Tweet
#248 by Jim Koehler.
[Consult the original if any questions because I could have made
errors in transcription - das]
BIFILAR WINDINGS
POSTING #226.
When I was using the spreadsheet tables in Jim Koehler's "Proton
Precession Magnetome-
ters") I noted that for any fixed size of fluid container; as the
number of turns of
any given wire size increased the signal level increased because of
the increased in-
ductance and coupling. At the same time the polarization current,
of course, de-
creased because of the increased inductance and coupling. At the
same time, the po-
larization current, decreased due to the increased resistance of the
longer wire. I
also noted that for any given wire size if I could double the
polarization current for
the same number of turns I would get a substantial increase in the
signal to noise ra-
tio.
This can be simply accomplished by using bifilar windings. That is,
two windings are
applied at the same time. The switching is a bit more complicated
and requires 4 FORM
C contacts rated in excess of the expected forward and backward
currents. The sche-
matic below shows the switching arrangement. Polarization current
is applied to the
two windings in parallel thus halving the resistance of the sensor.
This will double
the polarization current thus increasing precession probability.
The net result is a
stronger signal and a much improved signal to noise ratio.
The two coils are then connected in series for the detection cycle
giving the sensi-
tivity associated with the larger number of turns. Trifiliar and
even more numerous
windings are possible but there is a practical limit.
(Diagram in the original is at this point.)
Take care with polarity. Lee Fraser.
POSTING #227.
By George - I think you have something there Lee.
You can reduce the rating of the contacts by using a relay to do the
changeover thing.
But using a separate switching element (i.e. power transistor with
flywheel diode) to
apply the polarization current.
I guess this adds [to] the complexity of sequencing the switching
and rewiring relay.
Advantages could be that the reverse EMF induced in the polarization
coil, generated
as the polarizing field collapses is fed back to the battery
(assuming a battery with
low enough internal resistance) so preventing arcing and extending
battery life. (Not
sure how much though.) As switching contacts are not switching
current they can be
downrated.
Disadvantages would be that due to the amount of time required for
the sequencing,
(taking into account the time it takes for the polarizing field to
collapse via back
EMF into the battery), more of the precious time of precession could
be lost.
This effectively brings the discussion into a bit of a circle and
leaves us looking at
the proton rich fluid. Le Kirby
POSTING 234.
Been thinking some more. (V Sad sorry)
Using a relay to switch the windings input could be a bad idea.
Solenoids and magnetic fields and all that Tesla and Henry stuff.
A low power switching matrix (maybe using the analog MUX ICs from
Maxim or some such,
may be tenable. Alternatively use surface mount transistors. The
FET kind switch
very quickly and have a low on resistance.
Use a power transistor to switch the polarization supply (with
flywheel/protection di-
ode as discussed, to siphon off back EMF.
Providing the low power switches are either full on or full off and
are only switched
whilst the coil is effectively free of back or forward EMF. They
should survice fine.
The way the switching element can be rugged, and close up to the
coil arrangement and
cause minimal disruption to the sensed magnetic field.
Switching times should also be better than for those of the contact
change over for a
relay. Just thoughts. Le Kirby.
POSTING #239.
I think you are right about keeping relays away from the windings.
I also think that
if you consider winding-switching to maximize polarizing current,
you would be better
off using the same circuit elements (controls, transistors, etc.) to
generating the
optimal voltage and current to discharge into the simple monofilar
coil. After all,
if you double voltage and halve current, you are precisely at the
same polarizing ef-
ficacy. A voltage converter feeding a suitable capacitive storage
bank should do.
Peter Boetzkes.
POSTING #240.
I think this approach also has some merits.
Certainly increasing windings and increasing polarization voltage
should, if done pro-
portionately, keep the flux induced in the core the same, whilst
decreasing copper
losses due through resistive ions in the windings. And, as a result
of having more
windings you should increase the signal pickup.
There are some other factors to consider in this arrangement:
Increasing the windings increases the inductance of the coil and
therefore impedance
of the coil. (Impedance is to AC what pure resistance is to DC, but
is a fictional
effect that reflects the sum of the interactive effects that come
into play as soon as
you allow EMF to vary.) The net effect should be to increase the
time constants as-
sociated with establishing the polarizing magnetic field and the
time taken for it to
collapse sufficiently for our measurements to be taken. (Be glad to
take advice here,
my AC theory is not that good - too much math.)
I guess I am assuming that we are creating no more of a polarizing
field than is abso-
lutely necessary to do the job.
Think of the coil as being a bath tub and where the coil stores
electrical energy as a
magnetic field the bath tub stores mains water as bath water.
(Water is water, energy
is energy.)
In filling the bath you do it through a pipe of a given diameter
(coil winding diam-
eter) at a given pressure (Voltage) and flow rate (Current).
Unlike the bath analogy when you turn off the water supply it must
all rush out again,
magnetic coils are dynamic storage mechanisms. If you try to
prevent it, the pressure
will increase until something gives. (I.e. the voltage increases
until it can conduct
through something, when switched through relay contacts this is what
causes the spark-
ing or arcing). The same feature is put to good effect in joke
electric shock ma-
chines.
Filling the bath through a hose of smaller diameter at higher
pressure is OK but it
takes longer to do. And, when emptying the effect is the same.
The essential question is...
Does the increase of polarizing time constant for either technique
applied to a given
arrangement significantly cut into the preciously short decay
period?
The real aim of the game is...
to get access to as much of the decay period as we can (extend
through proton source,
maximize signal to noise, etc.) and polarize only when necessary
(max power efficiency
in a mobile set).
I am also assuming that you can not really tap off a usable signal
from the discharg-
ing polarization coil.
I think the man with the plan via a via magnetics must be Jim
Koehler. His math and
grip of the subject far surpasses mine.
Out of interest, has anyone got a storage scope and captured the
output from a magne-
tometer sensor???? Le Kirby.
POSTING #241.
1. The magnetic field is proportional to the current in the coil
and the number of
turns. The current in a coil is determined by Ohm's Law of I = E/R.
In this case I
would suggest the higher the current the better. Also, if you
double the voltage and
halve the current, as was suggested, the power remains the same but
apparently the re-
sistance has doubled.
2. The polarizing current should be applied for some multiple of
the spin relaxation
time (on the order of seconds). This more or less rules out
capacitive discharge as a
source of the polarization current.
3. The detection time need not be very long. Under ideal
conditions a phaselock loop
can acquire and lock in a few tens of cycles and the actual
frequency measurement done
in a period on the order of a tenth of a second or less.
The hardest things to deal with are noise and the high gain required
to bring the sig-
nal to a usable level. Lee Fraser.
POSTING #242.
Don't discard the idea of capacitive discharge so easily. These
days you can buy 50
Farads (!) in a couple of cubic inches for a few dollars. These
capacitors hold en-
ergy well over 100 joules. I suspect they can definitely be
considered for polarizing
current. Peter Boetzkes.
POSTING #243
When you think through the analogy you present here, you find that
by doubling voltage
and halving current, you get EXACTLY the same results. The coil
with both halves in
series (i.e. the normal winding, not the bifilar one) has four times
the resistance of
the bifilar one, since the coil halves are in series rather than in
parallel. Also,
since the number of turns is double (of the bifilar), the inductance
is quadruple (L
being proportional to turns squared). If you double the voltage,
the rate of rise of
current is still only half (since L is 4x) but the current you wish
to achieve is also
only half (since double turns) to achieve the same magnetic field.
Hence the time
constant is exactly the same. The same rationale hold for
resistance. There really
is no difference, it is just a matter of changing the impedance of
the power supply to
match that of the coil. Simple transformer action. Peter Boetzkes.
POSTING #244
Peter - Good stuff but I don't think its quite right. Remember
di/dt and the fact
that the ideal shape for the polarizing pulse is rectangular. This
is which I would
not bother with a charge pump, although it is an attractive idea.
Lee Fraser.
POSTING #245.
Peter: I was thinking about a capacitor discharge and drew out the
equivalent circuit
(an inductance, a resistance) ( didn't add one for the capacitor)
and a capacitor.
This is a tank circuit and most likely the only thing it will
produce is a exponen-
tially decaying sinusoidal wave form. Yes low cost BIG capacitors
are available but
check out their voltage rating. You have not lived until a
capacitor blows up in your
face.
Now if you think CD has merit then I would like to see a design that
I can build that
clearly deals with the turn off of the polarizing field. Also a
capacitor is a volt-
age device so when it is switched into the discharge path (through
the coil) its ter-
minal voltage will be determined mostly by the resistance of the
coil thus limiting
the current. Lee Fraser.
POSTING #246.
Lee: The sinusoidal (multicycle) wave form is easily eliminated with
a simple series
diode which prevents the current from flowing backwards in the coil.
I have found
that works well with capacitor-discharge magnetizing circuits for
permanent magnets.
So this turns off the polarizing field just as it tries to reverse.
If you want the
peak intensity to remain longer, you must also prevent the capacitor
from charging in
reverse, using a second diode (reverse biased) connected in parallel
with the ca-
pacitor. Basically this forces the peak stored magnetic-field
energy to be dissipated
in the coil (maintaining a decaying but non-alternating field)
rather than
reverse-charging the capacitor. Of course, it is necessary to match
the voltage with
the number of turns on the coil, its resistance, and the desired
flux density. If one
capacitor has sufficient energy (making series capacitors a waste)
then smaller-C ca-
pacitors can be connected in series to provide the same energy with
higher voltage.
Peter Boetzkes.
POSTING #247.
Lee: I am fully taking di/dt into account. The shape of the
polarizing pulse makes
no difference to the fact that NOTHING CHANGES going from a given
coil split into two
and paralleled, to the exact same coil connected in series, with the
voltage doubled
and the current halved. Both aspects of the impedance (resistance
and inductance)
change in the same proportion (4x) which is exactly the change of
voltage/current
(4x). Therefore wave shapes, min/max field intensities, energy
absorbed by the mag-
netic field, energy absorbed by the copper windings, are all exactly
the same. Actu-
ally this principle has applied to a lot of power conversion
equipment I have de-
signed. A charge pump is not necessarily the most efficient way to
charge the
capacitor, incidentally. In fact, conventional inductive chargers
usually have better
efficiency if they are well designed. Peter Boetzkes.
POSTING #248.
I've been reading the debate about bifilar windings with some
interest. I don't think
it is worth it for the reasons stated below. We will be comparing
two sensors; one
with a single winding and the second with two bifilar windings each
of which is half
the number of turns of the single winding. Then, let us connect
them in parallel for
polarization and then in series to extract the signal.
For each bifilar winding, the resistance is half that of a single
winding so, using
the same battery to polarize it, the current is double that of the
single winding.
The polarizing field is proportional to the number of turns, times
the current, so
each bifilar winding gives exactly the same polarizing field as the
single winding -
but, but because there are two of them in parallel, the total
polarizing field is
doubled.
When they are connected together in series, they will give a total
signal equal to
what the single winding would give for the SAME polarizing field.
Therefore, since
the polarizing field is twice as big, you get a total output of
twice what you would
get from a single winding.
However, the total polarizing power needed is four times what the
single winding needs
because there are two windings in parallel, each taking twice the
current of a single
winding.
SUMMARY: You get twice the output signal from a switched bifilar
winding, but you use
four times the power polarizing it. You therefore use up your
battery four times as
fast - probably faster because battery lifetime depends slightly on
current - the
greater the current, the smaller the number of amp-hours the battery
will supply. In
addition, you have the problem of the additional complexity.
However, if you are not
worried about power (i.e., if you are building an instrument for use
from a boat or a
vehicle) or complexity (more things to go wrong - and relays are the
weak link in
these sort of systems because they are electromechanical devices),
you DO get twice
the signal for the same number of turns of wire overall.
In general, the signal is proportional to polarizing current, times
the square of the
number of turns. If you want a bigger signal, it is easiest to get
it by just in-
creasing the number of turns. For a given battery voltage and wire
size, if you
double the number of turns, you double the resistance and so halve
the current. You
get, overall, a doubling of the signal. As well, because you've
reduced the polar-
izing current, you increase battery life. That is by simply
doubling the number of
turns, you get twice the signal and only use half the power. Jim
Koehler.
[Consult the original if any questions because I could have made
errors in transcription - das]
BIFILAR WINDINGS
POSTING #226.
When I was using the spreadsheet tables in Jim Koehler's "Proton
Precession Magnetome-
ters") I noted that for any fixed size of fluid container; as the
number of turns of
any given wire size increased the signal level increased because of
the increased in-
ductance and coupling. At the same time the polarization current,
of course, de-
creased because of the increased inductance and coupling. At the
same time, the po-
larization current, decreased due to the increased resistance of the
longer wire. I
also noted that for any given wire size if I could double the
polarization current for
the same number of turns I would get a substantial increase in the
signal to noise ra-
tio.
This can be simply accomplished by using bifilar windings. That is,
two windings are
applied at the same time. The switching is a bit more complicated
and requires 4 FORM
C contacts rated in excess of the expected forward and backward
currents. The sche-
matic below shows the switching arrangement. Polarization current
is applied to the
two windings in parallel thus halving the resistance of the sensor.
This will double
the polarization current thus increasing precession probability.
The net result is a
stronger signal and a much improved signal to noise ratio.
The two coils are then connected in series for the detection cycle
giving the sensi-
tivity associated with the larger number of turns. Trifiliar and
even more numerous
windings are possible but there is a practical limit.
(Diagram in the original is at this point.)
Take care with polarity. Lee Fraser.
POSTING #227.
By George - I think you have something there Lee.
You can reduce the rating of the contacts by using a relay to do the
changeover thing.
But using a separate switching element (i.e. power transistor with
flywheel diode) to
apply the polarization current.
I guess this adds [to] the complexity of sequencing the switching
and rewiring relay.
Advantages could be that the reverse EMF induced in the polarization
coil, generated
as the polarizing field collapses is fed back to the battery
(assuming a battery with
low enough internal resistance) so preventing arcing and extending
battery life. (Not
sure how much though.) As switching contacts are not switching
current they can be
downrated.
Disadvantages would be that due to the amount of time required for
the sequencing,
(taking into account the time it takes for the polarizing field to
collapse via back
EMF into the battery), more of the precious time of precession could
be lost.
This effectively brings the discussion into a bit of a circle and
leaves us looking at
the proton rich fluid. Le Kirby
POSTING 234.
Been thinking some more. (V Sad sorry)
Using a relay to switch the windings input could be a bad idea.
Solenoids and magnetic fields and all that Tesla and Henry stuff.
A low power switching matrix (maybe using the analog MUX ICs from
Maxim or some such,
may be tenable. Alternatively use surface mount transistors. The
FET kind switch
very quickly and have a low on resistance.
Use a power transistor to switch the polarization supply (with
flywheel/protection di-
ode as discussed, to siphon off back EMF.
Providing the low power switches are either full on or full off and
are only switched
whilst the coil is effectively free of back or forward EMF. They
should survice fine.
The way the switching element can be rugged, and close up to the
coil arrangement and
cause minimal disruption to the sensed magnetic field.
Switching times should also be better than for those of the contact
change over for a
relay. Just thoughts. Le Kirby.
POSTING #239.
I think you are right about keeping relays away from the windings.
I also think that
if you consider winding-switching to maximize polarizing current,
you would be better
off using the same circuit elements (controls, transistors, etc.) to
generating the
optimal voltage and current to discharge into the simple monofilar
coil. After all,
if you double voltage and halve current, you are precisely at the
same polarizing ef-
ficacy. A voltage converter feeding a suitable capacitive storage
bank should do.
Peter Boetzkes.
POSTING #240.
I think this approach also has some merits.
Certainly increasing windings and increasing polarization voltage
should, if done pro-
portionately, keep the flux induced in the core the same, whilst
decreasing copper
losses due through resistive ions in the windings. And, as a result
of having more
windings you should increase the signal pickup.
There are some other factors to consider in this arrangement:
Increasing the windings increases the inductance of the coil and
therefore impedance
of the coil. (Impedance is to AC what pure resistance is to DC, but
is a fictional
effect that reflects the sum of the interactive effects that come
into play as soon as
you allow EMF to vary.) The net effect should be to increase the
time constants as-
sociated with establishing the polarizing magnetic field and the
time taken for it to
collapse sufficiently for our measurements to be taken. (Be glad to
take advice here,
my AC theory is not that good - too much math.)
I guess I am assuming that we are creating no more of a polarizing
field than is abso-
lutely necessary to do the job.
Think of the coil as being a bath tub and where the coil stores
electrical energy as a
magnetic field the bath tub stores mains water as bath water.
(Water is water, energy
is energy.)
In filling the bath you do it through a pipe of a given diameter
(coil winding diam-
eter) at a given pressure (Voltage) and flow rate (Current).
Unlike the bath analogy when you turn off the water supply it must
all rush out again,
magnetic coils are dynamic storage mechanisms. If you try to
prevent it, the pressure
will increase until something gives. (I.e. the voltage increases
until it can conduct
through something, when switched through relay contacts this is what
causes the spark-
ing or arcing). The same feature is put to good effect in joke
electric shock ma-
chines.
Filling the bath through a hose of smaller diameter at higher
pressure is OK but it
takes longer to do. And, when emptying the effect is the same.
The essential question is...
Does the increase of polarizing time constant for either technique
applied to a given
arrangement significantly cut into the preciously short decay
period?
The real aim of the game is...
to get access to as much of the decay period as we can (extend
through proton source,
maximize signal to noise, etc.) and polarize only when necessary
(max power efficiency
in a mobile set).
I am also assuming that you can not really tap off a usable signal
from the discharg-
ing polarization coil.
I think the man with the plan via a via magnetics must be Jim
Koehler. His math and
grip of the subject far surpasses mine.
Out of interest, has anyone got a storage scope and captured the
output from a magne-
tometer sensor???? Le Kirby.
POSTING #241.
1. The magnetic field is proportional to the current in the coil
and the number of
turns. The current in a coil is determined by Ohm's Law of I = E/R.
In this case I
would suggest the higher the current the better. Also, if you
double the voltage and
halve the current, as was suggested, the power remains the same but
apparently the re-
sistance has doubled.
2. The polarizing current should be applied for some multiple of
the spin relaxation
time (on the order of seconds). This more or less rules out
capacitive discharge as a
source of the polarization current.
3. The detection time need not be very long. Under ideal
conditions a phaselock loop
can acquire and lock in a few tens of cycles and the actual
frequency measurement done
in a period on the order of a tenth of a second or less.
The hardest things to deal with are noise and the high gain required
to bring the sig-
nal to a usable level. Lee Fraser.
POSTING #242.
Don't discard the idea of capacitive discharge so easily. These
days you can buy 50
Farads (!) in a couple of cubic inches for a few dollars. These
capacitors hold en-
ergy well over 100 joules. I suspect they can definitely be
considered for polarizing
current. Peter Boetzkes.
POSTING #243
When you think through the analogy you present here, you find that
by doubling voltage
and halving current, you get EXACTLY the same results. The coil
with both halves in
series (i.e. the normal winding, not the bifilar one) has four times
the resistance of
the bifilar one, since the coil halves are in series rather than in
parallel. Also,
since the number of turns is double (of the bifilar), the inductance
is quadruple (L
being proportional to turns squared). If you double the voltage,
the rate of rise of
current is still only half (since L is 4x) but the current you wish
to achieve is also
only half (since double turns) to achieve the same magnetic field.
Hence the time
constant is exactly the same. The same rationale hold for
resistance. There really
is no difference, it is just a matter of changing the impedance of
the power supply to
match that of the coil. Simple transformer action. Peter Boetzkes.
POSTING #244
Peter - Good stuff but I don't think its quite right. Remember
di/dt and the fact
that the ideal shape for the polarizing pulse is rectangular. This
is which I would
not bother with a charge pump, although it is an attractive idea.
Lee Fraser.
POSTING #245.
Peter: I was thinking about a capacitor discharge and drew out the
equivalent circuit
(an inductance, a resistance) ( didn't add one for the capacitor)
and a capacitor.
This is a tank circuit and most likely the only thing it will
produce is a exponen-
tially decaying sinusoidal wave form. Yes low cost BIG capacitors
are available but
check out their voltage rating. You have not lived until a
capacitor blows up in your
face.
Now if you think CD has merit then I would like to see a design that
I can build that
clearly deals with the turn off of the polarizing field. Also a
capacitor is a volt-
age device so when it is switched into the discharge path (through
the coil) its ter-
minal voltage will be determined mostly by the resistance of the
coil thus limiting
the current. Lee Fraser.
POSTING #246.
Lee: The sinusoidal (multicycle) wave form is easily eliminated with
a simple series
diode which prevents the current from flowing backwards in the coil.
I have found
that works well with capacitor-discharge magnetizing circuits for
permanent magnets.
So this turns off the polarizing field just as it tries to reverse.
If you want the
peak intensity to remain longer, you must also prevent the capacitor
from charging in
reverse, using a second diode (reverse biased) connected in parallel
with the ca-
pacitor. Basically this forces the peak stored magnetic-field
energy to be dissipated
in the coil (maintaining a decaying but non-alternating field)
rather than
reverse-charging the capacitor. Of course, it is necessary to match
the voltage with
the number of turns on the coil, its resistance, and the desired
flux density. If one
capacitor has sufficient energy (making series capacitors a waste)
then smaller-C ca-
pacitors can be connected in series to provide the same energy with
higher voltage.
Peter Boetzkes.
POSTING #247.
Lee: I am fully taking di/dt into account. The shape of the
polarizing pulse makes
no difference to the fact that NOTHING CHANGES going from a given
coil split into two
and paralleled, to the exact same coil connected in series, with the
voltage doubled
and the current halved. Both aspects of the impedance (resistance
and inductance)
change in the same proportion (4x) which is exactly the change of
voltage/current
(4x). Therefore wave shapes, min/max field intensities, energy
absorbed by the mag-
netic field, energy absorbed by the copper windings, are all exactly
the same. Actu-
ally this principle has applied to a lot of power conversion
equipment I have de-
signed. A charge pump is not necessarily the most efficient way to
charge the
capacitor, incidentally. In fact, conventional inductive chargers
usually have better
efficiency if they are well designed. Peter Boetzkes.
POSTING #248.
I've been reading the debate about bifilar windings with some
interest. I don't think
it is worth it for the reasons stated below. We will be comparing
two sensors; one
with a single winding and the second with two bifilar windings each
of which is half
the number of turns of the single winding. Then, let us connect
them in parallel for
polarization and then in series to extract the signal.
For each bifilar winding, the resistance is half that of a single
winding so, using
the same battery to polarize it, the current is double that of the
single winding.
The polarizing field is proportional to the number of turns, times
the current, so
each bifilar winding gives exactly the same polarizing field as the
single winding -
but, but because there are two of them in parallel, the total
polarizing field is
doubled.
When they are connected together in series, they will give a total
signal equal to
what the single winding would give for the SAME polarizing field.
Therefore, since
the polarizing field is twice as big, you get a total output of
twice what you would
get from a single winding.
However, the total polarizing power needed is four times what the
single winding needs
because there are two windings in parallel, each taking twice the
current of a single
winding.
SUMMARY: You get twice the output signal from a switched bifilar
winding, but you use
four times the power polarizing it. You therefore use up your
battery four times as
fast - probably faster because battery lifetime depends slightly on
current - the
greater the current, the smaller the number of amp-hours the battery
will supply. In
addition, you have the problem of the additional complexity.
However, if you are not
worried about power (i.e., if you are building an instrument for use
from a boat or a
vehicle) or complexity (more things to go wrong - and relays are the
weak link in
these sort of systems because they are electromechanical devices),
you DO get twice
the signal for the same number of turns of wire overall.
In general, the signal is proportional to polarizing current, times
the square of the
number of turns. If you want a bigger signal, it is easiest to get
it by just in-
creasing the number of turns. For a given battery voltage and wire
size, if you
double the number of turns, you double the resistance and so halve
the current. You
get, overall, a doubling of the signal. As well, because you've
reduced the polar-
izing current, you increase battery life. That is by simply
doubling the number of
turns, you get twice the signal and only use half the power. Jim
Koehler.