Measuring during the Tx Pulse ... ?

[moodz]

Hi all ... I have designed a PI detector that processes target information during .. not after .. the TX pulse. I was originally attempting to minimise the back EMF post TX and happened to find a means of detecting target metal during the TX pulse.
My questions are as follows ...
Are there any detectors that do this or am I wasting my time.
How should I develop the design ... is there a commercialisation angle ? ( hey I need money to fund my experiments ).
Funnily enough the design also appears to provide descrimination ( ferrous / non ferrous ). ( i am working on this )

Any advice TIA.

 

[!-Rob-!]

Hi,
Sometimes things work on the bench on scopes etc,
but you need to prove the effect detecting outside
in different ground conditions or wet sand swinging a coil in the earths mag field
to prove if you have anything viable worth going further with
"try everything never & give up"
Cheers

 

[Dave J.]

Back in the 80's I built a PI which demodulated the signals induced by the transmit pulse, to provide reactive ground balance and discrimination like in a VLF IB machine. It came close to commercialization, but people who field tested it weren't favorably impressed. We abandoned it in favor of multiple frequency technology which became the Fisher CZ series which is still in production. The pulse induction platform we yanked the reactive signal stuff back out of and turned it into an all-metals fully static underwater machine-- the Impulse, which was taken out of production by Fisher in Los Banos CA prior to its purchase by the investor group here in Texas.

--Dave J.

 

[TheMarshall]

Hi all ... I have designed a PI detector that processes target information during .. not after .. the TX pulse. I was originally attempting to minimise the back EMF post TX and happened to find a means of detecting target metal during the TX pulse.
My questions are as follows ...
[size=x-large]Are there any detectors that do this or am I wasting my time.[/size]
How should I develop the design ... is there a commercialisation angle ? ( hey I need money to fund my experiments ).
Funnily enough the design also appears to provide discrimination ( ferrous / non ferrous ). ( i am working on this )

Any advice TIA.

~~~~~~~~~~~~~~~~~~~~~~~~

Hello TIA.

IN answer to the highlighted section of your quoted post.

Candy, of Minelab fame, he samples during the TX pulse ( according to one aspect of his FBS patent)

Hope this is of interest..........TheMarshall

 

[moodz]

Thanks for that .. I will read the Minelab patent and post my thoughts.

 

[moodz]

Thanks for that .. I will read the Minelab patent and post my thoughts.

Ok ... that Candy guy is sure one busy guy at the patent office.

However the FSB patent appears to relate to DSP processing of multiple sinusoidal signals
This is not a PI detector. IMHO
The process I am describing is measuring during the TX pulse not after the pulse as in traditional PI.

My reasoning is this ... if you transmit a pulse with a certain energy and a percentage of this is 'coupled' to the target and then
the target 'transmits' a certain amount of energy of which a certain percentage is 'coupled' back to the receiver then the total loss of signal
from Tx to Rx is .... (T x L) x (R x L) where T is the transmit energy, L is the loss ( same for Tx and Rx path ) and R is the energy in the target.
(the actual formula is quite a bit more complex but this will do by way of explanation )
This means that the Loss is squared. eg if loss L = 100 then Loss from Tx to target and back again to Rx is 100 x 100 = 10000.
This means huge amplifications required at the RX sampler of a PI detector.
However if you could measure the Loss at the TX pulse the loss factor is only 100 not 10000.
The clue here is you have to accurately know the energy contained in the TX pulse and measuring any changes to this energy any 'lost'
energy must have been transferred to a target. The larger more conductive the target the more energy will be lost. ( generally )

Anyway thats my theory on how it works ... I am willing to argue that politely with anyone.

 

[moodz]

Thanks Rob for the encouragement ... of course the proof of the pudding is in the eating.

 

[moodz]

Thanks Dave J for your reply.
Were you disappointed that the design was not progressed ?
Oh well. In relation to demodulation I would not say I am trying to do that.
See my post above regarding Pulse Energy Loss.
--

 

[Dave J.]

Okay, I understand now that you're actually looking at the effect that the target has on the transmit signal, as opposed to using an induction balance to look at the target during the transmit interval.

This is approximately equivalent to the eddy current loss principle of detection, most often found in industrial metal detectors for stationary installation, and in "pinpointers". From the standpoint of signal loss at first it looks attractive compared to an induction balance. However without an induction balance, the drift and noise on the transmit signal tend to bury weaker changes resulting from the proximity of metal masses. Some of the Gardiner (no relation to John Gardiner of FTP-Fisher) metal detectors of the early 1980's were quite sophisticated eddy current loss systems, even capable of multiple frequency discrimination, but the basic performance just wasn't there.

There is a fairly expensive pinpointing probe (sorry don't remember the name) which has a waveform something like PI but which demodulates the transmit reactive component in order to distinguish ferrous from nonferrous. Being a pinpoint probe designed to be very nearly in physical contact with the target, the inability to take advantage of induction balance is not an issue.

--Dave J.

 

[robby_h]

The Sovereign/Explorer/Etrac transmits a continuous wave, multi period, rectangular waveform (no off-period other than the brief transistion from one polarity to the other) and the rx measures the complex FFT to extract numerous R and X frequency components to ID targets and cancel minerals. A simple version, or a single frequency vlf only gets one shot at R and X whereas much more information is gathered and processed when transmitting multiperiod rectangular waveforms.

It is true that if you transmit a near linear current ramp, you can easily see changes to the TX waveform when you bring various metal objects to the coil but the ground (high X) in most places (at least here in Australia) also has a dramatic effect and modulates the coil's inductance.

You might gain a bit more info from this paper written by Bruce Candy.

http://www.minelab.com/consumer/xstandard/files/ML_08_007WhitePaper8.pdf

Rob.

 

[moodz]

Thats very interesting feedback Dave, the design I am investigating uses a sense coil to detect the tx pulse when it is active.
I am effectively measuring the strength of the magnetic field as it progresses from zero to some maximal valued dependant on
the pulse length. ( di/dt change of current with time)
The power supply driving the pulse switch is regulated.
The drift and noise is fairly low pulse to pulse .. so any loss from pulse to pulse should be measureable.
moodz.

 

[moodz]

Rob ... so the Sovereign/Explorer/Etrac use FFT analysis whilst transmitting continuously.
I have an another project that probably does something similiar.
I used the sound card i/p & o/p on a small laptop ( eee pc ) to do the tx and rx.
The Tx was amplified using a 6 watt audio amp and balanced coils.
You can do alot of DSP in Pentium class processor even though this one was
running at a comparitively 'slow' 900 Mhz.

moodz

 

[Monolith]

I have been experimenting with sampling during the TX pulse for some time now and got very significant results. There are also several patents, some going back many years, for this method. I would love to compare some numbers on sensitivity and depth. I find that I get much higher signal amplitude when sampling during the TX pulse and I also get FE differentiation.
I use an Induction Balanced coil. IB
I find very little information on the Web about the use of IB coils for PI or Time domain detectors. Feedback would be most welcome.

Monolith

 

[moodz]

Hiya Monolith .. the design i amusing basically measures the mag field strength during the tx pulse. I look at two things the tail end of the curve where the mag field is nearly at peak value / max current and the integral or area of the curve. A difference comparison of the two values will provide a descrimination and target size / depth information. I have found that I have to run the pulse out to 400us to get good numbers though.
I am nowhere near a field test yet .. my low level amplifiers need more work.

 

[Reg]

Hi Monolith,

Eric Foster designed a PI many years ago that used a coaxial coil in sort of an IB configuration to discriminate. I believe the design was called the PPD1. There might be some info on this forum if one did a search. The design looked at the transmit pulse for discrimination information.

The Pulse Devil is doing something similar I believe.

Now, Allan Westersten has a patent on looking late on the transmit pulse for discrimination purposes. I understand this patent is in process or maybe has been granted by now. I think I spelled his name right. So, it is something a person might want to look at.

Reg

 

[Monolith]

Hi Reg,
thanks for the feedback.
David Emery's patent has been posted today on Geotech.
I have seen Allan Westerstens patent, will try to locate it again to post a link.
I use a different method than either of them. The differentiation between Ferrous and non ferrous metals is produced by the IB coil. I wonder why this has not been further developed, since the fact is known since a very long time. George Payne also took out a patent on this many years ago.
I have been sampling very early, with a delay of 3.5uS or all along the TX pulse until just before switch OFF. The differentiation (negative for FE and positive for non FE) is present during the whole TX pulse, however, the best sampling point depends very much on the TC of the coil (300uH with 10 Ohm = 30uS)
I get higher target response amplitude during TX than after switch OFF.
This is not easily explained with traditional PI theory, since the di/dt is less during the switch ON transient than during the switch OFF transient.
Monolith

 

[Monolith]

Hiya Monolith .. the design i amusing basically measures the mag field strength during the tx pulse. I look at two things the tail end of the curve where the mag field is nearly at peak value / max current and the integral or area of the curve. A difference comparison of the two values will provide a descrimination and target size / depth information. I have found that I have to run the pulse out to 400us to get good numbers though.
I am nowhere near a field test yet .. my low level amplifiers need more work.

Hi moodz,

If I could find out how to post a schematic here I would post a front-end circuit for you to play with. In the meantime Google for the patents mentioned above.

Tinkerer

 

[Reg]

Hi Monolith,

I couldn't find the link to Dave Emery's patent. If you can send me the link or post the link, I would appreciate it. Also, I am not sure he tried to patent the basic design of the PD. Part of that came from Eric's design, I believe.



Reg

 

[Monolith]

Reg,
here is the link to David Emery's patent
http://www.geotech1.com/forums/showthread.php?p=86952&highlight=David+Emery#post86952

 

[moodz]

Monolith ... had a look at the Emery Patent .... If I had known that passive resonance was patentable I would have done it some years ago as I played with a very similiar system before I got onto this Tx Pulse.
I dont think this patent covers measuring during the Tx pulse as the measurements are done after the resonant system has been excited. He continually refers to the 'high voltage' flyback which seems to imply the coil response after a coil pulse event. The key detail seems to be the decay oscillation response of the coil after excitation by a pulse.
I take it you are using an IB coil setup .. whereas I am using a monocoil.