Tweet
Forgot to include my scope pictures post 14. Amplifier out gain about 450. 5mm square coke can side looks similar to no target at amplifier out but easy to see at integrator out.
-
-
Hi Green ...
I am still working out the math on it ...
but I am getting the strongest target signals during the current transitions (eg 2 microseconds ) ... not 7 microseconds later.
So I will be using samplfication ( sample then amplify ).
The response is prob due to fairly large targets just unbalancing the DD.
I can get much smaller delays ... theres a large capacitance across the TX coil ... BUT if you knock down the decay ... peak flyback goes up and you have to drop the peak TX currents.
WIth 2 amp swing the peak voltage at the SIC mosfets is 1100 volts ( rated at 1200 volts ).
At 1 amp peak to peak ... we could manage 1.6 microseconds.
No free lunch.
moodzComment
-
A normal PI looks at target decay after the target charge signal, first 4us. Don't know how ground balance could be done looking at signal during target charge. Wondering what is causing the no target decay signal after 4us+. Maybe coil wire? AWG24 has a TC near .8us about the no target decay signal.Originally posted by moodz View PostComment
-
Moodz, Sample during the transition (S1), then after the Tx current is constant (S2). S1 and S2 will have opposite polarities with metal targets. A differential integrator will sum these and output a usable signal. S1 will need to be adjusted in time to sample during the ferrite(GB) pivot. It's a lot more sensitive than normal subtractive pi sampling (S1 -EF)Originally posted by moodz View Post
This is the method I have been using for my THS pi's​Comment
-
Curious what you consider S1 and S2. Moodz scope picture post #14 has three times (0 to 1.5us) (1.5 to 4us) and 4+us.Originally posted by Altra View PostComment
-
In post 2, the sim scope shot1 "Rx waveform in PI mode", I assume he is using an induction balance coil. Sample S1 would be during the transition or voltage spike. Assuming a well nulled coil. Sample S2 would be to the right of the spike like a normal delayed pi sample. Post 14# looks like super early target responses which will include reactive ground signal.Originally posted by green View PostComment
-
...This circuit is not a regular PI circuit ... its a resonant system that ping pongs the coil energy at each current transition. The energy builds up in the circuit like a flywheel.
You can see this effect from the start up from power on. The Q of the resonant circuit on my test bed is about 100. ( there is no damping resistor )
The presence of a reactive ground will affect this Q and thus the energy stored in the coil each cycle.
This is where the current sensing comes in useful ... I am thinking I will use it to ground balance.
Below is a pic of the current waveform in the TX coil ... with (blue) and without(green) reactive ground.
The presence of ground causes a decrease in the peak coil current.
The RX signal shown below also shows the expected result for ground during the coil transition.
Comment
-
I will try your method first though Altra .... noting also the other comments from G.Originally posted by Altra View PostComment
-
Is the tilt polarity correct for the green trace post #22? Why less tilt for the blue trace?Comment
-
In theory it is possible what I proposed above. But after a second thought it may be a challenge. The very narrow sample window during the transition (<3us ) could be a limitation? A THS pi can have a typical half sine width of 30-200us to sample from, so lots of signal to work with. If anything, you may be able to do fe/non-fe determination? Worth exploring. Thanks for posting your projectsOriginally posted by moodz View PostComment
-
With the ground present the viscous matrix absorbs energy from the H field but does not return it during flyback. Any energy losses ( or returns ) will affect the tilt.Originally posted by green View Post
Thats my understanding anyway.Comment
-
I have modified a MAGPI 3 board to act as a controller / RX for the BPCCPI TX and can sample down to 100 nanoseconds minimum width with 20 nanosecond timing resolution.Originally posted by Altra View Post
Have breadboarded the circuit ... The pic below shows the gate switch ( yellow ) the output of the first preamp .. gain = 10 ( blue ) and the flyback voltage ( pink ) at the coil.
The sampler is followed by a high gain diff integrator and 12 bit ADC into the micro.
Will see what happens.
moodz
Comment
-
Question: On a DD coil ( this is a minelab ) ... what impedance should be loaded across the RX .. if any ?Comment
-
ML DD coils have a built-in damping resistor for the rx side. Don't remember the value?Comment
-
Comment