Properties of the Magnetic Wire Tension Meter ( WTM ) prototype .
25-1-2003 /  T. Sluijk , NIKHEF

Using the Wire Tension Meter electronics produced  by Iouri Gouz,  its criticalities where checked.
The following questions where treated.

1. how does the excitation work?
2. how critical is the magnet positioning?
3. signals on the oscilloscope

1 How does the excitation work?

There are 2 modes:

1. single shot mode  (the hammer pulse is repeated every 1 second)
2. Continuous mode: The hammer pulse  is given every new resonance period.

Single shot mode:
We use a long magnet under the wire. A current pulse of about 130 mA, 0.8mseclong, is send through the wire. In the magnetic field of 0.05 Tesla there is a force on the wire (.. gram each cm) pushing it to the rigth.
This depends on the polarities of coarse. The swing of the 25um wire is limited to less then  2.5 mm
to prevent the wire from hitting the wall of the r=2.5mm straw it is placed in. This is done by using the used current at 0.8ms with the used magnet length. After this the wire starts resonating like a string of a guitar, for many milli seconds. (decay time ~400 msec) The tone height is a measure for the wire tension of a certain length of wire. (frequency, or period time). The signal can be noticed as soon as the hammer pulse has disappeared. The moving wire in a magnetic fied is like a coil flipping from plus to minus, because the centre of gravity of the ground stays in the centre while the wire moves from + to -. Giving a reverse of polarity every period of the swing. The open voltage can be calculated from the loop surface ~10 cm2(effective)  and the magnetic field ~0.05T. 
U=d PHI/dt = B.dO/dt  = ~10mV (for a square wave). The real signal is more like 1-5 mV

Continuous Mode:
The excitation  current pulse in this mode is short and comes every zero crossing (or peak) of the resonance. The signal is monitored and the pulse duration (fixed current) is increased until there is enough signal of the rectified sine wave to produce detectable signal levels (1-4mV out of the wire).  The period time /frequency can be measured.

2. Is the magnet position critical?

Yes. 
Do not put the magnet perpendicular under the wire. The magnet should be in the length direction to avoid measurement errors like 150Hz instead of 140 Hz (This due to non symmetrical signal lobs)
The magnet should be placed within + or - 7.5cm away from the centre. For a wire length of 76 cm. (10%)
Not sensitive is the magnet distance.  The magnet can be moved away 7cm the WTM fails to indicate properly.
All the time the measurement is correct to within 1 Hz on 140Hz.

3. Scope Measurements, using the WTM prototype to generate and control the excitation signal.

All measurements where taken with the following conditions unless changed
in the sequence of photo' s taken.

*single shot mode with 1 pulse every second, 60V high , 0.8 msec long (constant)
- wire of 76 cm long between PCB pads in aluminum U-profile of 4*4cm 80 cm long, 2mm thick
- prototype connected through 180 ohm resistor to simulate 2m wire in series with 76cm piece.
-shorted end of wire.
-magnet in middle, in length direction, and against bottom of U profile.
- wire height= 2.5mm above aluminum
-meter reads 7.1 msec, ( or 140 Hz in continuous mode)
-magnet brick size: 15cm*7.5 cm *2cm. Big flat surface=north or south pole, ~1 kG on surface.
 see graph for decay.

fig 1 : 0.8msec 60V pulse measured on generator output

fig decay of signal in 1Hz sigle shot mode

fig2: signal just after "hammer pulse" 6mV peak to peak (pp) magnet against alu profile 

fig3  now 4 msec timebase. The period time is about 7msec. and decays. 6mVpp

fig4: Now magnet distance 2.0 cm away from alu profile bottom. U=2mVpp

fig5:  hammer pulse after 180 resistor, 10V, 0.8msec

fig6: now magnet to alu distance 4cm, signal 0.5mVpp, but meter still reliable reading.

fig7 2cm again, probe after 180 ohm for this photo.

fig8: now magnet at 2cm but perpendicular to alu profile. Signal much smaller.
Not so reliable value on meter.

fig 9: same averaged 16 times.

fig 10: 1cm between magnet and profile magnet in length direction 3Vpp

fig 11: Now magnet at 30% from end, in length direction still. Notice bottom reflections from 
both wire ends shift in time and don't add up as 1.

fig 12 : Now almost at end. Magnet hart 15cm  from far end (opposite to generator)

fig 13: dito but 20ms/div timebase

fig 14: 7.5 cm fom end (magnet hart)
Does the device use positive lob or negative? is polarity north south important?

fig 15: now 7.5 cm from beginning. looks the same again.

The usable area where the lobs still add, is 13.5 cm if we use the southpole up.

fig16: Now 1 magnet 1cm away and magnet2 on top of profile and 5mV/div instead of 1mV/div Upp=12mV

fig 17: now 1 magnet again at 1cm from profile and 1mV/div U=3mVpp 4 times lower!

fig 18 : Now meter in Hz mode= Continuous meter gives 140 Hz . A hammer pulse is given
on every positive top of the signal pulse duration is regulated for constant signal size of 4mVpp

fig :19 now short timebase to see the short 40usec. hammer pulse (max=800usec.) 
The time near to baseline is 200usec however! The pulse is going up although it does look
because of the undershoot.

fig 20: now 20usec/div Same polarity of scope! now 50V/div however.

fig22: now at end -7.5cm. This needs a ~100usec exitation. 
The meter now shows 330Hz however! (326..332Hz)  (280Hz on end-10cm, 60usec pulse)
Range seems 13.5 cm (+7, -7cm from center position) moving from outside.
Starting from middle gives "hold range of 40cm"before frequency goes wrong.
Hysteresis.

fig 23:  again 1cm distance for magnet. 30us needed.

fig 24: 2 cm distance, 40usec needed.
          4 cm needs 110 usec.

fig 27: now 7cm distance. Range still 13.5 cm for 140Hz
          7cm largest distance! meter then swaps to 800usec(see below)

fig 28: Magnet far away, Tmax=800usec.

fig 29: signal at 7cm distance still 1mV pp

fig 30: signal at 6cm magnet distance signal now 1.5mVpp instead of 1mVpp

fig 31: signal with magnet at 4cm distance, 4mVpp

fig 32: now 1 cm: U= 4mVpp

Do not put the magnet perpendicular. The meter can display 150 Hz istead of 140 7.23ms instead of 7.1ms

Magnet positioning sensitivity:

fig1:  magnet placement sensitivity. Magnet in centre in length direction

fig2 Magnet now in perpendicular position

fig3: magnet now shifted 7 cm away from centre

Frequency domain display:

fig bandwidth of peak=~1Hz = measurement precision.  Using a single shot of
10 divisions of 100msec/div= 1 second = 1 single shot before the next hammer pulse is fired.

frequency with Fourier Transform= 140Hz

Is there a limit to the measurement precision or resolution?

Signal in single shot mode:

fig: Decay time about 400msec= 57 periods, Qfactor~57

In continuous mode we can take ever increasing time bases. To have more signal periods,
and look at the frequency peak width.

fig : 10msec/div = 100msec signal peak width 10 Hz is equivalent. No problems with the hammer spike!
Peak= 140 Hz. Fourier 50Hz/div.

fig now 20ms/div=200msec of signal, peak width = 5Hz (3dB)

fig: now 40msec/div

fig 100msec/div= 1 second of signal, peak width 1 Hz

fig: 1sec/div= 10 sec signal= 0,1 Hz peak width see below for magnified frequency scale with same
sample time of 10 seconds

fig : same but F enlarged. Peak now 100 mHz (0.5 division) also measured with cursors.
The limit is not yet reached. A resolution of 0.1Hz would be possible!