf, lc, df, uf):
    return 1.0 / (1.0 + ((1.0*(lf - lc)/df)**1.4) / (5.0*uf) )

# Relative reluctance, outside, formulae 7.5, 7.6 (corrected in 8.2)
def k(lf, lc, df, dc, rel_flux):
    canf_d_e0 = 0.5 * pi* (lf-lc) / ( log(2.0*(lf+df)/df) - 1.0)
    return (rel_flux * canf_d_e0 + 2.0*df)/(2.0*dc)

# Formula 4.1
def x(lf, lc, df, dc):
    return 4.46 * lc / dc / (1 + 1.17*dc/lc)

# Inductance ratio, formula 5.1.1
def Lr(lf, lc, df, dc, uf):
    my_x = x(lf, lc, df, dc)
    rel_flux = rflux(lf, lc, df, uf)
    my_k = k(lf, lc, df, dc, rel_flux)
    return (1.0 + my_x) / ( 1.0 / my_k + my_x / uf)

# Apparent permeability correction, formula 5.2.1
def uapp(df, dc, uf):
    return ( (uf - 1.0) * ((1.0 * df) / dc)**4 ) + 1


#
# Adding it up:
#
def coil():
    dc = input("Coil diameter (mm)                   : ")
    dw = input("Wire diameter (mm)                   : ")
    N  = input("Number of turns                      : ")
    p  = input("Winding pitch (mm)(zero for minimum) : ")
    if (p == 0):
        p = p_min(dc ,dw)
        print ("-> Okay, pitch = %f mm" % p)
    lc = N * p
    lf = input("Ferrite length (mm)                  : ")
    df = input("Ferrite diameter (mm)                : ")
    uf = input("Ferrite permeability                 : ")
    f  = input("Frequency                            : ")
    my_L0 = L0(lc, dc, N, dw, f, thin=True, rosa=True) * 1e6
    my_uf = uapp(df, dc, uf)              # Coil-rod air gap correction.
    my_Lr = Lr(lf, lc, df, dc, my_uf)
    my_Li = Li(lc, dc, dw, N, f) * 1e6
    print "Coil length, in mm            : ", lc
    print "Winding pitch, in mm          : ", p
    print "Internal inductance, in uH    : ", my_Li
    print "Relative uf due to gap        : ", my_uf
    print "Air core inductance, L0, in uH: ", my_L0 + my_Li
    print "Inductance gain with rod L/L0 : ", my_Lr
    print "Inductance of ferrite rod, uH : ", my_L0 * my_Lr + my_Li
and an example of use:

Code: [Select]

>>> coil()
Coil diameter (mm)                   : 9.3
Wire diameter (mm)                   : 0.3
Number of turns                      : 80
Winding pitch (mm)(zero for minimum) : 0
-> Okay, pitch = 0.300016 mm
Ferrite length (mm)                  : 50
Ferrite diameter (mm)                : 9
Ferrite permeability                 : 100
Frequency                            : 1000000
Coil length, in mm            :  24.0012652969
Winding pitch, in mm          :  0.300015816212
Internal inductance, in uH    :  0.0991453072228
Relative uf due to gap        :  87.8307271843
Air core inductance, L0, in uH:  19.107760332
Inductance gain with rod L/L0 :  17.2919651051
Inductance of ferrite rod, uH :  328.795453012
Since there are so much parameters, and more to come, a GUI is inevitable, to include graphs (vs. frequency, permeability, and core position at least). I'll try to stick to matplotlib and gtk, which can be downloaded in a bundle with Portable Python and run at once.

I'll try to include, besides a direct inductance calculator:

-> Number of turns to achieve some inductance, given pitch, wire, coil radius and rod dimensions (permeability variable).
-> Permeability calculator, from inductance and dimensions for wire, coil and rod.
-> Off-center inductance calculator.

That's a lot of work, on top of my regular work and some heavy transistor testing I'm involved with, but I'll try to come trough.